From 990bc3f015a4f8fce2eb918375defcd44980a845 Mon Sep 17 00:00:00 2001 From: marha Date: Fri, 8 Jun 2012 09:33:13 +0200 Subject: Used synchronise script to update files --- mesalib/src/mesa/.gitignore | 5 + mesalib/src/mesa/drivers/common/driverfuncs.c | 20 +- mesalib/src/mesa/drivers/common/driverfuncs.h | 74 +- mesalib/src/mesa/drivers/common/meta.c | 80 +- mesalib/src/mesa/drivers/common/meta.h | 27 +- mesalib/src/mesa/drivers/dri/.gitignore | 1 + .../src/mesa/drivers/dri/common/xmlpool/.gitignore | 5 + .../src/mesa/drivers/dri/common/xmlpool/Makefile | 192 +- .../mesa/drivers/dri/common/xmlpool/gen_xmlpool.py | 382 +-- mesalib/src/mesa/drivers/dri/swrast/.gitignore | 1 + mesalib/src/mesa/gl.pc.in | 26 +- mesalib/src/mesa/main/.gitignore | 13 + mesalib/src/mesa/main/api_loopback.c | 3450 +++++++++---------- mesalib/src/mesa/main/atifragshader.c | 1592 ++++----- mesalib/src/mesa/main/blend.h | 212 +- mesalib/src/mesa/main/bufferobj.c | 19 + mesalib/src/mesa/main/clear.h | 128 +- mesalib/src/mesa/main/clip.c | 232 +- mesalib/src/mesa/main/clip.h | 94 +- mesalib/src/mesa/main/condrender.h | 90 +- mesalib/src/mesa/main/dd.h | 192 +- mesalib/src/mesa/main/depth.h | 136 +- mesalib/src/mesa/main/drawtex.h | 122 +- mesalib/src/mesa/main/enable.h | 142 +- mesalib/src/mesa/main/extensions.c | 1 + mesalib/src/mesa/main/ffvertex_prog.h | 80 +- mesalib/src/mesa/main/fog.h | 138 +- mesalib/src/mesa/main/formats.c | 8 +- mesalib/src/mesa/main/hint.h | 118 +- mesalib/src/mesa/main/lines.h | 98 +- mesalib/src/mesa/main/matrix.h | 226 +- mesalib/src/mesa/main/mipmap.c | 8 +- mesalib/src/mesa/main/mtypes.h | 1 + mesalib/src/mesa/main/multisample.c | 124 +- mesalib/src/mesa/main/multisample.h | 82 +- mesalib/src/mesa/main/pixelstore.h | 104 +- mesalib/src/mesa/main/points.c | 544 +-- mesalib/src/mesa/main/points.h | 118 +- mesalib/src/mesa/main/rastpos.c | 1132 +++---- mesalib/src/mesa/main/remap.c | 452 +-- mesalib/src/mesa/main/scissor.c | 200 +- mesalib/src/mesa/main/scissor.h | 94 +- mesalib/src/mesa/main/stencil.h | 162 +- mesalib/src/mesa/main/syncobj.c | 852 ++--- mesalib/src/mesa/main/texformat.h | 78 +- mesalib/src/mesa/main/texgen.c | 802 ++--- mesalib/src/mesa/main/teximage.c | 151 +- mesalib/src/mesa/main/texobj.c | 28 +- mesalib/src/mesa/main/texstore.c | 259 +- mesalib/src/mesa/main/texstore.h | 113 +- mesalib/src/mesa/main/texturebarrier.c | 108 +- mesalib/src/mesa/main/texturebarrier.h | 88 +- mesalib/src/mesa/main/transformfeedback.c | 21 +- mesalib/src/mesa/main/version.c | 1 + mesalib/src/mesa/main/viewport.c | 370 +-- mesalib/src/mesa/main/viewport.h | 116 +- mesalib/src/mesa/math/m_debug_clip.c | 818 ++--- mesalib/src/mesa/math/m_debug_norm.c | 766 ++--- mesalib/src/mesa/math/m_debug_xform.c | 678 ++-- mesalib/src/mesa/math/m_matrix.c | 3282 +++++++++--------- mesalib/src/mesa/math/m_vector.c | 368 +-- mesalib/src/mesa/program/.gitignore | 4 + mesalib/src/mesa/program/arbprogparse.h | 90 +- mesalib/src/mesa/program/ir_to_mesa.cpp | 5 + mesalib/src/mesa/program/nvfragparse.h | 94 +- mesalib/src/mesa/program/nvvertparse.h | 98 +- mesalib/src/mesa/program/prog_cache.c | 502 +-- mesalib/src/mesa/program/prog_cache.h | 136 +- mesalib/src/mesa/program/sampler.h | 58 +- mesalib/src/mesa/program/symbol_table.h | 124 +- mesalib/src/mesa/state_tracker/Makefile | 2 +- mesalib/src/mesa/state_tracker/st_atom.h | 146 +- mesalib/src/mesa/state_tracker/st_atom_constbuf.h | 82 +- mesalib/src/mesa/state_tracker/st_atom_msaa.c | 166 +- mesalib/src/mesa/state_tracker/st_atom_shader.h | 82 +- mesalib/src/mesa/state_tracker/st_atom_viewport.c | 188 +- mesalib/src/mesa/state_tracker/st_cache.h | 138 +- mesalib/src/mesa/state_tracker/st_cb_bitmap.h | 192 +- mesalib/src/mesa/state_tracker/st_cb_blit.h | 118 +- .../src/mesa/state_tracker/st_cb_bufferobjects.h | 136 +- mesalib/src/mesa/state_tracker/st_cb_condrender.h | 74 +- mesalib/src/mesa/state_tracker/st_cb_drawpixels.h | 140 +- mesalib/src/mesa/state_tracker/st_cb_drawtex.h | 82 +- mesalib/src/mesa/state_tracker/st_cb_eglimage.h | 102 +- mesalib/src/mesa/state_tracker/st_cb_feedback.c | 618 ++-- mesalib/src/mesa/state_tracker/st_cb_feedback.h | 104 +- mesalib/src/mesa/state_tracker/st_cb_flush.h | 102 +- mesalib/src/mesa/state_tracker/st_cb_program.h | 76 +- mesalib/src/mesa/state_tracker/st_cb_queryobj.h | 140 +- mesalib/src/mesa/state_tracker/st_cb_rasterpos.h | 100 +- mesalib/src/mesa/state_tracker/st_cb_strings.c | 146 +- mesalib/src/mesa/state_tracker/st_cb_strings.h | 80 +- mesalib/src/mesa/state_tracker/st_cb_syncobj.c | 244 +- mesalib/src/mesa/state_tracker/st_cb_texture.c | 135 +- .../src/mesa/state_tracker/st_cb_texturebarrier.c | 120 +- .../src/mesa/state_tracker/st_cb_texturebarrier.h | 74 +- mesalib/src/mesa/state_tracker/st_cb_viewport.h | 72 +- mesalib/src/mesa/state_tracker/st_debug.h | 144 +- mesalib/src/mesa/state_tracker/st_extensions.c | 4 + mesalib/src/mesa/state_tracker/st_extensions.h | 80 +- mesalib/src/mesa/state_tracker/st_gen_mipmap.h | 104 +- mesalib/src/mesa/state_tracker/st_gl_api.h | 14 +- mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp | 6 + mesalib/src/mesa/state_tracker/st_manager.h | 104 +- mesalib/src/mesa/swrast/NOTES | 108 +- mesalib/src/mesa/swrast/s_aaline.h | 76 +- mesalib/src/mesa/swrast/s_aatriangle.h | 76 +- mesalib/src/mesa/swrast/s_alpha.c | 320 +- mesalib/src/mesa/swrast/s_alpha.h | 78 +- mesalib/src/mesa/swrast/s_atifragshader.c | 1208 +++---- mesalib/src/mesa/swrast/s_atifragshader.h | 76 +- mesalib/src/mesa/swrast/s_blend.h | 90 +- mesalib/src/mesa/swrast/s_feedback.c | 276 +- mesalib/src/mesa/swrast/s_feedback.h | 100 +- mesalib/src/mesa/swrast/s_fog.h | 84 +- mesalib/src/mesa/swrast/s_lines.h | 82 +- mesalib/src/mesa/swrast/s_logic.h | 80 +- mesalib/src/mesa/swrast/s_masking.h | 82 +- mesalib/src/mesa/swrast/s_points.h | 78 +- mesalib/src/mesa/swrast/s_texcombine.h | 74 +- mesalib/src/mesa/swrast/s_triangle.h | 100 +- mesalib/src/mesa/swrast_setup/NOTES | 130 +- mesalib/src/mesa/swrast_setup/ss_triangle.c | 530 +-- mesalib/src/mesa/swrast_setup/ss_triangle.h | 76 +- mesalib/src/mesa/swrast_setup/ss_tritmp.h | 496 +-- mesalib/src/mesa/swrast_setup/ss_vb.h | 74 +- mesalib/src/mesa/swrast_setup/swrast_setup.h | 122 +- mesalib/src/mesa/tnl/t_pipeline.h | 146 +- mesalib/src/mesa/tnl/t_vb_fog.c | 552 ++-- mesalib/src/mesa/tnl/t_vb_normals.c | 374 +-- mesalib/src/mesa/tnl/t_vb_points.c | 228 +- mesalib/src/mesa/tnl/t_vb_render.c | 696 ++-- mesalib/src/mesa/tnl/t_vb_rendertmp.h | 972 +++--- mesalib/src/mesa/tnl/t_vb_texgen.c | 1218 +++---- mesalib/src/mesa/tnl/t_vb_texmat.c | 256 +- mesalib/src/mesa/tnl/t_vb_vertex.c | 566 ++-- mesalib/src/mesa/tnl/t_vertex.h | 368 +-- mesalib/src/mesa/tnl/t_vp_build.c | 120 +- mesalib/src/mesa/tnl/t_vp_build.h | 84 +- mesalib/src/mesa/vbo/vbo_exec_eval.c | 526 +-- mesalib/src/mesa/vbo/vbo_save_loopback.c | 392 +-- mesalib/src/mesa/vbo/vbo_split.h | 144 +- mesalib/src/mesa/x86/3dnow.c | 182 +- mesalib/src/mesa/x86/3dnow.h | 72 +- mesalib/src/mesa/x86/3dnow_normal.S | 1704 +++++----- mesalib/src/mesa/x86/3dnow_xform1.S | 874 ++--- mesalib/src/mesa/x86/3dnow_xform2.S | 954 +++--- mesalib/src/mesa/x86/3dnow_xform3.S | 1122 +++---- mesalib/src/mesa/x86/3dnow_xform4.S | 1140 +++---- mesalib/src/mesa/x86/Makefile | 92 +- mesalib/src/mesa/x86/assyntax.h | 3494 ++++++++++---------- mesalib/src/mesa/x86/clip_args.h | 118 +- mesalib/src/mesa/x86/common_x86.c | 672 ++-- mesalib/src/mesa/x86/common_x86_asm.S | 440 +-- mesalib/src/mesa/x86/common_x86_asm.h | 106 +- mesalib/src/mesa/x86/common_x86_features.h | 134 +- mesalib/src/mesa/x86/mmx.h | 118 +- mesalib/src/mesa/x86/mmx_blend.S | 804 ++--- mesalib/src/mesa/x86/mmx_blendtmp.h | 228 +- mesalib/src/mesa/x86/norm_args.h | 114 +- mesalib/src/mesa/x86/read_rgba_span_x86.S | 1372 ++++---- mesalib/src/mesa/x86/read_rgba_span_x86.h | 112 +- mesalib/src/mesa/x86/rtasm/x86sse.c | 2406 +++++++------- mesalib/src/mesa/x86/rtasm/x86sse.h | 512 +-- mesalib/src/mesa/x86/sse.c | 246 +- mesalib/src/mesa/x86/sse.h | 72 +- mesalib/src/mesa/x86/sse_normal.S | 522 +-- mesalib/src/mesa/x86/sse_xform1.S | 892 ++--- mesalib/src/mesa/x86/sse_xform2.S | 932 +++--- mesalib/src/mesa/x86/sse_xform3.S | 1024 +++--- mesalib/src/mesa/x86/sse_xform4.S | 470 +-- mesalib/src/mesa/x86/x86_cliptest.S | 814 ++--- mesalib/src/mesa/x86/x86_xform.c | 252 +- mesalib/src/mesa/x86/x86_xform.h | 212 +- mesalib/src/mesa/x86/x86_xform2.S | 1148 +++---- mesalib/src/mesa/x86/x86_xform3.S | 1288 ++++---- mesalib/src/mesa/x86/x86_xform4.S | 1354 ++++---- mesalib/src/mesa/x86/xform_args.h | 102 +- 178 files changed, 30840 insertions(+), 31366 deletions(-) create mode 100644 mesalib/src/mesa/.gitignore create mode 100644 mesalib/src/mesa/drivers/dri/.gitignore create mode 100644 mesalib/src/mesa/drivers/dri/common/xmlpool/.gitignore create mode 100644 mesalib/src/mesa/drivers/dri/swrast/.gitignore create mode 100644 mesalib/src/mesa/main/.gitignore create mode 100644 mesalib/src/mesa/program/.gitignore (limited to 'mesalib/src/mesa') diff --git a/mesalib/src/mesa/.gitignore b/mesalib/src/mesa/.gitignore new file mode 100644 index 000000000..ce83eaf47 --- /dev/null +++ b/mesalib/src/mesa/.gitignore @@ -0,0 +1,5 @@ +*/gen_matypes +*/matypes.h +depend.es* +depend.es* +objs-es* diff --git a/mesalib/src/mesa/drivers/common/driverfuncs.c b/mesalib/src/mesa/drivers/common/driverfuncs.c index ca120578d..93fa3c745 100644 --- a/mesalib/src/mesa/drivers/common/driverfuncs.c +++ b/mesalib/src/mesa/drivers/common/driverfuncs.c @@ -91,24 +91,14 @@ _mesa_init_driver_functions(struct dd_function_table *driver) /* Texture functions */ driver->ChooseTextureFormat = _mesa_choose_tex_format; - driver->TexImage1D = _mesa_store_teximage1d; - driver->TexImage2D = _mesa_store_teximage2d; - driver->TexImage3D = _mesa_store_teximage3d; - driver->TexSubImage1D = _mesa_store_texsubimage1d; - driver->TexSubImage2D = _mesa_store_texsubimage2d; - driver->TexSubImage3D = _mesa_store_texsubimage3d; + driver->TexImage = _mesa_store_teximage; + driver->TexSubImage = _mesa_store_texsubimage; driver->GetTexImage = _mesa_meta_GetTexImage; - driver->CopyTexSubImage1D = _mesa_meta_CopyTexSubImage1D; - driver->CopyTexSubImage2D = _mesa_meta_CopyTexSubImage2D; - driver->CopyTexSubImage3D = _mesa_meta_CopyTexSubImage3D; + driver->CopyTexSubImage = _mesa_meta_CopyTexSubImage; driver->GenerateMipmap = _mesa_meta_GenerateMipmap; driver->TestProxyTexImage = _mesa_test_proxy_teximage; - driver->CompressedTexImage1D = _mesa_store_compressed_teximage1d; - driver->CompressedTexImage2D = _mesa_store_compressed_teximage2d; - driver->CompressedTexImage3D = _mesa_store_compressed_teximage3d; - driver->CompressedTexSubImage1D = _mesa_store_compressed_texsubimage1d; - driver->CompressedTexSubImage2D = _mesa_store_compressed_texsubimage2d; - driver->CompressedTexSubImage3D = _mesa_store_compressed_texsubimage3d; + driver->CompressedTexImage = _mesa_store_compressed_teximage; + driver->CompressedTexSubImage = _mesa_store_compressed_texsubimage; driver->GetCompressedTexImage = _mesa_get_compressed_teximage; driver->BindTexture = NULL; driver->NewTextureObject = _mesa_new_texture_object; diff --git a/mesalib/src/mesa/drivers/common/driverfuncs.h b/mesalib/src/mesa/drivers/common/driverfuncs.h index 31549f65a..212f30742 100644 --- a/mesalib/src/mesa/drivers/common/driverfuncs.h +++ b/mesalib/src/mesa/drivers/common/driverfuncs.h @@ -1,37 +1,37 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef DRIVERFUNCS_H -#define DRIVERFUNCS_H - -extern void -_mesa_init_driver_functions(struct dd_function_table *driver); - - -extern void -_mesa_init_driver_state(struct gl_context *ctx); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef DRIVERFUNCS_H +#define DRIVERFUNCS_H + +extern void +_mesa_init_driver_functions(struct dd_function_table *driver); + + +extern void +_mesa_init_driver_state(struct gl_context *ctx); + + +#endif diff --git a/mesalib/src/mesa/drivers/common/meta.c b/mesalib/src/mesa/drivers/common/meta.c index a20e41972..8d7e90126 100644 --- a/mesalib/src/mesa/drivers/common/meta.c +++ b/mesalib/src/mesa/drivers/common/meta.c @@ -3117,17 +3117,15 @@ get_temp_image_type(struct gl_context *ctx, GLenum baseFormat) * Helper for _mesa_meta_CopyTexSubImage1/2/3D() functions. * Have to be careful with locking and meta state for pixel transfer. */ -static void -copy_tex_sub_image(struct gl_context *ctx, - GLuint dims, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, - GLsizei width, GLsizei height) +void +_mesa_meta_CopyTexSubImage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + struct gl_renderbuffer *rb, + GLint x, GLint y, + GLsizei width, GLsizei height) { struct gl_texture_object *texObj = texImage->TexObject; - const GLenum target = texObj->Target; GLenum format, type; GLint bpp; void *buf; @@ -3152,7 +3150,7 @@ copy_tex_sub_image(struct gl_context *ctx, type = get_temp_image_type(ctx, format); bpp = _mesa_bytes_per_pixel(format, type); if (bpp <= 0) { - _mesa_problem(ctx, "Bad bpp in meta copy_tex_sub_image()"); + _mesa_problem(ctx, "Bad bpp in _mesa_meta_CopyTexSubImage()"); return; } @@ -3181,21 +3179,11 @@ copy_tex_sub_image(struct gl_context *ctx, * Store texture data (with pixel transfer ops) */ _mesa_meta_begin(ctx, MESA_META_PIXEL_STORE); - if (target == GL_TEXTURE_1D) { - ctx->Driver.TexSubImage1D(ctx, texImage, - xoffset, width, - format, type, buf, &ctx->Unpack); - } - else if (target == GL_TEXTURE_3D) { - ctx->Driver.TexSubImage3D(ctx, texImage, - xoffset, yoffset, zoffset, width, height, 1, - format, type, buf, &ctx->Unpack); - } - else { - ctx->Driver.TexSubImage2D(ctx, texImage, - xoffset, yoffset, width, height, - format, type, buf, &ctx->Unpack); - } + + ctx->Driver.TexSubImage(ctx, dims, texImage, + xoffset, yoffset, zoffset, width, height, 1, + format, type, buf, &ctx->Unpack); + _mesa_meta_end(ctx); _mesa_lock_texture(ctx, texObj); /* re-lock */ @@ -3204,44 +3192,6 @@ copy_tex_sub_image(struct gl_context *ctx, } -void -_mesa_meta_CopyTexSubImage1D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, GLsizei width) -{ - copy_tex_sub_image(ctx, 1, texImage, xoffset, 0, 0, - rb, x, y, width, 1); -} - - -void -_mesa_meta_CopyTexSubImage2D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, - GLsizei width, GLsizei height) -{ - copy_tex_sub_image(ctx, 2, texImage, xoffset, yoffset, 0, - rb, x, y, width, height); -} - - -void -_mesa_meta_CopyTexSubImage3D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, - GLsizei width, GLsizei height) -{ - copy_tex_sub_image(ctx, 3, texImage, xoffset, yoffset, zoffset, - rb, x, y, width, height); -} - - /** * Decompress a texture image by drawing a quad with the compressed * texture and reading the pixels out of the color buffer. @@ -3269,6 +3219,7 @@ decompress_texture_image(struct gl_context *ctx, }; struct vertex verts[4]; GLuint fboDrawSave, fboReadSave; + GLuint rbSave; if (slice > 0) { assert(target == GL_TEXTURE_3D || @@ -3285,6 +3236,7 @@ decompress_texture_image(struct gl_context *ctx, /* save fbo bindings (not saved by _mesa_meta_begin()) */ fboDrawSave = ctx->DrawBuffer->Name; fboReadSave = ctx->ReadBuffer->Name; + rbSave = ctx->CurrentRenderbuffer ? ctx->CurrentRenderbuffer->Name : 0; _mesa_meta_begin(ctx, MESA_META_ALL & ~MESA_META_PIXEL_STORE); @@ -3305,6 +3257,7 @@ decompress_texture_image(struct gl_context *ctx, /* alloc dest surface */ if (width > decompress->Width || height > decompress->Height) { + _mesa_BindRenderbufferEXT(GL_RENDERBUFFER_EXT, decompress->RBO); _mesa_RenderbufferStorageEXT(GL_RENDERBUFFER_EXT, GL_RGBA, width, height); decompress->Width = width; @@ -3437,6 +3390,7 @@ decompress_texture_image(struct gl_context *ctx, _mesa_BindFramebufferEXT(GL_DRAW_FRAMEBUFFER_EXT, fboDrawSave); _mesa_BindFramebufferEXT(GL_READ_FRAMEBUFFER_EXT, fboReadSave); } + _mesa_BindRenderbufferEXT(GL_RENDERBUFFER_EXT, rbSave); } diff --git a/mesalib/src/mesa/drivers/common/meta.h b/mesalib/src/mesa/drivers/common/meta.h index de039b575..7a80b1dde 100644 --- a/mesalib/src/mesa/drivers/common/meta.h +++ b/mesalib/src/mesa/drivers/common/meta.h @@ -111,27 +111,12 @@ _mesa_meta_GenerateMipmap(struct gl_context *ctx, GLenum target, struct gl_texture_object *texObj); extern void -_mesa_meta_CopyTexSubImage1D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, GLsizei width); - -extern void -_mesa_meta_CopyTexSubImage2D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, - GLsizei width, GLsizei height); - -extern void -_mesa_meta_CopyTexSubImage3D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, - GLsizei width, GLsizei height); +_mesa_meta_CopyTexSubImage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + struct gl_renderbuffer *rb, + GLint x, GLint y, + GLsizei width, GLsizei height); extern void _mesa_meta_GetTexImage(struct gl_context *ctx, diff --git a/mesalib/src/mesa/drivers/dri/.gitignore b/mesalib/src/mesa/drivers/dri/.gitignore new file mode 100644 index 000000000..f3c7a7c5d --- /dev/null +++ b/mesalib/src/mesa/drivers/dri/.gitignore @@ -0,0 +1 @@ +Makefile diff --git a/mesalib/src/mesa/drivers/dri/common/xmlpool/.gitignore b/mesalib/src/mesa/drivers/dri/common/xmlpool/.gitignore new file mode 100644 index 000000000..a5a437849 --- /dev/null +++ b/mesalib/src/mesa/drivers/dri/common/xmlpool/.gitignore @@ -0,0 +1,5 @@ +de +es +fr +nl +sv diff --git a/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile b/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile index ae3d89938..b71629e9f 100644 --- a/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile +++ b/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile @@ -1,96 +1,96 @@ -# Convenient makefile for managing translations. - -# Prerequisites: -# - GNU gettext -# - Python - -# Adding new translations -# ----------------------- - -# To start working on a new translation edit the POS=... line -# below. If you want to add for example a french translation, add -# fr.po. - -# Then run "make po" to generate a fresh .po file from translatable -# strings in t_options.h. Now you can edit the new .po file (fr.po in -# the example above) to translate the strings. Please make sure that -# your editor encodes the file in UTF-8. - -# Updating existing translations -# ------------------------------ - -# Run "make po" to update .po files with new translatable strings from -# t_options.h. Now you can edit the .po files you're interested -# in. Please make sure that your editor encodes the file in UTF-8. - -# Updating options.h -# ------------------ - -# Finally run "make" to generate options.h from t_options.h with all -# translations. Now you can rebuild the drivers. Any common options -# used by the drivers will have option descriptions with the latest -# translations. - -# Publishing translations -# ----------------------- - -# To get your translation(s) into Mesa CVS, please send me your -# .po file. - -# More information: -# - info gettext - -# The set of supported languages. Add languages as needed. -POS=de.po es.po nl.po fr.po sv.po - -# -# Don't change anything below, unless you know what you're doing. -# -LANGS=$(POS:%.po=%) -MOS=$(POS:%.po=%/LC_MESSAGES/options.mo) -POT=xmlpool.pot - -.PHONY: all clean pot po mo - -all: options.h - -# Only intermediate files are cleaned up. options.h is not deleted because -# it's in CVS. -clean: - -rm -f $(POT) *~ - -rm -rf $(LANGS) - -# Default target options.h -options.h: t_options.h mo - $(PYTHON2) $(PYTHON_FLAGS) gen_xmlpool.py $(LANGS) > options.h - -# Update .mo files from the corresponding .po files. -mo: - @for mo in $(MOS); do \ - lang=$${mo%%/*}; \ - echo "Updating $$mo from $$lang.po."; \ - mkdir -p $${mo%/*}; \ - msgfmt -o $$mo $$lang.po; \ - done - -# Use this target to create or update .po files with new messages in -# driconf.py. -po: $(POS) - -pot: $(POT) - -# Extract message catalog from driconf.py. -$(POT): t_options.h - xgettext -L C --from-code utf-8 -o $(POT) t_options.h - -# Create or update a .po file for a specific language. -%.po: $(POT) - @if [ -f $@ ]; then \ - echo "Merging new strings from $(POT) into $@."; \ - mv $@ $@~; \ - msgmerge -o $@ $@~ $(POT); \ - else \ - echo "Initializing $@ from $(POT)."; \ - msginit -i $(POT) -o $@~ --locale=$*; \ - sed -e 's/charset=.*\\n/charset=UTF-8\\n/' $@~ > $@; \ - fi +# Convenient makefile for managing translations. + +# Prerequisites: +# - GNU gettext +# - Python + +# Adding new translations +# ----------------------- + +# To start working on a new translation edit the POS=... line +# below. If you want to add for example a french translation, add +# fr.po. + +# Then run "make po" to generate a fresh .po file from translatable +# strings in t_options.h. Now you can edit the new .po file (fr.po in +# the example above) to translate the strings. Please make sure that +# your editor encodes the file in UTF-8. + +# Updating existing translations +# ------------------------------ + +# Run "make po" to update .po files with new translatable strings from +# t_options.h. Now you can edit the .po files you're interested +# in. Please make sure that your editor encodes the file in UTF-8. + +# Updating options.h +# ------------------ + +# Finally run "make" to generate options.h from t_options.h with all +# translations. Now you can rebuild the drivers. Any common options +# used by the drivers will have option descriptions with the latest +# translations. + +# Publishing translations +# ----------------------- + +# To get your translation(s) into Mesa CVS, please send me your +# .po file. + +# More information: +# - info gettext + +# The set of supported languages. Add languages as needed. +POS=de.po es.po nl.po fr.po sv.po + +# +# Don't change anything below, unless you know what you're doing. +# +LANGS=$(POS:%.po=%) +MOS=$(POS:%.po=%/LC_MESSAGES/options.mo) +POT=xmlpool.pot + +.PHONY: all clean pot po mo + +all: options.h + +# Only intermediate files are cleaned up. options.h is not deleted because +# it's in CVS. +clean: + -rm -f $(POT) *~ + -rm -rf $(LANGS) + +# Default target options.h +options.h: t_options.h mo + $(PYTHON2) $(PYTHON_FLAGS) gen_xmlpool.py $(LANGS) > options.h + +# Update .mo files from the corresponding .po files. +mo: + @for mo in $(MOS); do \ + lang=$${mo%%/*}; \ + echo "Updating $$mo from $$lang.po."; \ + mkdir -p $${mo%/*}; \ + msgfmt -o $$mo $$lang.po; \ + done + +# Use this target to create or update .po files with new messages in +# driconf.py. +po: $(POS) + +pot: $(POT) + +# Extract message catalog from driconf.py. +$(POT): t_options.h + xgettext -L C --from-code utf-8 -o $(POT) t_options.h + +# Create or update a .po file for a specific language. +%.po: $(POT) + @if [ -f $@ ]; then \ + echo "Merging new strings from $(POT) into $@."; \ + mv $@ $@~; \ + msgmerge -o $@ $@~ $(POT); \ + else \ + echo "Initializing $@ from $(POT)."; \ + msginit -i $(POT) -o $@~ --locale=$*; \ + sed -e 's/charset=.*\\n/charset=UTF-8\\n/' $@~ > $@; \ + fi diff --git a/mesalib/src/mesa/drivers/dri/common/xmlpool/gen_xmlpool.py b/mesalib/src/mesa/drivers/dri/common/xmlpool/gen_xmlpool.py index f5ac5e25b..7398c4cd0 100644 --- a/mesalib/src/mesa/drivers/dri/common/xmlpool/gen_xmlpool.py +++ b/mesalib/src/mesa/drivers/dri/common/xmlpool/gen_xmlpool.py @@ -1,191 +1,191 @@ -#!/usr/bin/python - -import sys -import gettext -import re - -# List of supported languages -languages = sys.argv[1:] - -# Escape special characters in C strings -def escapeCString (s): - escapeSeqs = {'\a' : '\\a', '\b' : '\\b', '\f' : '\\f', '\n' : '\\n', - '\r' : '\\r', '\t' : '\\t', '\v' : '\\v', '\\' : '\\\\'} - # " -> '' is a hack. Quotes (") aren't possible in XML attributes. - # Better use Unicode characters for typographic quotes in option - # descriptions and translations. - i = 0 - r = '' - while i < len(s): - # Special case: escape double quote with \u201c or \u201d, depending - # on whether it's an open or close quote. This is needed because plain - # double quotes are not possible in XML attributes. - if s[i] == '"': - if i == len(s)-1 or s[i+1].isspace(): - # close quote - q = u'\u201c' - else: - # open quote - q = u'\u201d' - r = r + q - elif escapeSeqs.has_key(s[i]): - r = r + escapeSeqs[s[i]] - else: - r = r + s[i] - i = i + 1 - return r - -# Expand escape sequences in C strings (needed for gettext lookup) -def expandCString (s): - escapeSeqs = {'a' : '\a', 'b' : '\b', 'f' : '\f', 'n' : '\n', - 'r' : '\r', 't' : '\t', 'v' : '\v', - '"' : '"', '\\' : '\\'} - i = 0 - escape = False - hexa = False - octa = False - num = 0 - digits = 0 - r = '' - while i < len(s): - if not escape: - if s[i] == '\\': - escape = True - else: - r = r + s[i] - elif hexa: - if (s[i] >= '0' and s[i] <= '9') or \ - (s[i] >= 'a' and s[i] <= 'f') or \ - (s[i] >= 'A' and s[i] <= 'F'): - num = num * 16 + int(s[i],16) - digits = digits + 1 - else: - digits = 2 - if digits >= 2: - hexa = False - escape = False - r = r + chr(num) - elif octa: - if s[i] >= '0' and s[i] <= '7': - num = num * 8 + int(s[i],8) - digits = digits + 1 - else: - digits = 3 - if digits >= 3: - octa = False - escape = False - r = r + chr(num) - else: - if escapeSeqs.has_key(s[i]): - r = r + escapeSeqs[s[i]] - escape = False - elif s[i] >= '0' and s[i] <= '7': - octa = True - num = int(s[i],8) - if num <= 3: - digits = 1 - else: - digits = 2 - elif s[i] == 'x' or s[i] == 'X': - hexa = True - num = 0 - digits = 0 - else: - r = r + s[i] - escape = False - i = i + 1 - return r - -# Expand matches. The first match is always a DESC or DESC_BEGIN match. -# Subsequent matches are ENUM matches. -# -# DESC, DESC_BEGIN format: \1 \2= \3 \4=gettext(" \5= \6=") \7 -# ENUM format: \1 \2=gettext(" \3= \4=") \5 -def expandMatches (matches, translations, end=None): - assert len(matches) > 0 - nTranslations = len(translations) - i = 0 - # Expand the description+enums for all translations - for lang,trans in translations: - i = i + 1 - # Make sure that all but the last line of a simple description - # are extended with a backslash. - suffix = '' - if len(matches) == 1 and i < len(translations) and \ - not matches[0].expand (r'\7').endswith('\\'): - suffix = ' \\' - # Expand the description line. Need to use ugettext in order to allow - # non-ascii unicode chars in the original English descriptions. - text = escapeCString (trans.ugettext (unicode (expandCString ( - matches[0].expand (r'\5')), "utf-8"))).encode("utf-8") - print matches[0].expand (r'\1' + lang + r'\3"' + text + r'"\7') + suffix - # Expand any subsequent enum lines - for match in matches[1:]: - text = escapeCString (trans.ugettext (unicode (expandCString ( - match.expand (r'\3')), "utf-8"))).encode("utf-8") - print match.expand (r'\1"' + text + r'"\5') - - # Expand description end - if end: - print end, - -# Compile a list of translation classes to all supported languages. -# The first translation is always a NullTranslations. -translations = [("en", gettext.NullTranslations())] -for lang in languages: - try: - trans = gettext.translation ("options", ".", [lang]) - except IOError: - sys.stderr.write ("Warning: language '%s' not found.\n" % lang) - continue - translations.append ((lang, trans)) - -# Regular expressions: -reLibintl_h = re.compile (r'#\s*include\s*') -reDESC = re.compile (r'(\s*DRI_CONF_DESC\s*\(\s*)([a-z]+)(\s*,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$') -reDESC_BEGIN = re.compile (r'(\s*DRI_CONF_DESC_BEGIN\s*\(\s*)([a-z]+)(\s*,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$') -reENUM = re.compile (r'(\s*DRI_CONF_ENUM\s*\([^,]+,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$') -reDESC_END = re.compile (r'\s*DRI_CONF_DESC_END') - -# Print a header -print \ -"/***********************************************************************\n" \ -" *** THIS FILE IS GENERATED AUTOMATICALLY. DON'T EDIT! ***\n" \ -" ***********************************************************************/" - -# Process the options template and generate options.h with all -# translations. -template = file ("t_options.h", "r") -descMatches = [] -for line in template: - if len(descMatches) > 0: - matchENUM = reENUM .match (line) - matchDESC_END = reDESC_END.match (line) - if matchENUM: - descMatches.append (matchENUM) - elif matchDESC_END: - expandMatches (descMatches, translations, line) - descMatches = [] - else: - sys.stderr.write ( - "Warning: unexpected line inside description dropped:\n%s\n" \ - % line) - continue - if reLibintl_h.search (line): - # Ignore (comment out) #include - print "/* %s * commented out by gen_xmlpool.py */" % line - continue - matchDESC = reDESC .match (line) - matchDESC_BEGIN = reDESC_BEGIN.match (line) - if matchDESC: - assert len(descMatches) == 0 - expandMatches ([matchDESC], translations) - elif matchDESC_BEGIN: - assert len(descMatches) == 0 - descMatches = [matchDESC_BEGIN] - else: - print line, - -if len(descMatches) > 0: - sys.stderr.write ("Warning: unterminated description at end of file.\n") - expandMatches (descMatches, translations) +#!/usr/bin/python + +import sys +import gettext +import re + +# List of supported languages +languages = sys.argv[1:] + +# Escape special characters in C strings +def escapeCString (s): + escapeSeqs = {'\a' : '\\a', '\b' : '\\b', '\f' : '\\f', '\n' : '\\n', + '\r' : '\\r', '\t' : '\\t', '\v' : '\\v', '\\' : '\\\\'} + # " -> '' is a hack. Quotes (") aren't possible in XML attributes. + # Better use Unicode characters for typographic quotes in option + # descriptions and translations. + i = 0 + r = '' + while i < len(s): + # Special case: escape double quote with \u201c or \u201d, depending + # on whether it's an open or close quote. This is needed because plain + # double quotes are not possible in XML attributes. + if s[i] == '"': + if i == len(s)-1 or s[i+1].isspace(): + # close quote + q = u'\u201c' + else: + # open quote + q = u'\u201d' + r = r + q + elif escapeSeqs.has_key(s[i]): + r = r + escapeSeqs[s[i]] + else: + r = r + s[i] + i = i + 1 + return r + +# Expand escape sequences in C strings (needed for gettext lookup) +def expandCString (s): + escapeSeqs = {'a' : '\a', 'b' : '\b', 'f' : '\f', 'n' : '\n', + 'r' : '\r', 't' : '\t', 'v' : '\v', + '"' : '"', '\\' : '\\'} + i = 0 + escape = False + hexa = False + octa = False + num = 0 + digits = 0 + r = '' + while i < len(s): + if not escape: + if s[i] == '\\': + escape = True + else: + r = r + s[i] + elif hexa: + if (s[i] >= '0' and s[i] <= '9') or \ + (s[i] >= 'a' and s[i] <= 'f') or \ + (s[i] >= 'A' and s[i] <= 'F'): + num = num * 16 + int(s[i],16) + digits = digits + 1 + else: + digits = 2 + if digits >= 2: + hexa = False + escape = False + r = r + chr(num) + elif octa: + if s[i] >= '0' and s[i] <= '7': + num = num * 8 + int(s[i],8) + digits = digits + 1 + else: + digits = 3 + if digits >= 3: + octa = False + escape = False + r = r + chr(num) + else: + if escapeSeqs.has_key(s[i]): + r = r + escapeSeqs[s[i]] + escape = False + elif s[i] >= '0' and s[i] <= '7': + octa = True + num = int(s[i],8) + if num <= 3: + digits = 1 + else: + digits = 2 + elif s[i] == 'x' or s[i] == 'X': + hexa = True + num = 0 + digits = 0 + else: + r = r + s[i] + escape = False + i = i + 1 + return r + +# Expand matches. The first match is always a DESC or DESC_BEGIN match. +# Subsequent matches are ENUM matches. +# +# DESC, DESC_BEGIN format: \1 \2= \3 \4=gettext(" \5= \6=") \7 +# ENUM format: \1 \2=gettext(" \3= \4=") \5 +def expandMatches (matches, translations, end=None): + assert len(matches) > 0 + nTranslations = len(translations) + i = 0 + # Expand the description+enums for all translations + for lang,trans in translations: + i = i + 1 + # Make sure that all but the last line of a simple description + # are extended with a backslash. + suffix = '' + if len(matches) == 1 and i < len(translations) and \ + not matches[0].expand (r'\7').endswith('\\'): + suffix = ' \\' + # Expand the description line. Need to use ugettext in order to allow + # non-ascii unicode chars in the original English descriptions. + text = escapeCString (trans.ugettext (unicode (expandCString ( + matches[0].expand (r'\5')), "utf-8"))).encode("utf-8") + print matches[0].expand (r'\1' + lang + r'\3"' + text + r'"\7') + suffix + # Expand any subsequent enum lines + for match in matches[1:]: + text = escapeCString (trans.ugettext (unicode (expandCString ( + match.expand (r'\3')), "utf-8"))).encode("utf-8") + print match.expand (r'\1"' + text + r'"\5') + + # Expand description end + if end: + print end, + +# Compile a list of translation classes to all supported languages. +# The first translation is always a NullTranslations. +translations = [("en", gettext.NullTranslations())] +for lang in languages: + try: + trans = gettext.translation ("options", ".", [lang]) + except IOError: + sys.stderr.write ("Warning: language '%s' not found.\n" % lang) + continue + translations.append ((lang, trans)) + +# Regular expressions: +reLibintl_h = re.compile (r'#\s*include\s*') +reDESC = re.compile (r'(\s*DRI_CONF_DESC\s*\(\s*)([a-z]+)(\s*,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$') +reDESC_BEGIN = re.compile (r'(\s*DRI_CONF_DESC_BEGIN\s*\(\s*)([a-z]+)(\s*,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$') +reENUM = re.compile (r'(\s*DRI_CONF_ENUM\s*\([^,]+,\s*)(gettext\s*\(\s*")(.*)("\s*\))(\s*\)[ \t]*\\?)$') +reDESC_END = re.compile (r'\s*DRI_CONF_DESC_END') + +# Print a header +print \ +"/***********************************************************************\n" \ +" *** THIS FILE IS GENERATED AUTOMATICALLY. DON'T EDIT! ***\n" \ +" ***********************************************************************/" + +# Process the options template and generate options.h with all +# translations. +template = file ("t_options.h", "r") +descMatches = [] +for line in template: + if len(descMatches) > 0: + matchENUM = reENUM .match (line) + matchDESC_END = reDESC_END.match (line) + if matchENUM: + descMatches.append (matchENUM) + elif matchDESC_END: + expandMatches (descMatches, translations, line) + descMatches = [] + else: + sys.stderr.write ( + "Warning: unexpected line inside description dropped:\n%s\n" \ + % line) + continue + if reLibintl_h.search (line): + # Ignore (comment out) #include + print "/* %s * commented out by gen_xmlpool.py */" % line + continue + matchDESC = reDESC .match (line) + matchDESC_BEGIN = reDESC_BEGIN.match (line) + if matchDESC: + assert len(descMatches) == 0 + expandMatches ([matchDESC], translations) + elif matchDESC_BEGIN: + assert len(descMatches) == 0 + descMatches = [matchDESC_BEGIN] + else: + print line, + +if len(descMatches) > 0: + sys.stderr.write ("Warning: unterminated description at end of file.\n") + expandMatches (descMatches, translations) diff --git a/mesalib/src/mesa/drivers/dri/swrast/.gitignore b/mesalib/src/mesa/drivers/dri/swrast/.gitignore new file mode 100644 index 000000000..f3c7a7c5d --- /dev/null +++ b/mesalib/src/mesa/drivers/dri/swrast/.gitignore @@ -0,0 +1 @@ +Makefile diff --git a/mesalib/src/mesa/gl.pc.in b/mesalib/src/mesa/gl.pc.in index 616819063..2d3bc917e 100644 --- a/mesalib/src/mesa/gl.pc.in +++ b/mesalib/src/mesa/gl.pc.in @@ -1,13 +1,13 @@ -prefix=@INSTALL_DIR@ -exec_prefix=${prefix} -libdir=@INSTALL_LIB_DIR@ -includedir=@INSTALL_INC_DIR@ - -Name: gl -Description: Mesa OpenGL library -Requires.private: @GL_PC_REQ_PRIV@ -Version: @VERSION@ -Libs: -L${libdir} -l@GL_LIB@ -Libs.private: @GL_PC_LIB_PRIV@ -Cflags: -I${includedir} @GL_PC_CFLAGS@ -glx_tls: @GLX_TLS@ +prefix=@INSTALL_DIR@ +exec_prefix=${prefix} +libdir=@INSTALL_LIB_DIR@ +includedir=@INSTALL_INC_DIR@ + +Name: gl +Description: Mesa OpenGL library +Requires.private: @GL_PC_REQ_PRIV@ +Version: @VERSION@ +Libs: -L${libdir} -l@GL_LIB@ +Libs.private: @GL_PC_LIB_PRIV@ +Cflags: -I${includedir} @GL_PC_CFLAGS@ +glx_tls: @GLX_TLS@ diff --git a/mesalib/src/mesa/main/.gitignore b/mesalib/src/mesa/main/.gitignore new file mode 100644 index 000000000..caed74f40 --- /dev/null +++ b/mesalib/src/mesa/main/.gitignore @@ -0,0 +1,13 @@ +api_exec_es1.c +api_exec_es2.c +dispatch.h +enums.c +get_es1.c +get_es2.c +git_sha1.h +git_sha1.h.tmp +api_exec_es1_dispatch.h +api_exec_es1_remap_helper.h +api_exec_es2_dispatch.h +api_exec_es2_remap_helper.h +remap_helper.h diff --git a/mesalib/src/mesa/main/api_loopback.c b/mesalib/src/mesa/main/api_loopback.c index fb9d781f7..c438307d8 100644 --- a/mesalib/src/mesa/main/api_loopback.c +++ b/mesalib/src/mesa/main/api_loopback.c @@ -1,1725 +1,1725 @@ -/** - * \file api_loopback.c - * - * \author Keith Whitwell - */ - -/* - * Mesa 3-D graphics library - * Version: 6.3 - * - * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#include "glheader.h" -#include "macros.h" -#include "api_loopback.h" -#include "mtypes.h" -#include "glapi/glapi.h" -#include "glapi/glthread.h" -#include "main/dispatch.h" -#include "mfeatures.h" - -/* KW: A set of functions to convert unusual Color/Normal/Vertex/etc - * calls to a smaller set of driver-provided formats. Currently just - * go back to dispatch to find these (eg. call glNormal3f directly), - * hence 'loopback'. - * - * The driver must supply all of the remaining entry points, which are - * listed in dd.h. The easiest way for a driver to do this is to - * install the supplied software t&l module. - */ -#define COLORF(r,g,b,a) CALL_Color4f(GET_DISPATCH(), (r,g,b,a)) -#define VERTEX2(x,y) CALL_Vertex2f(GET_DISPATCH(), (x,y)) -#define VERTEX3(x,y,z) CALL_Vertex3f(GET_DISPATCH(), (x,y,z)) -#define VERTEX4(x,y,z,w) CALL_Vertex4f(GET_DISPATCH(), (x,y,z,w)) -#define NORMAL(x,y,z) CALL_Normal3f(GET_DISPATCH(), (x,y,z)) -#define TEXCOORD1(s) CALL_TexCoord1f(GET_DISPATCH(), (s)) -#define TEXCOORD2(s,t) CALL_TexCoord2f(GET_DISPATCH(), (s,t)) -#define TEXCOORD3(s,t,u) CALL_TexCoord3f(GET_DISPATCH(), (s,t,u)) -#define TEXCOORD4(s,t,u,v) CALL_TexCoord4f(GET_DISPATCH(), (s,t,u,v)) -#define INDEX(c) CALL_Indexf(GET_DISPATCH(), (c)) -#define MULTI_TEXCOORD1(z,s) CALL_MultiTexCoord1fARB(GET_DISPATCH(), (z,s)) -#define MULTI_TEXCOORD2(z,s,t) CALL_MultiTexCoord2fARB(GET_DISPATCH(), (z,s,t)) -#define MULTI_TEXCOORD3(z,s,t,u) CALL_MultiTexCoord3fARB(GET_DISPATCH(), (z,s,t,u)) -#define MULTI_TEXCOORD4(z,s,t,u,v) CALL_MultiTexCoord4fARB(GET_DISPATCH(), (z,s,t,u,v)) -#define EVALCOORD1(x) CALL_EvalCoord1f(GET_DISPATCH(), (x)) -#define EVALCOORD2(x,y) CALL_EvalCoord2f(GET_DISPATCH(), (x,y)) -#define MATERIALFV(a,b,c) CALL_Materialfv(GET_DISPATCH(), (a,b,c)) -#define RECTF(a,b,c,d) CALL_Rectf(GET_DISPATCH(), (a,b,c,d)) - -#define FOGCOORDF(x) CALL_FogCoordfEXT(GET_DISPATCH(), (x)) -#define SECONDARYCOLORF(a,b,c) CALL_SecondaryColor3fEXT(GET_DISPATCH(), (a,b,c)) - -#define ATTRIB1NV(index,x) CALL_VertexAttrib1fNV(GET_DISPATCH(), (index,x)) -#define ATTRIB2NV(index,x,y) CALL_VertexAttrib2fNV(GET_DISPATCH(), (index,x,y)) -#define ATTRIB3NV(index,x,y,z) CALL_VertexAttrib3fNV(GET_DISPATCH(), (index,x,y,z)) -#define ATTRIB4NV(index,x,y,z,w) CALL_VertexAttrib4fNV(GET_DISPATCH(), (index,x,y,z,w)) - -#define ATTRIB1ARB(index,x) CALL_VertexAttrib1fARB(GET_DISPATCH(), (index,x)) -#define ATTRIB2ARB(index,x,y) CALL_VertexAttrib2fARB(GET_DISPATCH(), (index,x,y)) -#define ATTRIB3ARB(index,x,y,z) CALL_VertexAttrib3fARB(GET_DISPATCH(), (index,x,y,z)) -#define ATTRIB4ARB(index,x,y,z,w) CALL_VertexAttrib4fARB(GET_DISPATCH(), (index,x,y,z,w)) - -#define ATTRIBI_1I(index,x) CALL_VertexAttribI1iEXT(GET_DISPATCH(), (index,x)) -#define ATTRIBI_1UI(index,x) CALL_VertexAttribI1uiEXT(GET_DISPATCH(), (index,x)) -#define ATTRIBI_4I(index,x,y,z,w) CALL_VertexAttribI4iEXT(GET_DISPATCH(), (index,x,y,z,w)) - -#define ATTRIBI_4UI(index,x,y,z,w) CALL_VertexAttribI4uiEXT(GET_DISPATCH(), (index,x,y,z,w)) - - -#if FEATURE_beginend - - -static void GLAPIENTRY -loopback_Color3b_f( GLbyte red, GLbyte green, GLbyte blue ) -{ - COLORF( BYTE_TO_FLOAT(red), - BYTE_TO_FLOAT(green), - BYTE_TO_FLOAT(blue), - 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3d_f( GLdouble red, GLdouble green, GLdouble blue ) -{ - COLORF( (GLfloat) red, (GLfloat) green, (GLfloat) blue, 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3i_f( GLint red, GLint green, GLint blue ) -{ - COLORF( INT_TO_FLOAT(red), INT_TO_FLOAT(green), - INT_TO_FLOAT(blue), 1.0); -} - -static void GLAPIENTRY -loopback_Color3s_f( GLshort red, GLshort green, GLshort blue ) -{ - COLORF( SHORT_TO_FLOAT(red), SHORT_TO_FLOAT(green), - SHORT_TO_FLOAT(blue), 1.0); -} - -static void GLAPIENTRY -loopback_Color3ui_f( GLuint red, GLuint green, GLuint blue ) -{ - COLORF( UINT_TO_FLOAT(red), UINT_TO_FLOAT(green), - UINT_TO_FLOAT(blue), 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3us_f( GLushort red, GLushort green, GLushort blue ) -{ - COLORF( USHORT_TO_FLOAT(red), USHORT_TO_FLOAT(green), - USHORT_TO_FLOAT(blue), 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3ub_f( GLubyte red, GLubyte green, GLubyte blue ) -{ - COLORF( UBYTE_TO_FLOAT(red), UBYTE_TO_FLOAT(green), - UBYTE_TO_FLOAT(blue), 1.0 ); -} - - -static void GLAPIENTRY -loopback_Color3bv_f( const GLbyte *v ) -{ - COLORF( BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), - BYTE_TO_FLOAT(v[2]), 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3dv_f( const GLdouble *v ) -{ - COLORF( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3iv_f( const GLint *v ) -{ - COLORF( INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), - INT_TO_FLOAT(v[2]), 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3sv_f( const GLshort *v ) -{ - COLORF( SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), - SHORT_TO_FLOAT(v[2]), 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3uiv_f( const GLuint *v ) -{ - COLORF( UINT_TO_FLOAT(v[0]), UINT_TO_FLOAT(v[1]), - UINT_TO_FLOAT(v[2]), 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3usv_f( const GLushort *v ) -{ - COLORF( USHORT_TO_FLOAT(v[0]), USHORT_TO_FLOAT(v[1]), - USHORT_TO_FLOAT(v[2]), 1.0 ); -} - -static void GLAPIENTRY -loopback_Color3ubv_f( const GLubyte *v ) -{ - COLORF( UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), - UBYTE_TO_FLOAT(v[2]), 1.0 ); -} - - -static void GLAPIENTRY -loopback_Color4b_f( GLbyte red, GLbyte green, GLbyte blue, - GLbyte alpha ) -{ - COLORF( BYTE_TO_FLOAT(red), BYTE_TO_FLOAT(green), - BYTE_TO_FLOAT(blue), BYTE_TO_FLOAT(alpha) ); -} - -static void GLAPIENTRY -loopback_Color4d_f( GLdouble red, GLdouble green, GLdouble blue, - GLdouble alpha ) -{ - COLORF( (GLfloat) red, (GLfloat) green, (GLfloat) blue, (GLfloat) alpha ); -} - -static void GLAPIENTRY -loopback_Color4i_f( GLint red, GLint green, GLint blue, GLint alpha ) -{ - COLORF( INT_TO_FLOAT(red), INT_TO_FLOAT(green), - INT_TO_FLOAT(blue), INT_TO_FLOAT(alpha) ); -} - -static void GLAPIENTRY -loopback_Color4s_f( GLshort red, GLshort green, GLshort blue, - GLshort alpha ) -{ - COLORF( SHORT_TO_FLOAT(red), SHORT_TO_FLOAT(green), - SHORT_TO_FLOAT(blue), SHORT_TO_FLOAT(alpha) ); -} - -static void GLAPIENTRY -loopback_Color4ui_f( GLuint red, GLuint green, GLuint blue, GLuint alpha ) -{ - COLORF( UINT_TO_FLOAT(red), UINT_TO_FLOAT(green), - UINT_TO_FLOAT(blue), UINT_TO_FLOAT(alpha) ); -} - -static void GLAPIENTRY -loopback_Color4us_f( GLushort red, GLushort green, GLushort blue, GLushort alpha ) -{ - COLORF( USHORT_TO_FLOAT(red), USHORT_TO_FLOAT(green), - USHORT_TO_FLOAT(blue), USHORT_TO_FLOAT(alpha) ); -} - -static void GLAPIENTRY -loopback_Color4ub_f( GLubyte red, GLubyte green, GLubyte blue, GLubyte alpha ) -{ - COLORF( UBYTE_TO_FLOAT(red), UBYTE_TO_FLOAT(green), - UBYTE_TO_FLOAT(blue), UBYTE_TO_FLOAT(alpha) ); -} - - -static void GLAPIENTRY -loopback_Color4iv_f( const GLint *v ) -{ - COLORF( INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), - INT_TO_FLOAT(v[2]), INT_TO_FLOAT(v[3]) ); -} - - -static void GLAPIENTRY -loopback_Color4bv_f( const GLbyte *v ) -{ - COLORF( BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), - BYTE_TO_FLOAT(v[2]), BYTE_TO_FLOAT(v[3]) ); -} - -static void GLAPIENTRY -loopback_Color4dv_f( const GLdouble *v ) -{ - COLORF( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3] ); -} - - -static void GLAPIENTRY -loopback_Color4sv_f( const GLshort *v) -{ - COLORF( SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), - SHORT_TO_FLOAT(v[2]), SHORT_TO_FLOAT(v[3]) ); -} - - -static void GLAPIENTRY -loopback_Color4uiv_f( const GLuint *v) -{ - COLORF( UINT_TO_FLOAT(v[0]), UINT_TO_FLOAT(v[1]), - UINT_TO_FLOAT(v[2]), UINT_TO_FLOAT(v[3]) ); -} - -static void GLAPIENTRY -loopback_Color4usv_f( const GLushort *v) -{ - COLORF( USHORT_TO_FLOAT(v[0]), USHORT_TO_FLOAT(v[1]), - USHORT_TO_FLOAT(v[2]), USHORT_TO_FLOAT(v[3]) ); -} - -static void GLAPIENTRY -loopback_Color4ubv_f( const GLubyte *v) -{ - COLORF( UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), - UBYTE_TO_FLOAT(v[2]), UBYTE_TO_FLOAT(v[3]) ); -} - - -static void GLAPIENTRY -loopback_FogCoorddEXT( GLdouble d ) -{ - FOGCOORDF( (GLfloat) d ); -} - -static void GLAPIENTRY -loopback_FogCoorddvEXT( const GLdouble *v ) -{ - FOGCOORDF( (GLfloat) *v ); -} - - -static void GLAPIENTRY -loopback_Indexd( GLdouble c ) -{ - INDEX( (GLfloat) c ); -} - -static void GLAPIENTRY -loopback_Indexi( GLint c ) -{ - INDEX( (GLfloat) c ); -} - -static void GLAPIENTRY -loopback_Indexs( GLshort c ) -{ - INDEX( (GLfloat) c ); -} - -static void GLAPIENTRY -loopback_Indexub( GLubyte c ) -{ - INDEX( (GLfloat) c ); -} - -static void GLAPIENTRY -loopback_Indexdv( const GLdouble *c ) -{ - INDEX( (GLfloat) *c ); -} - -static void GLAPIENTRY -loopback_Indexiv( const GLint *c ) -{ - INDEX( (GLfloat) *c ); -} - -static void GLAPIENTRY -loopback_Indexsv( const GLshort *c ) -{ - INDEX( (GLfloat) *c ); -} - -static void GLAPIENTRY -loopback_Indexubv( const GLubyte *c ) -{ - INDEX( (GLfloat) *c ); -} - - -static void GLAPIENTRY -loopback_EdgeFlagv(const GLboolean *flag) -{ - CALL_EdgeFlag(GET_DISPATCH(), (*flag)); -} - - -static void GLAPIENTRY -loopback_Normal3b( GLbyte nx, GLbyte ny, GLbyte nz ) -{ - NORMAL( BYTE_TO_FLOAT(nx), BYTE_TO_FLOAT(ny), BYTE_TO_FLOAT(nz) ); -} - -static void GLAPIENTRY -loopback_Normal3d( GLdouble nx, GLdouble ny, GLdouble nz ) -{ - NORMAL((GLfloat) nx, (GLfloat) ny, (GLfloat) nz); -} - -static void GLAPIENTRY -loopback_Normal3i( GLint nx, GLint ny, GLint nz ) -{ - NORMAL( INT_TO_FLOAT(nx), INT_TO_FLOAT(ny), INT_TO_FLOAT(nz) ); -} - -static void GLAPIENTRY -loopback_Normal3s( GLshort nx, GLshort ny, GLshort nz ) -{ - NORMAL( SHORT_TO_FLOAT(nx), SHORT_TO_FLOAT(ny), SHORT_TO_FLOAT(nz) ); -} - -static void GLAPIENTRY -loopback_Normal3bv( const GLbyte *v ) -{ - NORMAL( BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), BYTE_TO_FLOAT(v[2]) ); -} - -static void GLAPIENTRY -loopback_Normal3dv( const GLdouble *v ) -{ - NORMAL( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} - -static void GLAPIENTRY -loopback_Normal3iv( const GLint *v ) -{ - NORMAL( INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), INT_TO_FLOAT(v[2]) ); -} - -static void GLAPIENTRY -loopback_Normal3sv( const GLshort *v ) -{ - NORMAL( SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), SHORT_TO_FLOAT(v[2]) ); -} - -static void GLAPIENTRY -loopback_TexCoord1d( GLdouble s ) -{ - TEXCOORD1((GLfloat) s); -} - -static void GLAPIENTRY -loopback_TexCoord1i( GLint s ) -{ - TEXCOORD1((GLfloat) s); -} - -static void GLAPIENTRY -loopback_TexCoord1s( GLshort s ) -{ - TEXCOORD1((GLfloat) s); -} - -static void GLAPIENTRY -loopback_TexCoord2d( GLdouble s, GLdouble t ) -{ - TEXCOORD2((GLfloat) s,(GLfloat) t); -} - -static void GLAPIENTRY -loopback_TexCoord2s( GLshort s, GLshort t ) -{ - TEXCOORD2((GLfloat) s,(GLfloat) t); -} - -static void GLAPIENTRY -loopback_TexCoord2i( GLint s, GLint t ) -{ - TEXCOORD2((GLfloat) s,(GLfloat) t); -} - -static void GLAPIENTRY -loopback_TexCoord3d( GLdouble s, GLdouble t, GLdouble r ) -{ - TEXCOORD3((GLfloat) s,(GLfloat) t,(GLfloat) r); -} - -static void GLAPIENTRY -loopback_TexCoord3i( GLint s, GLint t, GLint r ) -{ - TEXCOORD3((GLfloat) s,(GLfloat) t,(GLfloat) r); -} - -static void GLAPIENTRY -loopback_TexCoord3s( GLshort s, GLshort t, GLshort r ) -{ - TEXCOORD3((GLfloat) s,(GLfloat) t,(GLfloat) r); -} - -static void GLAPIENTRY -loopback_TexCoord4d( GLdouble s, GLdouble t, GLdouble r, GLdouble q ) -{ - TEXCOORD4((GLfloat) s,(GLfloat) t,(GLfloat) r,(GLfloat) q); -} - -static void GLAPIENTRY -loopback_TexCoord4i( GLint s, GLint t, GLint r, GLint q ) -{ - TEXCOORD4((GLfloat) s,(GLfloat) t,(GLfloat) r,(GLfloat) q); -} - -static void GLAPIENTRY -loopback_TexCoord4s( GLshort s, GLshort t, GLshort r, GLshort q ) -{ - TEXCOORD4((GLfloat) s,(GLfloat) t,(GLfloat) r,(GLfloat) q); -} - -static void GLAPIENTRY -loopback_TexCoord1dv( const GLdouble *v ) -{ - TEXCOORD1((GLfloat) v[0]); -} - -static void GLAPIENTRY -loopback_TexCoord1iv( const GLint *v ) -{ - TEXCOORD1((GLfloat) v[0]); -} - -static void GLAPIENTRY -loopback_TexCoord1sv( const GLshort *v ) -{ - TEXCOORD1((GLfloat) v[0]); -} - -static void GLAPIENTRY -loopback_TexCoord2dv( const GLdouble *v ) -{ - TEXCOORD2((GLfloat) v[0],(GLfloat) v[1]); -} - -static void GLAPIENTRY -loopback_TexCoord2iv( const GLint *v ) -{ - TEXCOORD2((GLfloat) v[0],(GLfloat) v[1]); -} - -static void GLAPIENTRY -loopback_TexCoord2sv( const GLshort *v ) -{ - TEXCOORD2((GLfloat) v[0],(GLfloat) v[1]); -} - -static void GLAPIENTRY -loopback_TexCoord3dv( const GLdouble *v ) -{ - TEXCOORD3((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2]); -} - -static void GLAPIENTRY -loopback_TexCoord3iv( const GLint *v ) -{ - TEXCOORD3((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2]); -} - -static void GLAPIENTRY -loopback_TexCoord3sv( const GLshort *v ) -{ - TEXCOORD3((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2]); -} - -static void GLAPIENTRY -loopback_TexCoord4dv( const GLdouble *v ) -{ - TEXCOORD4((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2],(GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_TexCoord4iv( const GLint *v ) -{ - TEXCOORD4((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2],(GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_TexCoord4sv( const GLshort *v ) -{ - TEXCOORD4((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2],(GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_Vertex2d( GLdouble x, GLdouble y ) -{ - VERTEX2( (GLfloat) x, (GLfloat) y ); -} - -static void GLAPIENTRY -loopback_Vertex2i( GLint x, GLint y ) -{ - VERTEX2( (GLfloat) x, (GLfloat) y ); -} - -static void GLAPIENTRY -loopback_Vertex2s( GLshort x, GLshort y ) -{ - VERTEX2( (GLfloat) x, (GLfloat) y ); -} - -static void GLAPIENTRY -loopback_Vertex3d( GLdouble x, GLdouble y, GLdouble z ) -{ - VERTEX3( (GLfloat) x, (GLfloat) y, (GLfloat) z ); -} - -static void GLAPIENTRY -loopback_Vertex3i( GLint x, GLint y, GLint z ) -{ - VERTEX3( (GLfloat) x, (GLfloat) y, (GLfloat) z ); -} - -static void GLAPIENTRY -loopback_Vertex3s( GLshort x, GLshort y, GLshort z ) -{ - VERTEX3( (GLfloat) x, (GLfloat) y, (GLfloat) z ); -} - -static void GLAPIENTRY -loopback_Vertex4d( GLdouble x, GLdouble y, GLdouble z, GLdouble w ) -{ - VERTEX4( (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w ); -} - -static void GLAPIENTRY -loopback_Vertex4i( GLint x, GLint y, GLint z, GLint w ) -{ - VERTEX4( (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w ); -} - -static void GLAPIENTRY -loopback_Vertex4s( GLshort x, GLshort y, GLshort z, GLshort w ) -{ - VERTEX4( (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w ); -} - -static void GLAPIENTRY -loopback_Vertex2dv( const GLdouble *v ) -{ - VERTEX2( (GLfloat) v[0], (GLfloat) v[1] ); -} - -static void GLAPIENTRY -loopback_Vertex2iv( const GLint *v ) -{ - VERTEX2( (GLfloat) v[0], (GLfloat) v[1] ); -} - -static void GLAPIENTRY -loopback_Vertex2sv( const GLshort *v ) -{ - VERTEX2( (GLfloat) v[0], (GLfloat) v[1] ); -} - -static void GLAPIENTRY -loopback_Vertex3dv( const GLdouble *v ) -{ - VERTEX3( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} - -static void GLAPIENTRY -loopback_Vertex3iv( const GLint *v ) -{ - VERTEX3( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} - -static void GLAPIENTRY -loopback_Vertex3sv( const GLshort *v ) -{ - VERTEX3( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} - -static void GLAPIENTRY -loopback_Vertex4dv( const GLdouble *v ) -{ - VERTEX4( (GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3] ); -} - -static void GLAPIENTRY -loopback_Vertex4iv( const GLint *v ) -{ - VERTEX4( (GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3] ); -} - -static void GLAPIENTRY -loopback_Vertex4sv( const GLshort *v ) -{ - VERTEX4( (GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord1dARB(GLenum target, GLdouble s) -{ - MULTI_TEXCOORD1( target, (GLfloat) s ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord1dvARB(GLenum target, const GLdouble *v) -{ - MULTI_TEXCOORD1( target, (GLfloat) v[0] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord1iARB(GLenum target, GLint s) -{ - MULTI_TEXCOORD1( target, (GLfloat) s ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord1ivARB(GLenum target, const GLint *v) -{ - MULTI_TEXCOORD1( target, (GLfloat) v[0] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord1sARB(GLenum target, GLshort s) -{ - MULTI_TEXCOORD1( target, (GLfloat) s ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord1svARB(GLenum target, const GLshort *v) -{ - MULTI_TEXCOORD1( target, (GLfloat) v[0] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord2dARB(GLenum target, GLdouble s, GLdouble t) -{ - MULTI_TEXCOORD2( target, (GLfloat) s, (GLfloat) t ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord2dvARB(GLenum target, const GLdouble *v) -{ - MULTI_TEXCOORD2( target, (GLfloat) v[0], (GLfloat) v[1] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord2iARB(GLenum target, GLint s, GLint t) -{ - MULTI_TEXCOORD2( target, (GLfloat) s, (GLfloat) t ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord2ivARB(GLenum target, const GLint *v) -{ - MULTI_TEXCOORD2( target, (GLfloat) v[0], (GLfloat) v[1] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord2sARB(GLenum target, GLshort s, GLshort t) -{ - MULTI_TEXCOORD2( target, (GLfloat) s, (GLfloat) t ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord2svARB(GLenum target, const GLshort *v) -{ - MULTI_TEXCOORD2( target, (GLfloat) v[0], (GLfloat) v[1] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord3dARB(GLenum target, GLdouble s, GLdouble t, GLdouble r) -{ - MULTI_TEXCOORD3( target, (GLfloat) s, (GLfloat) t, (GLfloat) r ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord3dvARB(GLenum target, const GLdouble *v) -{ - MULTI_TEXCOORD3( target, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord3iARB(GLenum target, GLint s, GLint t, GLint r) -{ - MULTI_TEXCOORD3( target, (GLfloat) s, (GLfloat) t, (GLfloat) r ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord3ivARB(GLenum target, const GLint *v) -{ - MULTI_TEXCOORD3( target, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord3sARB(GLenum target, GLshort s, GLshort t, GLshort r) -{ - MULTI_TEXCOORD3( target, (GLfloat) s, (GLfloat) t, (GLfloat) r ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord3svARB(GLenum target, const GLshort *v) -{ - MULTI_TEXCOORD3( target, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord4dARB(GLenum target, GLdouble s, GLdouble t, GLdouble r, GLdouble q) -{ - MULTI_TEXCOORD4( target, (GLfloat) s, (GLfloat) t, - (GLfloat) r, (GLfloat) q ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord4dvARB(GLenum target, const GLdouble *v) -{ - MULTI_TEXCOORD4( target, (GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord4iARB(GLenum target, GLint s, GLint t, GLint r, GLint q) -{ - MULTI_TEXCOORD4( target, (GLfloat) s, (GLfloat) t, - (GLfloat) r, (GLfloat) q ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord4ivARB(GLenum target, const GLint *v) -{ - MULTI_TEXCOORD4( target, (GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3] ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord4sARB(GLenum target, GLshort s, GLshort t, GLshort r, GLshort q) -{ - MULTI_TEXCOORD4( target, (GLfloat) s, (GLfloat) t, - (GLfloat) r, (GLfloat) q ); -} - -static void GLAPIENTRY -loopback_MultiTexCoord4svARB(GLenum target, const GLshort *v) -{ - MULTI_TEXCOORD4( target, (GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3] ); -} - -static void GLAPIENTRY -loopback_EvalCoord2dv( const GLdouble *u ) -{ - EVALCOORD2( (GLfloat) u[0], (GLfloat) u[1] ); -} - -static void GLAPIENTRY -loopback_EvalCoord2fv( const GLfloat *u ) -{ - EVALCOORD2( u[0], u[1] ); -} - -static void GLAPIENTRY -loopback_EvalCoord2d( GLdouble u, GLdouble v ) -{ - EVALCOORD2( (GLfloat) u, (GLfloat) v ); -} - -static void GLAPIENTRY -loopback_EvalCoord1dv( const GLdouble *u ) -{ - EVALCOORD1( (GLfloat) *u ); -} - -static void GLAPIENTRY -loopback_EvalCoord1fv( const GLfloat *u ) -{ - EVALCOORD1( (GLfloat) *u ); -} - -static void GLAPIENTRY -loopback_EvalCoord1d( GLdouble u ) -{ - EVALCOORD1( (GLfloat) u ); -} - -static void GLAPIENTRY -loopback_Materialf( GLenum face, GLenum pname, GLfloat param ) -{ - GLfloat fparam[4]; - fparam[0] = param; - MATERIALFV( face, pname, fparam ); -} - -static void GLAPIENTRY -loopback_Materiali(GLenum face, GLenum pname, GLint param ) -{ - GLfloat p = (GLfloat) param; - MATERIALFV(face, pname, &p); -} - -static void GLAPIENTRY -loopback_Materialiv(GLenum face, GLenum pname, const GLint *params ) -{ - GLfloat fparam[4]; - switch (pname) { - case GL_AMBIENT: - case GL_DIFFUSE: - case GL_SPECULAR: - case GL_EMISSION: - case GL_AMBIENT_AND_DIFFUSE: - fparam[0] = INT_TO_FLOAT( params[0] ); - fparam[1] = INT_TO_FLOAT( params[1] ); - fparam[2] = INT_TO_FLOAT( params[2] ); - fparam[3] = INT_TO_FLOAT( params[3] ); - break; - case GL_SHININESS: - fparam[0] = (GLfloat) params[0]; - break; - case GL_COLOR_INDEXES: - fparam[0] = (GLfloat) params[0]; - fparam[1] = (GLfloat) params[1]; - fparam[2] = (GLfloat) params[2]; - break; - default: - ; - } - MATERIALFV(face, pname, fparam); -} - - -static void GLAPIENTRY -loopback_Rectd(GLdouble x1, GLdouble y1, GLdouble x2, GLdouble y2) -{ - RECTF((GLfloat) x1, (GLfloat) y1, (GLfloat) x2, (GLfloat) y2); -} - -static void GLAPIENTRY -loopback_Rectdv(const GLdouble *v1, const GLdouble *v2) -{ - RECTF((GLfloat) v1[0], (GLfloat) v1[1], (GLfloat) v2[0], (GLfloat) v2[1]); -} - -static void GLAPIENTRY -loopback_Rectfv(const GLfloat *v1, const GLfloat *v2) -{ - RECTF(v1[0], v1[1], v2[0], v2[1]); -} - -static void GLAPIENTRY -loopback_Recti(GLint x1, GLint y1, GLint x2, GLint y2) -{ - RECTF((GLfloat) x1, (GLfloat) y1, (GLfloat) x2, (GLfloat) y2); -} - -static void GLAPIENTRY -loopback_Rectiv(const GLint *v1, const GLint *v2) -{ - RECTF((GLfloat) v1[0], (GLfloat) v1[1], (GLfloat) v2[0], (GLfloat) v2[1]); -} - -static void GLAPIENTRY -loopback_Rects(GLshort x1, GLshort y1, GLshort x2, GLshort y2) -{ - RECTF((GLfloat) x1, (GLfloat) y1, (GLfloat) x2, (GLfloat) y2); -} - -static void GLAPIENTRY -loopback_Rectsv(const GLshort *v1, const GLshort *v2) -{ - RECTF((GLfloat) v1[0], (GLfloat) v1[1], (GLfloat) v2[0], (GLfloat) v2[1]); -} - -static void GLAPIENTRY -loopback_SecondaryColor3bEXT_f( GLbyte red, GLbyte green, GLbyte blue ) -{ - SECONDARYCOLORF( BYTE_TO_FLOAT(red), - BYTE_TO_FLOAT(green), - BYTE_TO_FLOAT(blue) ); -} - -static void GLAPIENTRY -loopback_SecondaryColor3dEXT_f( GLdouble red, GLdouble green, GLdouble blue ) -{ - SECONDARYCOLORF( (GLfloat) red, (GLfloat) green, (GLfloat) blue ); -} - -static void GLAPIENTRY -loopback_SecondaryColor3iEXT_f( GLint red, GLint green, GLint blue ) -{ - SECONDARYCOLORF( INT_TO_FLOAT(red), - INT_TO_FLOAT(green), - INT_TO_FLOAT(blue)); -} - -static void GLAPIENTRY -loopback_SecondaryColor3sEXT_f( GLshort red, GLshort green, GLshort blue ) -{ - SECONDARYCOLORF(SHORT_TO_FLOAT(red), - SHORT_TO_FLOAT(green), - SHORT_TO_FLOAT(blue)); -} - -static void GLAPIENTRY -loopback_SecondaryColor3uiEXT_f( GLuint red, GLuint green, GLuint blue ) -{ - SECONDARYCOLORF(UINT_TO_FLOAT(red), - UINT_TO_FLOAT(green), - UINT_TO_FLOAT(blue)); -} - -static void GLAPIENTRY -loopback_SecondaryColor3usEXT_f( GLushort red, GLushort green, GLushort blue ) -{ - SECONDARYCOLORF(USHORT_TO_FLOAT(red), - USHORT_TO_FLOAT(green), - USHORT_TO_FLOAT(blue)); -} - -static void GLAPIENTRY -loopback_SecondaryColor3ubEXT_f( GLubyte red, GLubyte green, GLubyte blue ) -{ - SECONDARYCOLORF(UBYTE_TO_FLOAT(red), - UBYTE_TO_FLOAT(green), - UBYTE_TO_FLOAT(blue)); -} - -static void GLAPIENTRY -loopback_SecondaryColor3bvEXT_f( const GLbyte *v ) -{ - SECONDARYCOLORF(BYTE_TO_FLOAT(v[0]), - BYTE_TO_FLOAT(v[1]), - BYTE_TO_FLOAT(v[2])); -} - -static void GLAPIENTRY -loopback_SecondaryColor3dvEXT_f( const GLdouble *v ) -{ - SECONDARYCOLORF( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); -} -static void GLAPIENTRY -loopback_SecondaryColor3ivEXT_f( const GLint *v ) -{ - SECONDARYCOLORF(INT_TO_FLOAT(v[0]), - INT_TO_FLOAT(v[1]), - INT_TO_FLOAT(v[2])); -} - -static void GLAPIENTRY -loopback_SecondaryColor3svEXT_f( const GLshort *v ) -{ - SECONDARYCOLORF(SHORT_TO_FLOAT(v[0]), - SHORT_TO_FLOAT(v[1]), - SHORT_TO_FLOAT(v[2])); -} - -static void GLAPIENTRY -loopback_SecondaryColor3uivEXT_f( const GLuint *v ) -{ - SECONDARYCOLORF(UINT_TO_FLOAT(v[0]), - UINT_TO_FLOAT(v[1]), - UINT_TO_FLOAT(v[2])); -} - -static void GLAPIENTRY -loopback_SecondaryColor3usvEXT_f( const GLushort *v ) -{ - SECONDARYCOLORF(USHORT_TO_FLOAT(v[0]), - USHORT_TO_FLOAT(v[1]), - USHORT_TO_FLOAT(v[2])); -} - -static void GLAPIENTRY -loopback_SecondaryColor3ubvEXT_f( const GLubyte *v ) -{ - SECONDARYCOLORF(UBYTE_TO_FLOAT(v[0]), - UBYTE_TO_FLOAT(v[1]), - UBYTE_TO_FLOAT(v[2])); -} - - -/* - * GL_NV_vertex_program: - * Always loop-back to one of the VertexAttrib[1234]f[v]NV functions. - * Note that attribute indexes DO alias conventional vertex attributes. - */ - -static void GLAPIENTRY -loopback_VertexAttrib1sNV(GLuint index, GLshort x) -{ - ATTRIB1NV(index, (GLfloat) x); -} - -static void GLAPIENTRY -loopback_VertexAttrib1dNV(GLuint index, GLdouble x) -{ - ATTRIB1NV(index, (GLfloat) x); -} - -static void GLAPIENTRY -loopback_VertexAttrib2sNV(GLuint index, GLshort x, GLshort y) -{ - ATTRIB2NV(index, (GLfloat) x, y); -} - -static void GLAPIENTRY -loopback_VertexAttrib2dNV(GLuint index, GLdouble x, GLdouble y) -{ - ATTRIB2NV(index, (GLfloat) x, (GLfloat) y); -} - -static void GLAPIENTRY -loopback_VertexAttrib3sNV(GLuint index, GLshort x, GLshort y, GLshort z) -{ - ATTRIB3NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z); -} - -static void GLAPIENTRY -loopback_VertexAttrib3dNV(GLuint index, GLdouble x, GLdouble y, GLdouble z) -{ - ATTRIB4NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); -} - -static void GLAPIENTRY -loopback_VertexAttrib4sNV(GLuint index, GLshort x, GLshort y, GLshort z, GLshort w) -{ - ATTRIB4NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -loopback_VertexAttrib4dNV(GLuint index, GLdouble x, GLdouble y, GLdouble z, GLdouble w) -{ - ATTRIB4NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -loopback_VertexAttrib4ubNV(GLuint index, GLubyte x, GLubyte y, GLubyte z, GLubyte w) -{ - ATTRIB4NV(index, UBYTE_TO_FLOAT(x), UBYTE_TO_FLOAT(y), - UBYTE_TO_FLOAT(z), UBYTE_TO_FLOAT(w)); -} - -static void GLAPIENTRY -loopback_VertexAttrib1svNV(GLuint index, const GLshort *v) -{ - ATTRIB1NV(index, (GLfloat) v[0]); -} - -static void GLAPIENTRY -loopback_VertexAttrib1dvNV(GLuint index, const GLdouble *v) -{ - ATTRIB1NV(index, (GLfloat) v[0]); -} - -static void GLAPIENTRY -loopback_VertexAttrib2svNV(GLuint index, const GLshort *v) -{ - ATTRIB2NV(index, (GLfloat) v[0], (GLfloat) v[1]); -} - -static void GLAPIENTRY -loopback_VertexAttrib2dvNV(GLuint index, const GLdouble *v) -{ - ATTRIB2NV(index, (GLfloat) v[0], (GLfloat) v[1]); -} - -static void GLAPIENTRY -loopback_VertexAttrib3svNV(GLuint index, const GLshort *v) -{ - ATTRIB3NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); -} - -static void GLAPIENTRY -loopback_VertexAttrib3dvNV(GLuint index, const GLdouble *v) -{ - ATTRIB3NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4svNV(GLuint index, const GLshort *v) -{ - ATTRIB4NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], - (GLfloat)v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4dvNV(GLuint index, const GLdouble *v) -{ - ATTRIB4NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4ubvNV(GLuint index, const GLubyte *v) -{ - ATTRIB4NV(index, UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), - UBYTE_TO_FLOAT(v[2]), UBYTE_TO_FLOAT(v[3])); -} - - -static void GLAPIENTRY -loopback_VertexAttribs1svNV(GLuint index, GLsizei n, const GLshort *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib1svNV(index + i, v + i); -} - -static void GLAPIENTRY -loopback_VertexAttribs1fvNV(GLuint index, GLsizei n, const GLfloat *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - ATTRIB1NV(index + i, v[i]); -} - -static void GLAPIENTRY -loopback_VertexAttribs1dvNV(GLuint index, GLsizei n, const GLdouble *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib1dvNV(index + i, v + i); -} - -static void GLAPIENTRY -loopback_VertexAttribs2svNV(GLuint index, GLsizei n, const GLshort *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib2svNV(index + i, v + 2 * i); -} - -static void GLAPIENTRY -loopback_VertexAttribs2fvNV(GLuint index, GLsizei n, const GLfloat *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - ATTRIB2NV(index + i, v[2 * i], v[2 * i + 1]); -} - -static void GLAPIENTRY -loopback_VertexAttribs2dvNV(GLuint index, GLsizei n, const GLdouble *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib2dvNV(index + i, v + 2 * i); -} - -static void GLAPIENTRY -loopback_VertexAttribs3svNV(GLuint index, GLsizei n, const GLshort *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib3svNV(index + i, v + 3 * i); -} - -static void GLAPIENTRY -loopback_VertexAttribs3fvNV(GLuint index, GLsizei n, const GLfloat *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - ATTRIB3NV(index + i, v[3 * i], v[3 * i + 1], v[3 * i + 2]); -} - -static void GLAPIENTRY -loopback_VertexAttribs3dvNV(GLuint index, GLsizei n, const GLdouble *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib3dvNV(index + i, v + 3 * i); -} - -static void GLAPIENTRY -loopback_VertexAttribs4svNV(GLuint index, GLsizei n, const GLshort *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib4svNV(index + i, v + 4 * i); -} - -static void GLAPIENTRY -loopback_VertexAttribs4fvNV(GLuint index, GLsizei n, const GLfloat *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - ATTRIB4NV(index + i, v[4 * i], v[4 * i + 1], v[4 * i + 2], v[4 * i + 3]); -} - -static void GLAPIENTRY -loopback_VertexAttribs4dvNV(GLuint index, GLsizei n, const GLdouble *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib4dvNV(index + i, v + 4 * i); -} - -static void GLAPIENTRY -loopback_VertexAttribs4ubvNV(GLuint index, GLsizei n, const GLubyte *v) -{ - GLint i; - for (i = n - 1; i >= 0; i--) - loopback_VertexAttrib4ubvNV(index + i, v + 4 * i); -} - - -/* - * GL_ARB_vertex_program - * Always loop-back to one of the VertexAttrib[1234]f[v]ARB functions. - * Note that attribute indexes do NOT alias conventional attributes. - */ - -static void GLAPIENTRY -loopback_VertexAttrib1sARB(GLuint index, GLshort x) -{ - ATTRIB1ARB(index, (GLfloat) x); -} - -static void GLAPIENTRY -loopback_VertexAttrib1dARB(GLuint index, GLdouble x) -{ - ATTRIB1ARB(index, (GLfloat) x); -} - -static void GLAPIENTRY -loopback_VertexAttrib2sARB(GLuint index, GLshort x, GLshort y) -{ - ATTRIB2ARB(index, (GLfloat) x, y); -} - -static void GLAPIENTRY -loopback_VertexAttrib2dARB(GLuint index, GLdouble x, GLdouble y) -{ - ATTRIB2ARB(index, (GLfloat) x, (GLfloat) y); -} - -static void GLAPIENTRY -loopback_VertexAttrib3sARB(GLuint index, GLshort x, GLshort y, GLshort z) -{ - ATTRIB3ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z); -} - -static void GLAPIENTRY -loopback_VertexAttrib3dARB(GLuint index, GLdouble x, GLdouble y, GLdouble z) -{ - ATTRIB4ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); -} - -static void GLAPIENTRY -loopback_VertexAttrib4sARB(GLuint index, GLshort x, GLshort y, GLshort z, GLshort w) -{ - ATTRIB4ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -loopback_VertexAttrib4dARB(GLuint index, GLdouble x, GLdouble y, GLdouble z, GLdouble w) -{ - ATTRIB4ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -loopback_VertexAttrib1svARB(GLuint index, const GLshort *v) -{ - ATTRIB1ARB(index, (GLfloat) v[0]); -} - -static void GLAPIENTRY -loopback_VertexAttrib1dvARB(GLuint index, const GLdouble *v) -{ - ATTRIB1ARB(index, (GLfloat) v[0]); -} - -static void GLAPIENTRY -loopback_VertexAttrib2svARB(GLuint index, const GLshort *v) -{ - ATTRIB2ARB(index, (GLfloat) v[0], (GLfloat) v[1]); -} - -static void GLAPIENTRY -loopback_VertexAttrib2dvARB(GLuint index, const GLdouble *v) -{ - ATTRIB2ARB(index, (GLfloat) v[0], (GLfloat) v[1]); -} - -static void GLAPIENTRY -loopback_VertexAttrib3svARB(GLuint index, const GLshort *v) -{ - ATTRIB3ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); -} - -static void GLAPIENTRY -loopback_VertexAttrib3dvARB(GLuint index, const GLdouble *v) -{ - ATTRIB3ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4svARB(GLuint index, const GLshort *v) -{ - ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], - (GLfloat)v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4dvARB(GLuint index, const GLdouble *v) -{ - ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4bvARB(GLuint index, const GLbyte * v) -{ - ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4ivARB(GLuint index, const GLint * v) -{ - ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4ubvARB(GLuint index, const GLubyte * v) -{ - ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4usvARB(GLuint index, const GLushort * v) -{ - ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4uivARB(GLuint index, const GLuint * v) -{ - ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttrib4NbvARB(GLuint index, const GLbyte * v) -{ - ATTRIB4ARB(index, BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), - BYTE_TO_FLOAT(v[2]), BYTE_TO_FLOAT(v[3])); -} - -static void GLAPIENTRY -loopback_VertexAttrib4NsvARB(GLuint index, const GLshort * v) -{ - ATTRIB4ARB(index, SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), - SHORT_TO_FLOAT(v[2]), SHORT_TO_FLOAT(v[3])); -} - -static void GLAPIENTRY -loopback_VertexAttrib4NivARB(GLuint index, const GLint * v) -{ - ATTRIB4ARB(index, INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), - INT_TO_FLOAT(v[2]), INT_TO_FLOAT(v[3])); -} - -static void GLAPIENTRY -loopback_VertexAttrib4NubARB(GLuint index, GLubyte x, GLubyte y, GLubyte z, GLubyte w) -{ - ATTRIB4ARB(index, UBYTE_TO_FLOAT(x), UBYTE_TO_FLOAT(y), - UBYTE_TO_FLOAT(z), UBYTE_TO_FLOAT(w)); -} - -static void GLAPIENTRY -loopback_VertexAttrib4NubvARB(GLuint index, const GLubyte * v) -{ - ATTRIB4ARB(index, UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), - UBYTE_TO_FLOAT(v[2]), UBYTE_TO_FLOAT(v[3])); -} - -static void GLAPIENTRY -loopback_VertexAttrib4NusvARB(GLuint index, const GLushort * v) -{ - ATTRIB4ARB(index, USHORT_TO_FLOAT(v[0]), USHORT_TO_FLOAT(v[1]), - USHORT_TO_FLOAT(v[2]), USHORT_TO_FLOAT(v[3])); -} - -static void GLAPIENTRY -loopback_VertexAttrib4NuivARB(GLuint index, const GLuint * v) -{ - ATTRIB4ARB(index, UINT_TO_FLOAT(v[0]), UINT_TO_FLOAT(v[1]), - UINT_TO_FLOAT(v[2]), UINT_TO_FLOAT(v[3])); -} - - - -/** - * GL_EXT_gpu_shader / GL 3.0 signed/unsigned integer-valued attributes. - * Note that attribute indexes do NOT alias conventional attributes. - */ - -static void GLAPIENTRY -loopback_VertexAttribI1iv(GLuint index, const GLint *v) -{ - ATTRIBI_1I(index, v[0]); -} - -static void GLAPIENTRY -loopback_VertexAttribI1uiv(GLuint index, const GLuint *v) -{ - ATTRIBI_1UI(index, v[0]); -} - -static void GLAPIENTRY -loopback_VertexAttribI4bv(GLuint index, const GLbyte *v) -{ - ATTRIBI_4I(index, v[0], v[1], v[2], v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttribI4sv(GLuint index, const GLshort *v) -{ - ATTRIBI_4I(index, v[0], v[1], v[2], v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttribI4ubv(GLuint index, const GLubyte *v) -{ - ATTRIBI_4UI(index, v[0], v[1], v[2], v[3]); -} - -static void GLAPIENTRY -loopback_VertexAttribI4usv(GLuint index, const GLushort *v) -{ - ATTRIBI_4UI(index, v[0], v[1], v[2], v[3]); -} - - - - -/* - * This code never registers handlers for any of the entry points - * listed in vtxfmt.h. - */ -void -_mesa_loopback_init_api_table( struct _glapi_table *dest ) -{ - SET_Color3b(dest, loopback_Color3b_f); - SET_Color3d(dest, loopback_Color3d_f); - SET_Color3i(dest, loopback_Color3i_f); - SET_Color3s(dest, loopback_Color3s_f); - SET_Color3ui(dest, loopback_Color3ui_f); - SET_Color3us(dest, loopback_Color3us_f); - SET_Color3ub(dest, loopback_Color3ub_f); - SET_Color4b(dest, loopback_Color4b_f); - SET_Color4d(dest, loopback_Color4d_f); - SET_Color4i(dest, loopback_Color4i_f); - SET_Color4s(dest, loopback_Color4s_f); - SET_Color4ui(dest, loopback_Color4ui_f); - SET_Color4us(dest, loopback_Color4us_f); - SET_Color4ub(dest, loopback_Color4ub_f); - SET_Color3bv(dest, loopback_Color3bv_f); - SET_Color3dv(dest, loopback_Color3dv_f); - SET_Color3iv(dest, loopback_Color3iv_f); - SET_Color3sv(dest, loopback_Color3sv_f); - SET_Color3uiv(dest, loopback_Color3uiv_f); - SET_Color3usv(dest, loopback_Color3usv_f); - SET_Color3ubv(dest, loopback_Color3ubv_f); - SET_Color4bv(dest, loopback_Color4bv_f); - SET_Color4dv(dest, loopback_Color4dv_f); - SET_Color4iv(dest, loopback_Color4iv_f); - SET_Color4sv(dest, loopback_Color4sv_f); - SET_Color4uiv(dest, loopback_Color4uiv_f); - SET_Color4usv(dest, loopback_Color4usv_f); - SET_Color4ubv(dest, loopback_Color4ubv_f); - - SET_SecondaryColor3bEXT(dest, loopback_SecondaryColor3bEXT_f); - SET_SecondaryColor3dEXT(dest, loopback_SecondaryColor3dEXT_f); - SET_SecondaryColor3iEXT(dest, loopback_SecondaryColor3iEXT_f); - SET_SecondaryColor3sEXT(dest, loopback_SecondaryColor3sEXT_f); - SET_SecondaryColor3uiEXT(dest, loopback_SecondaryColor3uiEXT_f); - SET_SecondaryColor3usEXT(dest, loopback_SecondaryColor3usEXT_f); - SET_SecondaryColor3ubEXT(dest, loopback_SecondaryColor3ubEXT_f); - SET_SecondaryColor3bvEXT(dest, loopback_SecondaryColor3bvEXT_f); - SET_SecondaryColor3dvEXT(dest, loopback_SecondaryColor3dvEXT_f); - SET_SecondaryColor3ivEXT(dest, loopback_SecondaryColor3ivEXT_f); - SET_SecondaryColor3svEXT(dest, loopback_SecondaryColor3svEXT_f); - SET_SecondaryColor3uivEXT(dest, loopback_SecondaryColor3uivEXT_f); - SET_SecondaryColor3usvEXT(dest, loopback_SecondaryColor3usvEXT_f); - SET_SecondaryColor3ubvEXT(dest, loopback_SecondaryColor3ubvEXT_f); - - SET_EdgeFlagv(dest, loopback_EdgeFlagv); - - SET_Indexd(dest, loopback_Indexd); - SET_Indexi(dest, loopback_Indexi); - SET_Indexs(dest, loopback_Indexs); - SET_Indexub(dest, loopback_Indexub); - SET_Indexdv(dest, loopback_Indexdv); - SET_Indexiv(dest, loopback_Indexiv); - SET_Indexsv(dest, loopback_Indexsv); - SET_Indexubv(dest, loopback_Indexubv); - SET_Normal3b(dest, loopback_Normal3b); - SET_Normal3d(dest, loopback_Normal3d); - SET_Normal3i(dest, loopback_Normal3i); - SET_Normal3s(dest, loopback_Normal3s); - SET_Normal3bv(dest, loopback_Normal3bv); - SET_Normal3dv(dest, loopback_Normal3dv); - SET_Normal3iv(dest, loopback_Normal3iv); - SET_Normal3sv(dest, loopback_Normal3sv); - SET_TexCoord1d(dest, loopback_TexCoord1d); - SET_TexCoord1i(dest, loopback_TexCoord1i); - SET_TexCoord1s(dest, loopback_TexCoord1s); - SET_TexCoord2d(dest, loopback_TexCoord2d); - SET_TexCoord2s(dest, loopback_TexCoord2s); - SET_TexCoord2i(dest, loopback_TexCoord2i); - SET_TexCoord3d(dest, loopback_TexCoord3d); - SET_TexCoord3i(dest, loopback_TexCoord3i); - SET_TexCoord3s(dest, loopback_TexCoord3s); - SET_TexCoord4d(dest, loopback_TexCoord4d); - SET_TexCoord4i(dest, loopback_TexCoord4i); - SET_TexCoord4s(dest, loopback_TexCoord4s); - SET_TexCoord1dv(dest, loopback_TexCoord1dv); - SET_TexCoord1iv(dest, loopback_TexCoord1iv); - SET_TexCoord1sv(dest, loopback_TexCoord1sv); - SET_TexCoord2dv(dest, loopback_TexCoord2dv); - SET_TexCoord2iv(dest, loopback_TexCoord2iv); - SET_TexCoord2sv(dest, loopback_TexCoord2sv); - SET_TexCoord3dv(dest, loopback_TexCoord3dv); - SET_TexCoord3iv(dest, loopback_TexCoord3iv); - SET_TexCoord3sv(dest, loopback_TexCoord3sv); - SET_TexCoord4dv(dest, loopback_TexCoord4dv); - SET_TexCoord4iv(dest, loopback_TexCoord4iv); - SET_TexCoord4sv(dest, loopback_TexCoord4sv); - SET_Vertex2d(dest, loopback_Vertex2d); - SET_Vertex2i(dest, loopback_Vertex2i); - SET_Vertex2s(dest, loopback_Vertex2s); - SET_Vertex3d(dest, loopback_Vertex3d); - SET_Vertex3i(dest, loopback_Vertex3i); - SET_Vertex3s(dest, loopback_Vertex3s); - SET_Vertex4d(dest, loopback_Vertex4d); - SET_Vertex4i(dest, loopback_Vertex4i); - SET_Vertex4s(dest, loopback_Vertex4s); - SET_Vertex2dv(dest, loopback_Vertex2dv); - SET_Vertex2iv(dest, loopback_Vertex2iv); - SET_Vertex2sv(dest, loopback_Vertex2sv); - SET_Vertex3dv(dest, loopback_Vertex3dv); - SET_Vertex3iv(dest, loopback_Vertex3iv); - SET_Vertex3sv(dest, loopback_Vertex3sv); - SET_Vertex4dv(dest, loopback_Vertex4dv); - SET_Vertex4iv(dest, loopback_Vertex4iv); - SET_Vertex4sv(dest, loopback_Vertex4sv); - SET_MultiTexCoord1dARB(dest, loopback_MultiTexCoord1dARB); - SET_MultiTexCoord1dvARB(dest, loopback_MultiTexCoord1dvARB); - SET_MultiTexCoord1iARB(dest, loopback_MultiTexCoord1iARB); - SET_MultiTexCoord1ivARB(dest, loopback_MultiTexCoord1ivARB); - SET_MultiTexCoord1sARB(dest, loopback_MultiTexCoord1sARB); - SET_MultiTexCoord1svARB(dest, loopback_MultiTexCoord1svARB); - SET_MultiTexCoord2dARB(dest, loopback_MultiTexCoord2dARB); - SET_MultiTexCoord2dvARB(dest, loopback_MultiTexCoord2dvARB); - SET_MultiTexCoord2iARB(dest, loopback_MultiTexCoord2iARB); - SET_MultiTexCoord2ivARB(dest, loopback_MultiTexCoord2ivARB); - SET_MultiTexCoord2sARB(dest, loopback_MultiTexCoord2sARB); - SET_MultiTexCoord2svARB(dest, loopback_MultiTexCoord2svARB); - SET_MultiTexCoord3dARB(dest, loopback_MultiTexCoord3dARB); - SET_MultiTexCoord3dvARB(dest, loopback_MultiTexCoord3dvARB); - SET_MultiTexCoord3iARB(dest, loopback_MultiTexCoord3iARB); - SET_MultiTexCoord3ivARB(dest, loopback_MultiTexCoord3ivARB); - SET_MultiTexCoord3sARB(dest, loopback_MultiTexCoord3sARB); - SET_MultiTexCoord3svARB(dest, loopback_MultiTexCoord3svARB); - SET_MultiTexCoord4dARB(dest, loopback_MultiTexCoord4dARB); - SET_MultiTexCoord4dvARB(dest, loopback_MultiTexCoord4dvARB); - SET_MultiTexCoord4iARB(dest, loopback_MultiTexCoord4iARB); - SET_MultiTexCoord4ivARB(dest, loopback_MultiTexCoord4ivARB); - SET_MultiTexCoord4sARB(dest, loopback_MultiTexCoord4sARB); - SET_MultiTexCoord4svARB(dest, loopback_MultiTexCoord4svARB); - SET_EvalCoord2dv(dest, loopback_EvalCoord2dv); - SET_EvalCoord2fv(dest, loopback_EvalCoord2fv); - SET_EvalCoord2d(dest, loopback_EvalCoord2d); - SET_EvalCoord1dv(dest, loopback_EvalCoord1dv); - SET_EvalCoord1fv(dest, loopback_EvalCoord1fv); - SET_EvalCoord1d(dest, loopback_EvalCoord1d); - SET_Materialf(dest, loopback_Materialf); - SET_Materiali(dest, loopback_Materiali); - SET_Materialiv(dest, loopback_Materialiv); - SET_Rectd(dest, loopback_Rectd); - SET_Rectdv(dest, loopback_Rectdv); - SET_Rectfv(dest, loopback_Rectfv); - SET_Recti(dest, loopback_Recti); - SET_Rectiv(dest, loopback_Rectiv); - SET_Rects(dest, loopback_Rects); - SET_Rectsv(dest, loopback_Rectsv); - SET_FogCoorddEXT(dest, loopback_FogCoorddEXT); - SET_FogCoorddvEXT(dest, loopback_FogCoorddvEXT); - - SET_VertexAttrib1sNV(dest, loopback_VertexAttrib1sNV); - SET_VertexAttrib1dNV(dest, loopback_VertexAttrib1dNV); - SET_VertexAttrib2sNV(dest, loopback_VertexAttrib2sNV); - SET_VertexAttrib2dNV(dest, loopback_VertexAttrib2dNV); - SET_VertexAttrib3sNV(dest, loopback_VertexAttrib3sNV); - SET_VertexAttrib3dNV(dest, loopback_VertexAttrib3dNV); - SET_VertexAttrib4sNV(dest, loopback_VertexAttrib4sNV); - SET_VertexAttrib4dNV(dest, loopback_VertexAttrib4dNV); - SET_VertexAttrib4ubNV(dest, loopback_VertexAttrib4ubNV); - SET_VertexAttrib1svNV(dest, loopback_VertexAttrib1svNV); - SET_VertexAttrib1dvNV(dest, loopback_VertexAttrib1dvNV); - SET_VertexAttrib2svNV(dest, loopback_VertexAttrib2svNV); - SET_VertexAttrib2dvNV(dest, loopback_VertexAttrib2dvNV); - SET_VertexAttrib3svNV(dest, loopback_VertexAttrib3svNV); - SET_VertexAttrib3dvNV(dest, loopback_VertexAttrib3dvNV); - SET_VertexAttrib4svNV(dest, loopback_VertexAttrib4svNV); - SET_VertexAttrib4dvNV(dest, loopback_VertexAttrib4dvNV); - SET_VertexAttrib4ubvNV(dest, loopback_VertexAttrib4ubvNV); - SET_VertexAttribs1svNV(dest, loopback_VertexAttribs1svNV); - SET_VertexAttribs1fvNV(dest, loopback_VertexAttribs1fvNV); - SET_VertexAttribs1dvNV(dest, loopback_VertexAttribs1dvNV); - SET_VertexAttribs2svNV(dest, loopback_VertexAttribs2svNV); - SET_VertexAttribs2fvNV(dest, loopback_VertexAttribs2fvNV); - SET_VertexAttribs2dvNV(dest, loopback_VertexAttribs2dvNV); - SET_VertexAttribs3svNV(dest, loopback_VertexAttribs3svNV); - SET_VertexAttribs3fvNV(dest, loopback_VertexAttribs3fvNV); - SET_VertexAttribs3dvNV(dest, loopback_VertexAttribs3dvNV); - SET_VertexAttribs4svNV(dest, loopback_VertexAttribs4svNV); - SET_VertexAttribs4fvNV(dest, loopback_VertexAttribs4fvNV); - SET_VertexAttribs4dvNV(dest, loopback_VertexAttribs4dvNV); - SET_VertexAttribs4ubvNV(dest, loopback_VertexAttribs4ubvNV); - - SET_VertexAttrib1sARB(dest, loopback_VertexAttrib1sARB); - SET_VertexAttrib1dARB(dest, loopback_VertexAttrib1dARB); - SET_VertexAttrib2sARB(dest, loopback_VertexAttrib2sARB); - SET_VertexAttrib2dARB(dest, loopback_VertexAttrib2dARB); - SET_VertexAttrib3sARB(dest, loopback_VertexAttrib3sARB); - SET_VertexAttrib3dARB(dest, loopback_VertexAttrib3dARB); - SET_VertexAttrib4sARB(dest, loopback_VertexAttrib4sARB); - SET_VertexAttrib4dARB(dest, loopback_VertexAttrib4dARB); - SET_VertexAttrib1svARB(dest, loopback_VertexAttrib1svARB); - SET_VertexAttrib1dvARB(dest, loopback_VertexAttrib1dvARB); - SET_VertexAttrib2svARB(dest, loopback_VertexAttrib2svARB); - SET_VertexAttrib2dvARB(dest, loopback_VertexAttrib2dvARB); - SET_VertexAttrib3svARB(dest, loopback_VertexAttrib3svARB); - SET_VertexAttrib3dvARB(dest, loopback_VertexAttrib3dvARB); - SET_VertexAttrib4svARB(dest, loopback_VertexAttrib4svARB); - SET_VertexAttrib4dvARB(dest, loopback_VertexAttrib4dvARB); - SET_VertexAttrib4NubARB(dest, loopback_VertexAttrib4NubARB); - SET_VertexAttrib4NubvARB(dest, loopback_VertexAttrib4NubvARB); - SET_VertexAttrib4bvARB(dest, loopback_VertexAttrib4bvARB); - SET_VertexAttrib4ivARB(dest, loopback_VertexAttrib4ivARB); - SET_VertexAttrib4ubvARB(dest, loopback_VertexAttrib4ubvARB); - SET_VertexAttrib4usvARB(dest, loopback_VertexAttrib4usvARB); - SET_VertexAttrib4uivARB(dest, loopback_VertexAttrib4uivARB); - SET_VertexAttrib4NbvARB(dest, loopback_VertexAttrib4NbvARB); - SET_VertexAttrib4NsvARB(dest, loopback_VertexAttrib4NsvARB); - SET_VertexAttrib4NivARB(dest, loopback_VertexAttrib4NivARB); - SET_VertexAttrib4NusvARB(dest, loopback_VertexAttrib4NusvARB); - SET_VertexAttrib4NuivARB(dest, loopback_VertexAttrib4NuivARB); - - /* GL_EXT_gpu_shader4, GL 3.0 */ - SET_VertexAttribI1ivEXT(dest, loopback_VertexAttribI1iv); - SET_VertexAttribI1uivEXT(dest, loopback_VertexAttribI1uiv); - SET_VertexAttribI4bvEXT(dest, loopback_VertexAttribI4bv); - SET_VertexAttribI4svEXT(dest, loopback_VertexAttribI4sv); - SET_VertexAttribI4ubvEXT(dest, loopback_VertexAttribI4ubv); - SET_VertexAttribI4usvEXT(dest, loopback_VertexAttribI4usv); -} - - -#endif /* FEATURE_beginend */ +/** + * \file api_loopback.c + * + * \author Keith Whitwell + */ + +/* + * Mesa 3-D graphics library + * Version: 6.3 + * + * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#include "glheader.h" +#include "macros.h" +#include "api_loopback.h" +#include "mtypes.h" +#include "glapi/glapi.h" +#include "glapi/glthread.h" +#include "main/dispatch.h" +#include "mfeatures.h" + +/* KW: A set of functions to convert unusual Color/Normal/Vertex/etc + * calls to a smaller set of driver-provided formats. Currently just + * go back to dispatch to find these (eg. call glNormal3f directly), + * hence 'loopback'. + * + * The driver must supply all of the remaining entry points, which are + * listed in dd.h. The easiest way for a driver to do this is to + * install the supplied software t&l module. + */ +#define COLORF(r,g,b,a) CALL_Color4f(GET_DISPATCH(), (r,g,b,a)) +#define VERTEX2(x,y) CALL_Vertex2f(GET_DISPATCH(), (x,y)) +#define VERTEX3(x,y,z) CALL_Vertex3f(GET_DISPATCH(), (x,y,z)) +#define VERTEX4(x,y,z,w) CALL_Vertex4f(GET_DISPATCH(), (x,y,z,w)) +#define NORMAL(x,y,z) CALL_Normal3f(GET_DISPATCH(), (x,y,z)) +#define TEXCOORD1(s) CALL_TexCoord1f(GET_DISPATCH(), (s)) +#define TEXCOORD2(s,t) CALL_TexCoord2f(GET_DISPATCH(), (s,t)) +#define TEXCOORD3(s,t,u) CALL_TexCoord3f(GET_DISPATCH(), (s,t,u)) +#define TEXCOORD4(s,t,u,v) CALL_TexCoord4f(GET_DISPATCH(), (s,t,u,v)) +#define INDEX(c) CALL_Indexf(GET_DISPATCH(), (c)) +#define MULTI_TEXCOORD1(z,s) CALL_MultiTexCoord1fARB(GET_DISPATCH(), (z,s)) +#define MULTI_TEXCOORD2(z,s,t) CALL_MultiTexCoord2fARB(GET_DISPATCH(), (z,s,t)) +#define MULTI_TEXCOORD3(z,s,t,u) CALL_MultiTexCoord3fARB(GET_DISPATCH(), (z,s,t,u)) +#define MULTI_TEXCOORD4(z,s,t,u,v) CALL_MultiTexCoord4fARB(GET_DISPATCH(), (z,s,t,u,v)) +#define EVALCOORD1(x) CALL_EvalCoord1f(GET_DISPATCH(), (x)) +#define EVALCOORD2(x,y) CALL_EvalCoord2f(GET_DISPATCH(), (x,y)) +#define MATERIALFV(a,b,c) CALL_Materialfv(GET_DISPATCH(), (a,b,c)) +#define RECTF(a,b,c,d) CALL_Rectf(GET_DISPATCH(), (a,b,c,d)) + +#define FOGCOORDF(x) CALL_FogCoordfEXT(GET_DISPATCH(), (x)) +#define SECONDARYCOLORF(a,b,c) CALL_SecondaryColor3fEXT(GET_DISPATCH(), (a,b,c)) + +#define ATTRIB1NV(index,x) CALL_VertexAttrib1fNV(GET_DISPATCH(), (index,x)) +#define ATTRIB2NV(index,x,y) CALL_VertexAttrib2fNV(GET_DISPATCH(), (index,x,y)) +#define ATTRIB3NV(index,x,y,z) CALL_VertexAttrib3fNV(GET_DISPATCH(), (index,x,y,z)) +#define ATTRIB4NV(index,x,y,z,w) CALL_VertexAttrib4fNV(GET_DISPATCH(), (index,x,y,z,w)) + +#define ATTRIB1ARB(index,x) CALL_VertexAttrib1fARB(GET_DISPATCH(), (index,x)) +#define ATTRIB2ARB(index,x,y) CALL_VertexAttrib2fARB(GET_DISPATCH(), (index,x,y)) +#define ATTRIB3ARB(index,x,y,z) CALL_VertexAttrib3fARB(GET_DISPATCH(), (index,x,y,z)) +#define ATTRIB4ARB(index,x,y,z,w) CALL_VertexAttrib4fARB(GET_DISPATCH(), (index,x,y,z,w)) + +#define ATTRIBI_1I(index,x) CALL_VertexAttribI1iEXT(GET_DISPATCH(), (index,x)) +#define ATTRIBI_1UI(index,x) CALL_VertexAttribI1uiEXT(GET_DISPATCH(), (index,x)) +#define ATTRIBI_4I(index,x,y,z,w) CALL_VertexAttribI4iEXT(GET_DISPATCH(), (index,x,y,z,w)) + +#define ATTRIBI_4UI(index,x,y,z,w) CALL_VertexAttribI4uiEXT(GET_DISPATCH(), (index,x,y,z,w)) + + +#if FEATURE_beginend + + +static void GLAPIENTRY +loopback_Color3b_f( GLbyte red, GLbyte green, GLbyte blue ) +{ + COLORF( BYTE_TO_FLOAT(red), + BYTE_TO_FLOAT(green), + BYTE_TO_FLOAT(blue), + 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3d_f( GLdouble red, GLdouble green, GLdouble blue ) +{ + COLORF( (GLfloat) red, (GLfloat) green, (GLfloat) blue, 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3i_f( GLint red, GLint green, GLint blue ) +{ + COLORF( INT_TO_FLOAT(red), INT_TO_FLOAT(green), + INT_TO_FLOAT(blue), 1.0); +} + +static void GLAPIENTRY +loopback_Color3s_f( GLshort red, GLshort green, GLshort blue ) +{ + COLORF( SHORT_TO_FLOAT(red), SHORT_TO_FLOAT(green), + SHORT_TO_FLOAT(blue), 1.0); +} + +static void GLAPIENTRY +loopback_Color3ui_f( GLuint red, GLuint green, GLuint blue ) +{ + COLORF( UINT_TO_FLOAT(red), UINT_TO_FLOAT(green), + UINT_TO_FLOAT(blue), 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3us_f( GLushort red, GLushort green, GLushort blue ) +{ + COLORF( USHORT_TO_FLOAT(red), USHORT_TO_FLOAT(green), + USHORT_TO_FLOAT(blue), 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3ub_f( GLubyte red, GLubyte green, GLubyte blue ) +{ + COLORF( UBYTE_TO_FLOAT(red), UBYTE_TO_FLOAT(green), + UBYTE_TO_FLOAT(blue), 1.0 ); +} + + +static void GLAPIENTRY +loopback_Color3bv_f( const GLbyte *v ) +{ + COLORF( BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), + BYTE_TO_FLOAT(v[2]), 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3dv_f( const GLdouble *v ) +{ + COLORF( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3iv_f( const GLint *v ) +{ + COLORF( INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), + INT_TO_FLOAT(v[2]), 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3sv_f( const GLshort *v ) +{ + COLORF( SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), + SHORT_TO_FLOAT(v[2]), 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3uiv_f( const GLuint *v ) +{ + COLORF( UINT_TO_FLOAT(v[0]), UINT_TO_FLOAT(v[1]), + UINT_TO_FLOAT(v[2]), 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3usv_f( const GLushort *v ) +{ + COLORF( USHORT_TO_FLOAT(v[0]), USHORT_TO_FLOAT(v[1]), + USHORT_TO_FLOAT(v[2]), 1.0 ); +} + +static void GLAPIENTRY +loopback_Color3ubv_f( const GLubyte *v ) +{ + COLORF( UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), + UBYTE_TO_FLOAT(v[2]), 1.0 ); +} + + +static void GLAPIENTRY +loopback_Color4b_f( GLbyte red, GLbyte green, GLbyte blue, + GLbyte alpha ) +{ + COLORF( BYTE_TO_FLOAT(red), BYTE_TO_FLOAT(green), + BYTE_TO_FLOAT(blue), BYTE_TO_FLOAT(alpha) ); +} + +static void GLAPIENTRY +loopback_Color4d_f( GLdouble red, GLdouble green, GLdouble blue, + GLdouble alpha ) +{ + COLORF( (GLfloat) red, (GLfloat) green, (GLfloat) blue, (GLfloat) alpha ); +} + +static void GLAPIENTRY +loopback_Color4i_f( GLint red, GLint green, GLint blue, GLint alpha ) +{ + COLORF( INT_TO_FLOAT(red), INT_TO_FLOAT(green), + INT_TO_FLOAT(blue), INT_TO_FLOAT(alpha) ); +} + +static void GLAPIENTRY +loopback_Color4s_f( GLshort red, GLshort green, GLshort blue, + GLshort alpha ) +{ + COLORF( SHORT_TO_FLOAT(red), SHORT_TO_FLOAT(green), + SHORT_TO_FLOAT(blue), SHORT_TO_FLOAT(alpha) ); +} + +static void GLAPIENTRY +loopback_Color4ui_f( GLuint red, GLuint green, GLuint blue, GLuint alpha ) +{ + COLORF( UINT_TO_FLOAT(red), UINT_TO_FLOAT(green), + UINT_TO_FLOAT(blue), UINT_TO_FLOAT(alpha) ); +} + +static void GLAPIENTRY +loopback_Color4us_f( GLushort red, GLushort green, GLushort blue, GLushort alpha ) +{ + COLORF( USHORT_TO_FLOAT(red), USHORT_TO_FLOAT(green), + USHORT_TO_FLOAT(blue), USHORT_TO_FLOAT(alpha) ); +} + +static void GLAPIENTRY +loopback_Color4ub_f( GLubyte red, GLubyte green, GLubyte blue, GLubyte alpha ) +{ + COLORF( UBYTE_TO_FLOAT(red), UBYTE_TO_FLOAT(green), + UBYTE_TO_FLOAT(blue), UBYTE_TO_FLOAT(alpha) ); +} + + +static void GLAPIENTRY +loopback_Color4iv_f( const GLint *v ) +{ + COLORF( INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), + INT_TO_FLOAT(v[2]), INT_TO_FLOAT(v[3]) ); +} + + +static void GLAPIENTRY +loopback_Color4bv_f( const GLbyte *v ) +{ + COLORF( BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), + BYTE_TO_FLOAT(v[2]), BYTE_TO_FLOAT(v[3]) ); +} + +static void GLAPIENTRY +loopback_Color4dv_f( const GLdouble *v ) +{ + COLORF( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3] ); +} + + +static void GLAPIENTRY +loopback_Color4sv_f( const GLshort *v) +{ + COLORF( SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), + SHORT_TO_FLOAT(v[2]), SHORT_TO_FLOAT(v[3]) ); +} + + +static void GLAPIENTRY +loopback_Color4uiv_f( const GLuint *v) +{ + COLORF( UINT_TO_FLOAT(v[0]), UINT_TO_FLOAT(v[1]), + UINT_TO_FLOAT(v[2]), UINT_TO_FLOAT(v[3]) ); +} + +static void GLAPIENTRY +loopback_Color4usv_f( const GLushort *v) +{ + COLORF( USHORT_TO_FLOAT(v[0]), USHORT_TO_FLOAT(v[1]), + USHORT_TO_FLOAT(v[2]), USHORT_TO_FLOAT(v[3]) ); +} + +static void GLAPIENTRY +loopback_Color4ubv_f( const GLubyte *v) +{ + COLORF( UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), + UBYTE_TO_FLOAT(v[2]), UBYTE_TO_FLOAT(v[3]) ); +} + + +static void GLAPIENTRY +loopback_FogCoorddEXT( GLdouble d ) +{ + FOGCOORDF( (GLfloat) d ); +} + +static void GLAPIENTRY +loopback_FogCoorddvEXT( const GLdouble *v ) +{ + FOGCOORDF( (GLfloat) *v ); +} + + +static void GLAPIENTRY +loopback_Indexd( GLdouble c ) +{ + INDEX( (GLfloat) c ); +} + +static void GLAPIENTRY +loopback_Indexi( GLint c ) +{ + INDEX( (GLfloat) c ); +} + +static void GLAPIENTRY +loopback_Indexs( GLshort c ) +{ + INDEX( (GLfloat) c ); +} + +static void GLAPIENTRY +loopback_Indexub( GLubyte c ) +{ + INDEX( (GLfloat) c ); +} + +static void GLAPIENTRY +loopback_Indexdv( const GLdouble *c ) +{ + INDEX( (GLfloat) *c ); +} + +static void GLAPIENTRY +loopback_Indexiv( const GLint *c ) +{ + INDEX( (GLfloat) *c ); +} + +static void GLAPIENTRY +loopback_Indexsv( const GLshort *c ) +{ + INDEX( (GLfloat) *c ); +} + +static void GLAPIENTRY +loopback_Indexubv( const GLubyte *c ) +{ + INDEX( (GLfloat) *c ); +} + + +static void GLAPIENTRY +loopback_EdgeFlagv(const GLboolean *flag) +{ + CALL_EdgeFlag(GET_DISPATCH(), (*flag)); +} + + +static void GLAPIENTRY +loopback_Normal3b( GLbyte nx, GLbyte ny, GLbyte nz ) +{ + NORMAL( BYTE_TO_FLOAT(nx), BYTE_TO_FLOAT(ny), BYTE_TO_FLOAT(nz) ); +} + +static void GLAPIENTRY +loopback_Normal3d( GLdouble nx, GLdouble ny, GLdouble nz ) +{ + NORMAL((GLfloat) nx, (GLfloat) ny, (GLfloat) nz); +} + +static void GLAPIENTRY +loopback_Normal3i( GLint nx, GLint ny, GLint nz ) +{ + NORMAL( INT_TO_FLOAT(nx), INT_TO_FLOAT(ny), INT_TO_FLOAT(nz) ); +} + +static void GLAPIENTRY +loopback_Normal3s( GLshort nx, GLshort ny, GLshort nz ) +{ + NORMAL( SHORT_TO_FLOAT(nx), SHORT_TO_FLOAT(ny), SHORT_TO_FLOAT(nz) ); +} + +static void GLAPIENTRY +loopback_Normal3bv( const GLbyte *v ) +{ + NORMAL( BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), BYTE_TO_FLOAT(v[2]) ); +} + +static void GLAPIENTRY +loopback_Normal3dv( const GLdouble *v ) +{ + NORMAL( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} + +static void GLAPIENTRY +loopback_Normal3iv( const GLint *v ) +{ + NORMAL( INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), INT_TO_FLOAT(v[2]) ); +} + +static void GLAPIENTRY +loopback_Normal3sv( const GLshort *v ) +{ + NORMAL( SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), SHORT_TO_FLOAT(v[2]) ); +} + +static void GLAPIENTRY +loopback_TexCoord1d( GLdouble s ) +{ + TEXCOORD1((GLfloat) s); +} + +static void GLAPIENTRY +loopback_TexCoord1i( GLint s ) +{ + TEXCOORD1((GLfloat) s); +} + +static void GLAPIENTRY +loopback_TexCoord1s( GLshort s ) +{ + TEXCOORD1((GLfloat) s); +} + +static void GLAPIENTRY +loopback_TexCoord2d( GLdouble s, GLdouble t ) +{ + TEXCOORD2((GLfloat) s,(GLfloat) t); +} + +static void GLAPIENTRY +loopback_TexCoord2s( GLshort s, GLshort t ) +{ + TEXCOORD2((GLfloat) s,(GLfloat) t); +} + +static void GLAPIENTRY +loopback_TexCoord2i( GLint s, GLint t ) +{ + TEXCOORD2((GLfloat) s,(GLfloat) t); +} + +static void GLAPIENTRY +loopback_TexCoord3d( GLdouble s, GLdouble t, GLdouble r ) +{ + TEXCOORD3((GLfloat) s,(GLfloat) t,(GLfloat) r); +} + +static void GLAPIENTRY +loopback_TexCoord3i( GLint s, GLint t, GLint r ) +{ + TEXCOORD3((GLfloat) s,(GLfloat) t,(GLfloat) r); +} + +static void GLAPIENTRY +loopback_TexCoord3s( GLshort s, GLshort t, GLshort r ) +{ + TEXCOORD3((GLfloat) s,(GLfloat) t,(GLfloat) r); +} + +static void GLAPIENTRY +loopback_TexCoord4d( GLdouble s, GLdouble t, GLdouble r, GLdouble q ) +{ + TEXCOORD4((GLfloat) s,(GLfloat) t,(GLfloat) r,(GLfloat) q); +} + +static void GLAPIENTRY +loopback_TexCoord4i( GLint s, GLint t, GLint r, GLint q ) +{ + TEXCOORD4((GLfloat) s,(GLfloat) t,(GLfloat) r,(GLfloat) q); +} + +static void GLAPIENTRY +loopback_TexCoord4s( GLshort s, GLshort t, GLshort r, GLshort q ) +{ + TEXCOORD4((GLfloat) s,(GLfloat) t,(GLfloat) r,(GLfloat) q); +} + +static void GLAPIENTRY +loopback_TexCoord1dv( const GLdouble *v ) +{ + TEXCOORD1((GLfloat) v[0]); +} + +static void GLAPIENTRY +loopback_TexCoord1iv( const GLint *v ) +{ + TEXCOORD1((GLfloat) v[0]); +} + +static void GLAPIENTRY +loopback_TexCoord1sv( const GLshort *v ) +{ + TEXCOORD1((GLfloat) v[0]); +} + +static void GLAPIENTRY +loopback_TexCoord2dv( const GLdouble *v ) +{ + TEXCOORD2((GLfloat) v[0],(GLfloat) v[1]); +} + +static void GLAPIENTRY +loopback_TexCoord2iv( const GLint *v ) +{ + TEXCOORD2((GLfloat) v[0],(GLfloat) v[1]); +} + +static void GLAPIENTRY +loopback_TexCoord2sv( const GLshort *v ) +{ + TEXCOORD2((GLfloat) v[0],(GLfloat) v[1]); +} + +static void GLAPIENTRY +loopback_TexCoord3dv( const GLdouble *v ) +{ + TEXCOORD3((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2]); +} + +static void GLAPIENTRY +loopback_TexCoord3iv( const GLint *v ) +{ + TEXCOORD3((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2]); +} + +static void GLAPIENTRY +loopback_TexCoord3sv( const GLshort *v ) +{ + TEXCOORD3((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2]); +} + +static void GLAPIENTRY +loopback_TexCoord4dv( const GLdouble *v ) +{ + TEXCOORD4((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2],(GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_TexCoord4iv( const GLint *v ) +{ + TEXCOORD4((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2],(GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_TexCoord4sv( const GLshort *v ) +{ + TEXCOORD4((GLfloat) v[0],(GLfloat) v[1],(GLfloat) v[2],(GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_Vertex2d( GLdouble x, GLdouble y ) +{ + VERTEX2( (GLfloat) x, (GLfloat) y ); +} + +static void GLAPIENTRY +loopback_Vertex2i( GLint x, GLint y ) +{ + VERTEX2( (GLfloat) x, (GLfloat) y ); +} + +static void GLAPIENTRY +loopback_Vertex2s( GLshort x, GLshort y ) +{ + VERTEX2( (GLfloat) x, (GLfloat) y ); +} + +static void GLAPIENTRY +loopback_Vertex3d( GLdouble x, GLdouble y, GLdouble z ) +{ + VERTEX3( (GLfloat) x, (GLfloat) y, (GLfloat) z ); +} + +static void GLAPIENTRY +loopback_Vertex3i( GLint x, GLint y, GLint z ) +{ + VERTEX3( (GLfloat) x, (GLfloat) y, (GLfloat) z ); +} + +static void GLAPIENTRY +loopback_Vertex3s( GLshort x, GLshort y, GLshort z ) +{ + VERTEX3( (GLfloat) x, (GLfloat) y, (GLfloat) z ); +} + +static void GLAPIENTRY +loopback_Vertex4d( GLdouble x, GLdouble y, GLdouble z, GLdouble w ) +{ + VERTEX4( (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w ); +} + +static void GLAPIENTRY +loopback_Vertex4i( GLint x, GLint y, GLint z, GLint w ) +{ + VERTEX4( (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w ); +} + +static void GLAPIENTRY +loopback_Vertex4s( GLshort x, GLshort y, GLshort z, GLshort w ) +{ + VERTEX4( (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w ); +} + +static void GLAPIENTRY +loopback_Vertex2dv( const GLdouble *v ) +{ + VERTEX2( (GLfloat) v[0], (GLfloat) v[1] ); +} + +static void GLAPIENTRY +loopback_Vertex2iv( const GLint *v ) +{ + VERTEX2( (GLfloat) v[0], (GLfloat) v[1] ); +} + +static void GLAPIENTRY +loopback_Vertex2sv( const GLshort *v ) +{ + VERTEX2( (GLfloat) v[0], (GLfloat) v[1] ); +} + +static void GLAPIENTRY +loopback_Vertex3dv( const GLdouble *v ) +{ + VERTEX3( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} + +static void GLAPIENTRY +loopback_Vertex3iv( const GLint *v ) +{ + VERTEX3( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} + +static void GLAPIENTRY +loopback_Vertex3sv( const GLshort *v ) +{ + VERTEX3( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} + +static void GLAPIENTRY +loopback_Vertex4dv( const GLdouble *v ) +{ + VERTEX4( (GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3] ); +} + +static void GLAPIENTRY +loopback_Vertex4iv( const GLint *v ) +{ + VERTEX4( (GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3] ); +} + +static void GLAPIENTRY +loopback_Vertex4sv( const GLshort *v ) +{ + VERTEX4( (GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord1dARB(GLenum target, GLdouble s) +{ + MULTI_TEXCOORD1( target, (GLfloat) s ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord1dvARB(GLenum target, const GLdouble *v) +{ + MULTI_TEXCOORD1( target, (GLfloat) v[0] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord1iARB(GLenum target, GLint s) +{ + MULTI_TEXCOORD1( target, (GLfloat) s ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord1ivARB(GLenum target, const GLint *v) +{ + MULTI_TEXCOORD1( target, (GLfloat) v[0] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord1sARB(GLenum target, GLshort s) +{ + MULTI_TEXCOORD1( target, (GLfloat) s ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord1svARB(GLenum target, const GLshort *v) +{ + MULTI_TEXCOORD1( target, (GLfloat) v[0] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord2dARB(GLenum target, GLdouble s, GLdouble t) +{ + MULTI_TEXCOORD2( target, (GLfloat) s, (GLfloat) t ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord2dvARB(GLenum target, const GLdouble *v) +{ + MULTI_TEXCOORD2( target, (GLfloat) v[0], (GLfloat) v[1] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord2iARB(GLenum target, GLint s, GLint t) +{ + MULTI_TEXCOORD2( target, (GLfloat) s, (GLfloat) t ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord2ivARB(GLenum target, const GLint *v) +{ + MULTI_TEXCOORD2( target, (GLfloat) v[0], (GLfloat) v[1] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord2sARB(GLenum target, GLshort s, GLshort t) +{ + MULTI_TEXCOORD2( target, (GLfloat) s, (GLfloat) t ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord2svARB(GLenum target, const GLshort *v) +{ + MULTI_TEXCOORD2( target, (GLfloat) v[0], (GLfloat) v[1] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord3dARB(GLenum target, GLdouble s, GLdouble t, GLdouble r) +{ + MULTI_TEXCOORD3( target, (GLfloat) s, (GLfloat) t, (GLfloat) r ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord3dvARB(GLenum target, const GLdouble *v) +{ + MULTI_TEXCOORD3( target, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord3iARB(GLenum target, GLint s, GLint t, GLint r) +{ + MULTI_TEXCOORD3( target, (GLfloat) s, (GLfloat) t, (GLfloat) r ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord3ivARB(GLenum target, const GLint *v) +{ + MULTI_TEXCOORD3( target, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord3sARB(GLenum target, GLshort s, GLshort t, GLshort r) +{ + MULTI_TEXCOORD3( target, (GLfloat) s, (GLfloat) t, (GLfloat) r ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord3svARB(GLenum target, const GLshort *v) +{ + MULTI_TEXCOORD3( target, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord4dARB(GLenum target, GLdouble s, GLdouble t, GLdouble r, GLdouble q) +{ + MULTI_TEXCOORD4( target, (GLfloat) s, (GLfloat) t, + (GLfloat) r, (GLfloat) q ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord4dvARB(GLenum target, const GLdouble *v) +{ + MULTI_TEXCOORD4( target, (GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord4iARB(GLenum target, GLint s, GLint t, GLint r, GLint q) +{ + MULTI_TEXCOORD4( target, (GLfloat) s, (GLfloat) t, + (GLfloat) r, (GLfloat) q ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord4ivARB(GLenum target, const GLint *v) +{ + MULTI_TEXCOORD4( target, (GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3] ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord4sARB(GLenum target, GLshort s, GLshort t, GLshort r, GLshort q) +{ + MULTI_TEXCOORD4( target, (GLfloat) s, (GLfloat) t, + (GLfloat) r, (GLfloat) q ); +} + +static void GLAPIENTRY +loopback_MultiTexCoord4svARB(GLenum target, const GLshort *v) +{ + MULTI_TEXCOORD4( target, (GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3] ); +} + +static void GLAPIENTRY +loopback_EvalCoord2dv( const GLdouble *u ) +{ + EVALCOORD2( (GLfloat) u[0], (GLfloat) u[1] ); +} + +static void GLAPIENTRY +loopback_EvalCoord2fv( const GLfloat *u ) +{ + EVALCOORD2( u[0], u[1] ); +} + +static void GLAPIENTRY +loopback_EvalCoord2d( GLdouble u, GLdouble v ) +{ + EVALCOORD2( (GLfloat) u, (GLfloat) v ); +} + +static void GLAPIENTRY +loopback_EvalCoord1dv( const GLdouble *u ) +{ + EVALCOORD1( (GLfloat) *u ); +} + +static void GLAPIENTRY +loopback_EvalCoord1fv( const GLfloat *u ) +{ + EVALCOORD1( (GLfloat) *u ); +} + +static void GLAPIENTRY +loopback_EvalCoord1d( GLdouble u ) +{ + EVALCOORD1( (GLfloat) u ); +} + +static void GLAPIENTRY +loopback_Materialf( GLenum face, GLenum pname, GLfloat param ) +{ + GLfloat fparam[4]; + fparam[0] = param; + MATERIALFV( face, pname, fparam ); +} + +static void GLAPIENTRY +loopback_Materiali(GLenum face, GLenum pname, GLint param ) +{ + GLfloat p = (GLfloat) param; + MATERIALFV(face, pname, &p); +} + +static void GLAPIENTRY +loopback_Materialiv(GLenum face, GLenum pname, const GLint *params ) +{ + GLfloat fparam[4]; + switch (pname) { + case GL_AMBIENT: + case GL_DIFFUSE: + case GL_SPECULAR: + case GL_EMISSION: + case GL_AMBIENT_AND_DIFFUSE: + fparam[0] = INT_TO_FLOAT( params[0] ); + fparam[1] = INT_TO_FLOAT( params[1] ); + fparam[2] = INT_TO_FLOAT( params[2] ); + fparam[3] = INT_TO_FLOAT( params[3] ); + break; + case GL_SHININESS: + fparam[0] = (GLfloat) params[0]; + break; + case GL_COLOR_INDEXES: + fparam[0] = (GLfloat) params[0]; + fparam[1] = (GLfloat) params[1]; + fparam[2] = (GLfloat) params[2]; + break; + default: + ; + } + MATERIALFV(face, pname, fparam); +} + + +static void GLAPIENTRY +loopback_Rectd(GLdouble x1, GLdouble y1, GLdouble x2, GLdouble y2) +{ + RECTF((GLfloat) x1, (GLfloat) y1, (GLfloat) x2, (GLfloat) y2); +} + +static void GLAPIENTRY +loopback_Rectdv(const GLdouble *v1, const GLdouble *v2) +{ + RECTF((GLfloat) v1[0], (GLfloat) v1[1], (GLfloat) v2[0], (GLfloat) v2[1]); +} + +static void GLAPIENTRY +loopback_Rectfv(const GLfloat *v1, const GLfloat *v2) +{ + RECTF(v1[0], v1[1], v2[0], v2[1]); +} + +static void GLAPIENTRY +loopback_Recti(GLint x1, GLint y1, GLint x2, GLint y2) +{ + RECTF((GLfloat) x1, (GLfloat) y1, (GLfloat) x2, (GLfloat) y2); +} + +static void GLAPIENTRY +loopback_Rectiv(const GLint *v1, const GLint *v2) +{ + RECTF((GLfloat) v1[0], (GLfloat) v1[1], (GLfloat) v2[0], (GLfloat) v2[1]); +} + +static void GLAPIENTRY +loopback_Rects(GLshort x1, GLshort y1, GLshort x2, GLshort y2) +{ + RECTF((GLfloat) x1, (GLfloat) y1, (GLfloat) x2, (GLfloat) y2); +} + +static void GLAPIENTRY +loopback_Rectsv(const GLshort *v1, const GLshort *v2) +{ + RECTF((GLfloat) v1[0], (GLfloat) v1[1], (GLfloat) v2[0], (GLfloat) v2[1]); +} + +static void GLAPIENTRY +loopback_SecondaryColor3bEXT_f( GLbyte red, GLbyte green, GLbyte blue ) +{ + SECONDARYCOLORF( BYTE_TO_FLOAT(red), + BYTE_TO_FLOAT(green), + BYTE_TO_FLOAT(blue) ); +} + +static void GLAPIENTRY +loopback_SecondaryColor3dEXT_f( GLdouble red, GLdouble green, GLdouble blue ) +{ + SECONDARYCOLORF( (GLfloat) red, (GLfloat) green, (GLfloat) blue ); +} + +static void GLAPIENTRY +loopback_SecondaryColor3iEXT_f( GLint red, GLint green, GLint blue ) +{ + SECONDARYCOLORF( INT_TO_FLOAT(red), + INT_TO_FLOAT(green), + INT_TO_FLOAT(blue)); +} + +static void GLAPIENTRY +loopback_SecondaryColor3sEXT_f( GLshort red, GLshort green, GLshort blue ) +{ + SECONDARYCOLORF(SHORT_TO_FLOAT(red), + SHORT_TO_FLOAT(green), + SHORT_TO_FLOAT(blue)); +} + +static void GLAPIENTRY +loopback_SecondaryColor3uiEXT_f( GLuint red, GLuint green, GLuint blue ) +{ + SECONDARYCOLORF(UINT_TO_FLOAT(red), + UINT_TO_FLOAT(green), + UINT_TO_FLOAT(blue)); +} + +static void GLAPIENTRY +loopback_SecondaryColor3usEXT_f( GLushort red, GLushort green, GLushort blue ) +{ + SECONDARYCOLORF(USHORT_TO_FLOAT(red), + USHORT_TO_FLOAT(green), + USHORT_TO_FLOAT(blue)); +} + +static void GLAPIENTRY +loopback_SecondaryColor3ubEXT_f( GLubyte red, GLubyte green, GLubyte blue ) +{ + SECONDARYCOLORF(UBYTE_TO_FLOAT(red), + UBYTE_TO_FLOAT(green), + UBYTE_TO_FLOAT(blue)); +} + +static void GLAPIENTRY +loopback_SecondaryColor3bvEXT_f( const GLbyte *v ) +{ + SECONDARYCOLORF(BYTE_TO_FLOAT(v[0]), + BYTE_TO_FLOAT(v[1]), + BYTE_TO_FLOAT(v[2])); +} + +static void GLAPIENTRY +loopback_SecondaryColor3dvEXT_f( const GLdouble *v ) +{ + SECONDARYCOLORF( (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2] ); +} +static void GLAPIENTRY +loopback_SecondaryColor3ivEXT_f( const GLint *v ) +{ + SECONDARYCOLORF(INT_TO_FLOAT(v[0]), + INT_TO_FLOAT(v[1]), + INT_TO_FLOAT(v[2])); +} + +static void GLAPIENTRY +loopback_SecondaryColor3svEXT_f( const GLshort *v ) +{ + SECONDARYCOLORF(SHORT_TO_FLOAT(v[0]), + SHORT_TO_FLOAT(v[1]), + SHORT_TO_FLOAT(v[2])); +} + +static void GLAPIENTRY +loopback_SecondaryColor3uivEXT_f( const GLuint *v ) +{ + SECONDARYCOLORF(UINT_TO_FLOAT(v[0]), + UINT_TO_FLOAT(v[1]), + UINT_TO_FLOAT(v[2])); +} + +static void GLAPIENTRY +loopback_SecondaryColor3usvEXT_f( const GLushort *v ) +{ + SECONDARYCOLORF(USHORT_TO_FLOAT(v[0]), + USHORT_TO_FLOAT(v[1]), + USHORT_TO_FLOAT(v[2])); +} + +static void GLAPIENTRY +loopback_SecondaryColor3ubvEXT_f( const GLubyte *v ) +{ + SECONDARYCOLORF(UBYTE_TO_FLOAT(v[0]), + UBYTE_TO_FLOAT(v[1]), + UBYTE_TO_FLOAT(v[2])); +} + + +/* + * GL_NV_vertex_program: + * Always loop-back to one of the VertexAttrib[1234]f[v]NV functions. + * Note that attribute indexes DO alias conventional vertex attributes. + */ + +static void GLAPIENTRY +loopback_VertexAttrib1sNV(GLuint index, GLshort x) +{ + ATTRIB1NV(index, (GLfloat) x); +} + +static void GLAPIENTRY +loopback_VertexAttrib1dNV(GLuint index, GLdouble x) +{ + ATTRIB1NV(index, (GLfloat) x); +} + +static void GLAPIENTRY +loopback_VertexAttrib2sNV(GLuint index, GLshort x, GLshort y) +{ + ATTRIB2NV(index, (GLfloat) x, y); +} + +static void GLAPIENTRY +loopback_VertexAttrib2dNV(GLuint index, GLdouble x, GLdouble y) +{ + ATTRIB2NV(index, (GLfloat) x, (GLfloat) y); +} + +static void GLAPIENTRY +loopback_VertexAttrib3sNV(GLuint index, GLshort x, GLshort y, GLshort z) +{ + ATTRIB3NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z); +} + +static void GLAPIENTRY +loopback_VertexAttrib3dNV(GLuint index, GLdouble x, GLdouble y, GLdouble z) +{ + ATTRIB4NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); +} + +static void GLAPIENTRY +loopback_VertexAttrib4sNV(GLuint index, GLshort x, GLshort y, GLshort z, GLshort w) +{ + ATTRIB4NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +loopback_VertexAttrib4dNV(GLuint index, GLdouble x, GLdouble y, GLdouble z, GLdouble w) +{ + ATTRIB4NV(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +loopback_VertexAttrib4ubNV(GLuint index, GLubyte x, GLubyte y, GLubyte z, GLubyte w) +{ + ATTRIB4NV(index, UBYTE_TO_FLOAT(x), UBYTE_TO_FLOAT(y), + UBYTE_TO_FLOAT(z), UBYTE_TO_FLOAT(w)); +} + +static void GLAPIENTRY +loopback_VertexAttrib1svNV(GLuint index, const GLshort *v) +{ + ATTRIB1NV(index, (GLfloat) v[0]); +} + +static void GLAPIENTRY +loopback_VertexAttrib1dvNV(GLuint index, const GLdouble *v) +{ + ATTRIB1NV(index, (GLfloat) v[0]); +} + +static void GLAPIENTRY +loopback_VertexAttrib2svNV(GLuint index, const GLshort *v) +{ + ATTRIB2NV(index, (GLfloat) v[0], (GLfloat) v[1]); +} + +static void GLAPIENTRY +loopback_VertexAttrib2dvNV(GLuint index, const GLdouble *v) +{ + ATTRIB2NV(index, (GLfloat) v[0], (GLfloat) v[1]); +} + +static void GLAPIENTRY +loopback_VertexAttrib3svNV(GLuint index, const GLshort *v) +{ + ATTRIB3NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); +} + +static void GLAPIENTRY +loopback_VertexAttrib3dvNV(GLuint index, const GLdouble *v) +{ + ATTRIB3NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4svNV(GLuint index, const GLshort *v) +{ + ATTRIB4NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], + (GLfloat)v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4dvNV(GLuint index, const GLdouble *v) +{ + ATTRIB4NV(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4ubvNV(GLuint index, const GLubyte *v) +{ + ATTRIB4NV(index, UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), + UBYTE_TO_FLOAT(v[2]), UBYTE_TO_FLOAT(v[3])); +} + + +static void GLAPIENTRY +loopback_VertexAttribs1svNV(GLuint index, GLsizei n, const GLshort *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib1svNV(index + i, v + i); +} + +static void GLAPIENTRY +loopback_VertexAttribs1fvNV(GLuint index, GLsizei n, const GLfloat *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + ATTRIB1NV(index + i, v[i]); +} + +static void GLAPIENTRY +loopback_VertexAttribs1dvNV(GLuint index, GLsizei n, const GLdouble *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib1dvNV(index + i, v + i); +} + +static void GLAPIENTRY +loopback_VertexAttribs2svNV(GLuint index, GLsizei n, const GLshort *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib2svNV(index + i, v + 2 * i); +} + +static void GLAPIENTRY +loopback_VertexAttribs2fvNV(GLuint index, GLsizei n, const GLfloat *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + ATTRIB2NV(index + i, v[2 * i], v[2 * i + 1]); +} + +static void GLAPIENTRY +loopback_VertexAttribs2dvNV(GLuint index, GLsizei n, const GLdouble *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib2dvNV(index + i, v + 2 * i); +} + +static void GLAPIENTRY +loopback_VertexAttribs3svNV(GLuint index, GLsizei n, const GLshort *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib3svNV(index + i, v + 3 * i); +} + +static void GLAPIENTRY +loopback_VertexAttribs3fvNV(GLuint index, GLsizei n, const GLfloat *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + ATTRIB3NV(index + i, v[3 * i], v[3 * i + 1], v[3 * i + 2]); +} + +static void GLAPIENTRY +loopback_VertexAttribs3dvNV(GLuint index, GLsizei n, const GLdouble *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib3dvNV(index + i, v + 3 * i); +} + +static void GLAPIENTRY +loopback_VertexAttribs4svNV(GLuint index, GLsizei n, const GLshort *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib4svNV(index + i, v + 4 * i); +} + +static void GLAPIENTRY +loopback_VertexAttribs4fvNV(GLuint index, GLsizei n, const GLfloat *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + ATTRIB4NV(index + i, v[4 * i], v[4 * i + 1], v[4 * i + 2], v[4 * i + 3]); +} + +static void GLAPIENTRY +loopback_VertexAttribs4dvNV(GLuint index, GLsizei n, const GLdouble *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib4dvNV(index + i, v + 4 * i); +} + +static void GLAPIENTRY +loopback_VertexAttribs4ubvNV(GLuint index, GLsizei n, const GLubyte *v) +{ + GLint i; + for (i = n - 1; i >= 0; i--) + loopback_VertexAttrib4ubvNV(index + i, v + 4 * i); +} + + +/* + * GL_ARB_vertex_program + * Always loop-back to one of the VertexAttrib[1234]f[v]ARB functions. + * Note that attribute indexes do NOT alias conventional attributes. + */ + +static void GLAPIENTRY +loopback_VertexAttrib1sARB(GLuint index, GLshort x) +{ + ATTRIB1ARB(index, (GLfloat) x); +} + +static void GLAPIENTRY +loopback_VertexAttrib1dARB(GLuint index, GLdouble x) +{ + ATTRIB1ARB(index, (GLfloat) x); +} + +static void GLAPIENTRY +loopback_VertexAttrib2sARB(GLuint index, GLshort x, GLshort y) +{ + ATTRIB2ARB(index, (GLfloat) x, y); +} + +static void GLAPIENTRY +loopback_VertexAttrib2dARB(GLuint index, GLdouble x, GLdouble y) +{ + ATTRIB2ARB(index, (GLfloat) x, (GLfloat) y); +} + +static void GLAPIENTRY +loopback_VertexAttrib3sARB(GLuint index, GLshort x, GLshort y, GLshort z) +{ + ATTRIB3ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z); +} + +static void GLAPIENTRY +loopback_VertexAttrib3dARB(GLuint index, GLdouble x, GLdouble y, GLdouble z) +{ + ATTRIB4ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); +} + +static void GLAPIENTRY +loopback_VertexAttrib4sARB(GLuint index, GLshort x, GLshort y, GLshort z, GLshort w) +{ + ATTRIB4ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +loopback_VertexAttrib4dARB(GLuint index, GLdouble x, GLdouble y, GLdouble z, GLdouble w) +{ + ATTRIB4ARB(index, (GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +loopback_VertexAttrib1svARB(GLuint index, const GLshort *v) +{ + ATTRIB1ARB(index, (GLfloat) v[0]); +} + +static void GLAPIENTRY +loopback_VertexAttrib1dvARB(GLuint index, const GLdouble *v) +{ + ATTRIB1ARB(index, (GLfloat) v[0]); +} + +static void GLAPIENTRY +loopback_VertexAttrib2svARB(GLuint index, const GLshort *v) +{ + ATTRIB2ARB(index, (GLfloat) v[0], (GLfloat) v[1]); +} + +static void GLAPIENTRY +loopback_VertexAttrib2dvARB(GLuint index, const GLdouble *v) +{ + ATTRIB2ARB(index, (GLfloat) v[0], (GLfloat) v[1]); +} + +static void GLAPIENTRY +loopback_VertexAttrib3svARB(GLuint index, const GLshort *v) +{ + ATTRIB3ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); +} + +static void GLAPIENTRY +loopback_VertexAttrib3dvARB(GLuint index, const GLdouble *v) +{ + ATTRIB3ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4svARB(GLuint index, const GLshort *v) +{ + ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], + (GLfloat)v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4dvARB(GLuint index, const GLdouble *v) +{ + ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4bvARB(GLuint index, const GLbyte * v) +{ + ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4ivARB(GLuint index, const GLint * v) +{ + ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4ubvARB(GLuint index, const GLubyte * v) +{ + ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4usvARB(GLuint index, const GLushort * v) +{ + ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4uivARB(GLuint index, const GLuint * v) +{ + ATTRIB4ARB(index, (GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttrib4NbvARB(GLuint index, const GLbyte * v) +{ + ATTRIB4ARB(index, BYTE_TO_FLOAT(v[0]), BYTE_TO_FLOAT(v[1]), + BYTE_TO_FLOAT(v[2]), BYTE_TO_FLOAT(v[3])); +} + +static void GLAPIENTRY +loopback_VertexAttrib4NsvARB(GLuint index, const GLshort * v) +{ + ATTRIB4ARB(index, SHORT_TO_FLOAT(v[0]), SHORT_TO_FLOAT(v[1]), + SHORT_TO_FLOAT(v[2]), SHORT_TO_FLOAT(v[3])); +} + +static void GLAPIENTRY +loopback_VertexAttrib4NivARB(GLuint index, const GLint * v) +{ + ATTRIB4ARB(index, INT_TO_FLOAT(v[0]), INT_TO_FLOAT(v[1]), + INT_TO_FLOAT(v[2]), INT_TO_FLOAT(v[3])); +} + +static void GLAPIENTRY +loopback_VertexAttrib4NubARB(GLuint index, GLubyte x, GLubyte y, GLubyte z, GLubyte w) +{ + ATTRIB4ARB(index, UBYTE_TO_FLOAT(x), UBYTE_TO_FLOAT(y), + UBYTE_TO_FLOAT(z), UBYTE_TO_FLOAT(w)); +} + +static void GLAPIENTRY +loopback_VertexAttrib4NubvARB(GLuint index, const GLubyte * v) +{ + ATTRIB4ARB(index, UBYTE_TO_FLOAT(v[0]), UBYTE_TO_FLOAT(v[1]), + UBYTE_TO_FLOAT(v[2]), UBYTE_TO_FLOAT(v[3])); +} + +static void GLAPIENTRY +loopback_VertexAttrib4NusvARB(GLuint index, const GLushort * v) +{ + ATTRIB4ARB(index, USHORT_TO_FLOAT(v[0]), USHORT_TO_FLOAT(v[1]), + USHORT_TO_FLOAT(v[2]), USHORT_TO_FLOAT(v[3])); +} + +static void GLAPIENTRY +loopback_VertexAttrib4NuivARB(GLuint index, const GLuint * v) +{ + ATTRIB4ARB(index, UINT_TO_FLOAT(v[0]), UINT_TO_FLOAT(v[1]), + UINT_TO_FLOAT(v[2]), UINT_TO_FLOAT(v[3])); +} + + + +/** + * GL_EXT_gpu_shader / GL 3.0 signed/unsigned integer-valued attributes. + * Note that attribute indexes do NOT alias conventional attributes. + */ + +static void GLAPIENTRY +loopback_VertexAttribI1iv(GLuint index, const GLint *v) +{ + ATTRIBI_1I(index, v[0]); +} + +static void GLAPIENTRY +loopback_VertexAttribI1uiv(GLuint index, const GLuint *v) +{ + ATTRIBI_1UI(index, v[0]); +} + +static void GLAPIENTRY +loopback_VertexAttribI4bv(GLuint index, const GLbyte *v) +{ + ATTRIBI_4I(index, v[0], v[1], v[2], v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttribI4sv(GLuint index, const GLshort *v) +{ + ATTRIBI_4I(index, v[0], v[1], v[2], v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttribI4ubv(GLuint index, const GLubyte *v) +{ + ATTRIBI_4UI(index, v[0], v[1], v[2], v[3]); +} + +static void GLAPIENTRY +loopback_VertexAttribI4usv(GLuint index, const GLushort *v) +{ + ATTRIBI_4UI(index, v[0], v[1], v[2], v[3]); +} + + + + +/* + * This code never registers handlers for any of the entry points + * listed in vtxfmt.h. + */ +void +_mesa_loopback_init_api_table( struct _glapi_table *dest ) +{ + SET_Color3b(dest, loopback_Color3b_f); + SET_Color3d(dest, loopback_Color3d_f); + SET_Color3i(dest, loopback_Color3i_f); + SET_Color3s(dest, loopback_Color3s_f); + SET_Color3ui(dest, loopback_Color3ui_f); + SET_Color3us(dest, loopback_Color3us_f); + SET_Color3ub(dest, loopback_Color3ub_f); + SET_Color4b(dest, loopback_Color4b_f); + SET_Color4d(dest, loopback_Color4d_f); + SET_Color4i(dest, loopback_Color4i_f); + SET_Color4s(dest, loopback_Color4s_f); + SET_Color4ui(dest, loopback_Color4ui_f); + SET_Color4us(dest, loopback_Color4us_f); + SET_Color4ub(dest, loopback_Color4ub_f); + SET_Color3bv(dest, loopback_Color3bv_f); + SET_Color3dv(dest, loopback_Color3dv_f); + SET_Color3iv(dest, loopback_Color3iv_f); + SET_Color3sv(dest, loopback_Color3sv_f); + SET_Color3uiv(dest, loopback_Color3uiv_f); + SET_Color3usv(dest, loopback_Color3usv_f); + SET_Color3ubv(dest, loopback_Color3ubv_f); + SET_Color4bv(dest, loopback_Color4bv_f); + SET_Color4dv(dest, loopback_Color4dv_f); + SET_Color4iv(dest, loopback_Color4iv_f); + SET_Color4sv(dest, loopback_Color4sv_f); + SET_Color4uiv(dest, loopback_Color4uiv_f); + SET_Color4usv(dest, loopback_Color4usv_f); + SET_Color4ubv(dest, loopback_Color4ubv_f); + + SET_SecondaryColor3bEXT(dest, loopback_SecondaryColor3bEXT_f); + SET_SecondaryColor3dEXT(dest, loopback_SecondaryColor3dEXT_f); + SET_SecondaryColor3iEXT(dest, loopback_SecondaryColor3iEXT_f); + SET_SecondaryColor3sEXT(dest, loopback_SecondaryColor3sEXT_f); + SET_SecondaryColor3uiEXT(dest, loopback_SecondaryColor3uiEXT_f); + SET_SecondaryColor3usEXT(dest, loopback_SecondaryColor3usEXT_f); + SET_SecondaryColor3ubEXT(dest, loopback_SecondaryColor3ubEXT_f); + SET_SecondaryColor3bvEXT(dest, loopback_SecondaryColor3bvEXT_f); + SET_SecondaryColor3dvEXT(dest, loopback_SecondaryColor3dvEXT_f); + SET_SecondaryColor3ivEXT(dest, loopback_SecondaryColor3ivEXT_f); + SET_SecondaryColor3svEXT(dest, loopback_SecondaryColor3svEXT_f); + SET_SecondaryColor3uivEXT(dest, loopback_SecondaryColor3uivEXT_f); + SET_SecondaryColor3usvEXT(dest, loopback_SecondaryColor3usvEXT_f); + SET_SecondaryColor3ubvEXT(dest, loopback_SecondaryColor3ubvEXT_f); + + SET_EdgeFlagv(dest, loopback_EdgeFlagv); + + SET_Indexd(dest, loopback_Indexd); + SET_Indexi(dest, loopback_Indexi); + SET_Indexs(dest, loopback_Indexs); + SET_Indexub(dest, loopback_Indexub); + SET_Indexdv(dest, loopback_Indexdv); + SET_Indexiv(dest, loopback_Indexiv); + SET_Indexsv(dest, loopback_Indexsv); + SET_Indexubv(dest, loopback_Indexubv); + SET_Normal3b(dest, loopback_Normal3b); + SET_Normal3d(dest, loopback_Normal3d); + SET_Normal3i(dest, loopback_Normal3i); + SET_Normal3s(dest, loopback_Normal3s); + SET_Normal3bv(dest, loopback_Normal3bv); + SET_Normal3dv(dest, loopback_Normal3dv); + SET_Normal3iv(dest, loopback_Normal3iv); + SET_Normal3sv(dest, loopback_Normal3sv); + SET_TexCoord1d(dest, loopback_TexCoord1d); + SET_TexCoord1i(dest, loopback_TexCoord1i); + SET_TexCoord1s(dest, loopback_TexCoord1s); + SET_TexCoord2d(dest, loopback_TexCoord2d); + SET_TexCoord2s(dest, loopback_TexCoord2s); + SET_TexCoord2i(dest, loopback_TexCoord2i); + SET_TexCoord3d(dest, loopback_TexCoord3d); + SET_TexCoord3i(dest, loopback_TexCoord3i); + SET_TexCoord3s(dest, loopback_TexCoord3s); + SET_TexCoord4d(dest, loopback_TexCoord4d); + SET_TexCoord4i(dest, loopback_TexCoord4i); + SET_TexCoord4s(dest, loopback_TexCoord4s); + SET_TexCoord1dv(dest, loopback_TexCoord1dv); + SET_TexCoord1iv(dest, loopback_TexCoord1iv); + SET_TexCoord1sv(dest, loopback_TexCoord1sv); + SET_TexCoord2dv(dest, loopback_TexCoord2dv); + SET_TexCoord2iv(dest, loopback_TexCoord2iv); + SET_TexCoord2sv(dest, loopback_TexCoord2sv); + SET_TexCoord3dv(dest, loopback_TexCoord3dv); + SET_TexCoord3iv(dest, loopback_TexCoord3iv); + SET_TexCoord3sv(dest, loopback_TexCoord3sv); + SET_TexCoord4dv(dest, loopback_TexCoord4dv); + SET_TexCoord4iv(dest, loopback_TexCoord4iv); + SET_TexCoord4sv(dest, loopback_TexCoord4sv); + SET_Vertex2d(dest, loopback_Vertex2d); + SET_Vertex2i(dest, loopback_Vertex2i); + SET_Vertex2s(dest, loopback_Vertex2s); + SET_Vertex3d(dest, loopback_Vertex3d); + SET_Vertex3i(dest, loopback_Vertex3i); + SET_Vertex3s(dest, loopback_Vertex3s); + SET_Vertex4d(dest, loopback_Vertex4d); + SET_Vertex4i(dest, loopback_Vertex4i); + SET_Vertex4s(dest, loopback_Vertex4s); + SET_Vertex2dv(dest, loopback_Vertex2dv); + SET_Vertex2iv(dest, loopback_Vertex2iv); + SET_Vertex2sv(dest, loopback_Vertex2sv); + SET_Vertex3dv(dest, loopback_Vertex3dv); + SET_Vertex3iv(dest, loopback_Vertex3iv); + SET_Vertex3sv(dest, loopback_Vertex3sv); + SET_Vertex4dv(dest, loopback_Vertex4dv); + SET_Vertex4iv(dest, loopback_Vertex4iv); + SET_Vertex4sv(dest, loopback_Vertex4sv); + SET_MultiTexCoord1dARB(dest, loopback_MultiTexCoord1dARB); + SET_MultiTexCoord1dvARB(dest, loopback_MultiTexCoord1dvARB); + SET_MultiTexCoord1iARB(dest, loopback_MultiTexCoord1iARB); + SET_MultiTexCoord1ivARB(dest, loopback_MultiTexCoord1ivARB); + SET_MultiTexCoord1sARB(dest, loopback_MultiTexCoord1sARB); + SET_MultiTexCoord1svARB(dest, loopback_MultiTexCoord1svARB); + SET_MultiTexCoord2dARB(dest, loopback_MultiTexCoord2dARB); + SET_MultiTexCoord2dvARB(dest, loopback_MultiTexCoord2dvARB); + SET_MultiTexCoord2iARB(dest, loopback_MultiTexCoord2iARB); + SET_MultiTexCoord2ivARB(dest, loopback_MultiTexCoord2ivARB); + SET_MultiTexCoord2sARB(dest, loopback_MultiTexCoord2sARB); + SET_MultiTexCoord2svARB(dest, loopback_MultiTexCoord2svARB); + SET_MultiTexCoord3dARB(dest, loopback_MultiTexCoord3dARB); + SET_MultiTexCoord3dvARB(dest, loopback_MultiTexCoord3dvARB); + SET_MultiTexCoord3iARB(dest, loopback_MultiTexCoord3iARB); + SET_MultiTexCoord3ivARB(dest, loopback_MultiTexCoord3ivARB); + SET_MultiTexCoord3sARB(dest, loopback_MultiTexCoord3sARB); + SET_MultiTexCoord3svARB(dest, loopback_MultiTexCoord3svARB); + SET_MultiTexCoord4dARB(dest, loopback_MultiTexCoord4dARB); + SET_MultiTexCoord4dvARB(dest, loopback_MultiTexCoord4dvARB); + SET_MultiTexCoord4iARB(dest, loopback_MultiTexCoord4iARB); + SET_MultiTexCoord4ivARB(dest, loopback_MultiTexCoord4ivARB); + SET_MultiTexCoord4sARB(dest, loopback_MultiTexCoord4sARB); + SET_MultiTexCoord4svARB(dest, loopback_MultiTexCoord4svARB); + SET_EvalCoord2dv(dest, loopback_EvalCoord2dv); + SET_EvalCoord2fv(dest, loopback_EvalCoord2fv); + SET_EvalCoord2d(dest, loopback_EvalCoord2d); + SET_EvalCoord1dv(dest, loopback_EvalCoord1dv); + SET_EvalCoord1fv(dest, loopback_EvalCoord1fv); + SET_EvalCoord1d(dest, loopback_EvalCoord1d); + SET_Materialf(dest, loopback_Materialf); + SET_Materiali(dest, loopback_Materiali); + SET_Materialiv(dest, loopback_Materialiv); + SET_Rectd(dest, loopback_Rectd); + SET_Rectdv(dest, loopback_Rectdv); + SET_Rectfv(dest, loopback_Rectfv); + SET_Recti(dest, loopback_Recti); + SET_Rectiv(dest, loopback_Rectiv); + SET_Rects(dest, loopback_Rects); + SET_Rectsv(dest, loopback_Rectsv); + SET_FogCoorddEXT(dest, loopback_FogCoorddEXT); + SET_FogCoorddvEXT(dest, loopback_FogCoorddvEXT); + + SET_VertexAttrib1sNV(dest, loopback_VertexAttrib1sNV); + SET_VertexAttrib1dNV(dest, loopback_VertexAttrib1dNV); + SET_VertexAttrib2sNV(dest, loopback_VertexAttrib2sNV); + SET_VertexAttrib2dNV(dest, loopback_VertexAttrib2dNV); + SET_VertexAttrib3sNV(dest, loopback_VertexAttrib3sNV); + SET_VertexAttrib3dNV(dest, loopback_VertexAttrib3dNV); + SET_VertexAttrib4sNV(dest, loopback_VertexAttrib4sNV); + SET_VertexAttrib4dNV(dest, loopback_VertexAttrib4dNV); + SET_VertexAttrib4ubNV(dest, loopback_VertexAttrib4ubNV); + SET_VertexAttrib1svNV(dest, loopback_VertexAttrib1svNV); + SET_VertexAttrib1dvNV(dest, loopback_VertexAttrib1dvNV); + SET_VertexAttrib2svNV(dest, loopback_VertexAttrib2svNV); + SET_VertexAttrib2dvNV(dest, loopback_VertexAttrib2dvNV); + SET_VertexAttrib3svNV(dest, loopback_VertexAttrib3svNV); + SET_VertexAttrib3dvNV(dest, loopback_VertexAttrib3dvNV); + SET_VertexAttrib4svNV(dest, loopback_VertexAttrib4svNV); + SET_VertexAttrib4dvNV(dest, loopback_VertexAttrib4dvNV); + SET_VertexAttrib4ubvNV(dest, loopback_VertexAttrib4ubvNV); + SET_VertexAttribs1svNV(dest, loopback_VertexAttribs1svNV); + SET_VertexAttribs1fvNV(dest, loopback_VertexAttribs1fvNV); + SET_VertexAttribs1dvNV(dest, loopback_VertexAttribs1dvNV); + SET_VertexAttribs2svNV(dest, loopback_VertexAttribs2svNV); + SET_VertexAttribs2fvNV(dest, loopback_VertexAttribs2fvNV); + SET_VertexAttribs2dvNV(dest, loopback_VertexAttribs2dvNV); + SET_VertexAttribs3svNV(dest, loopback_VertexAttribs3svNV); + SET_VertexAttribs3fvNV(dest, loopback_VertexAttribs3fvNV); + SET_VertexAttribs3dvNV(dest, loopback_VertexAttribs3dvNV); + SET_VertexAttribs4svNV(dest, loopback_VertexAttribs4svNV); + SET_VertexAttribs4fvNV(dest, loopback_VertexAttribs4fvNV); + SET_VertexAttribs4dvNV(dest, loopback_VertexAttribs4dvNV); + SET_VertexAttribs4ubvNV(dest, loopback_VertexAttribs4ubvNV); + + SET_VertexAttrib1sARB(dest, loopback_VertexAttrib1sARB); + SET_VertexAttrib1dARB(dest, loopback_VertexAttrib1dARB); + SET_VertexAttrib2sARB(dest, loopback_VertexAttrib2sARB); + SET_VertexAttrib2dARB(dest, loopback_VertexAttrib2dARB); + SET_VertexAttrib3sARB(dest, loopback_VertexAttrib3sARB); + SET_VertexAttrib3dARB(dest, loopback_VertexAttrib3dARB); + SET_VertexAttrib4sARB(dest, loopback_VertexAttrib4sARB); + SET_VertexAttrib4dARB(dest, loopback_VertexAttrib4dARB); + SET_VertexAttrib1svARB(dest, loopback_VertexAttrib1svARB); + SET_VertexAttrib1dvARB(dest, loopback_VertexAttrib1dvARB); + SET_VertexAttrib2svARB(dest, loopback_VertexAttrib2svARB); + SET_VertexAttrib2dvARB(dest, loopback_VertexAttrib2dvARB); + SET_VertexAttrib3svARB(dest, loopback_VertexAttrib3svARB); + SET_VertexAttrib3dvARB(dest, loopback_VertexAttrib3dvARB); + SET_VertexAttrib4svARB(dest, loopback_VertexAttrib4svARB); + SET_VertexAttrib4dvARB(dest, loopback_VertexAttrib4dvARB); + SET_VertexAttrib4NubARB(dest, loopback_VertexAttrib4NubARB); + SET_VertexAttrib4NubvARB(dest, loopback_VertexAttrib4NubvARB); + SET_VertexAttrib4bvARB(dest, loopback_VertexAttrib4bvARB); + SET_VertexAttrib4ivARB(dest, loopback_VertexAttrib4ivARB); + SET_VertexAttrib4ubvARB(dest, loopback_VertexAttrib4ubvARB); + SET_VertexAttrib4usvARB(dest, loopback_VertexAttrib4usvARB); + SET_VertexAttrib4uivARB(dest, loopback_VertexAttrib4uivARB); + SET_VertexAttrib4NbvARB(dest, loopback_VertexAttrib4NbvARB); + SET_VertexAttrib4NsvARB(dest, loopback_VertexAttrib4NsvARB); + SET_VertexAttrib4NivARB(dest, loopback_VertexAttrib4NivARB); + SET_VertexAttrib4NusvARB(dest, loopback_VertexAttrib4NusvARB); + SET_VertexAttrib4NuivARB(dest, loopback_VertexAttrib4NuivARB); + + /* GL_EXT_gpu_shader4, GL 3.0 */ + SET_VertexAttribI1ivEXT(dest, loopback_VertexAttribI1iv); + SET_VertexAttribI1uivEXT(dest, loopback_VertexAttribI1uiv); + SET_VertexAttribI4bvEXT(dest, loopback_VertexAttribI4bv); + SET_VertexAttribI4svEXT(dest, loopback_VertexAttribI4sv); + SET_VertexAttribI4ubvEXT(dest, loopback_VertexAttribI4ubv); + SET_VertexAttribI4usvEXT(dest, loopback_VertexAttribI4usv); +} + + +#endif /* FEATURE_beginend */ diff --git a/mesalib/src/mesa/main/atifragshader.c b/mesalib/src/mesa/main/atifragshader.c index ff7b042d2..c74c999f8 100644 --- a/mesalib/src/mesa/main/atifragshader.c +++ b/mesalib/src/mesa/main/atifragshader.c @@ -1,796 +1,796 @@ -/** - * \file atifragshader.c - * \author David Airlie - * Copyright (C) 2004 David Airlie All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * DAVID AIRLIE BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "main/hash.h" -#include "main/imports.h" -#include "main/macros.h" -#include "main/mfeatures.h" -#include "main/enums.h" -#include "main/mtypes.h" -#include "main/dispatch.h" -#include "main/atifragshader.h" - -#if FEATURE_ATI_fragment_shader - -#define MESA_DEBUG_ATI_FS 0 - -static struct ati_fragment_shader DummyShader; - - -void -_mesa_init_ati_fragment_shader_dispatch(struct _glapi_table *disp) -{ - SET_GenFragmentShadersATI(disp, _mesa_GenFragmentShadersATI); - SET_BindFragmentShaderATI(disp, _mesa_BindFragmentShaderATI); - SET_DeleteFragmentShaderATI(disp, _mesa_DeleteFragmentShaderATI); - SET_BeginFragmentShaderATI(disp, _mesa_BeginFragmentShaderATI); - SET_EndFragmentShaderATI(disp, _mesa_EndFragmentShaderATI); - SET_PassTexCoordATI(disp, _mesa_PassTexCoordATI); - SET_SampleMapATI(disp, _mesa_SampleMapATI); - SET_ColorFragmentOp1ATI(disp, _mesa_ColorFragmentOp1ATI); - SET_ColorFragmentOp2ATI(disp, _mesa_ColorFragmentOp2ATI); - SET_ColorFragmentOp3ATI(disp, _mesa_ColorFragmentOp3ATI); - SET_AlphaFragmentOp1ATI(disp, _mesa_AlphaFragmentOp1ATI); - SET_AlphaFragmentOp2ATI(disp, _mesa_AlphaFragmentOp2ATI); - SET_AlphaFragmentOp3ATI(disp, _mesa_AlphaFragmentOp3ATI); - SET_SetFragmentShaderConstantATI(disp, _mesa_SetFragmentShaderConstantATI); -} - - -/** - * Allocate and initialize a new ATI fragment shader object. - */ -struct ati_fragment_shader * -_mesa_new_ati_fragment_shader(struct gl_context *ctx, GLuint id) -{ - struct ati_fragment_shader *s = CALLOC_STRUCT(ati_fragment_shader); - (void) ctx; - if (s) { - s->Id = id; - s->RefCount = 1; - } - return s; -} - - -/** - * Delete the given ati fragment shader - */ -void -_mesa_delete_ati_fragment_shader(struct gl_context *ctx, struct ati_fragment_shader *s) -{ - GLuint i; - for (i = 0; i < MAX_NUM_PASSES_ATI; i++) { - if (s->Instructions[i]) - free(s->Instructions[i]); - if (s->SetupInst[i]) - free(s->SetupInst[i]); - } - free(s); -} - - - -static void -new_arith_inst(struct ati_fragment_shader *prog) -{ -/* set "default" instruction as not all may get defined. - there is no specified way to express a nop with ati fragment shaders we use - GL_NONE as the op enum and just set some params to 0 - so nothing to do here */ - prog->numArithInstr[prog->cur_pass >> 1]++; -} - -static void -new_tex_inst(struct ati_fragment_shader *prog) -{ -} - -static void match_pair_inst(struct ati_fragment_shader *curProg, GLuint optype) -{ - if (optype == curProg->last_optype) { - curProg->last_optype = 1; - } -} - -#if MESA_DEBUG_ATI_FS -static char * -create_dst_mod_str(GLuint mod) -{ - static char ret_str[1024]; - - memset(ret_str, 0, 1024); - if (mod & GL_2X_BIT_ATI) - strncat(ret_str, "|2X", 1024); - - if (mod & GL_4X_BIT_ATI) - strncat(ret_str, "|4X", 1024); - - if (mod & GL_8X_BIT_ATI) - strncat(ret_str, "|8X", 1024); - if (mod & GL_HALF_BIT_ATI) - strncat(ret_str, "|HA", 1024); - if (mod & GL_QUARTER_BIT_ATI) - strncat(ret_str, "|QU", 1024); - if (mod & GL_EIGHTH_BIT_ATI) - strncat(ret_str, "|EI", 1024); - - if (mod & GL_SATURATE_BIT_ATI) - strncat(ret_str, "|SAT", 1024); - - if (strlen(ret_str) == 0) - strncat(ret_str, "NONE", 1024); - return ret_str; -} - -static char *atifs_ops[] = {"ColorFragmentOp1ATI", "ColorFragmentOp2ATI", "ColorFragmentOp3ATI", - "AlphaFragmentOp1ATI", "AlphaFragmentOp2ATI", "AlphaFragmentOp3ATI" }; - -static void debug_op(GLint optype, GLuint arg_count, GLenum op, GLuint dst, - GLuint dstMask, GLuint dstMod, GLuint arg1, - GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, - GLuint arg2Rep, GLuint arg2Mod, GLuint arg3, - GLuint arg3Rep, GLuint arg3Mod) -{ - char *op_name; - - op_name = atifs_ops[(arg_count-1)+(optype?3:0)]; - - fprintf(stderr, "%s(%s, %s", op_name, _mesa_lookup_enum_by_nr(op), - _mesa_lookup_enum_by_nr(dst)); - if (!optype) - fprintf(stderr, ", %d", dstMask); - - fprintf(stderr, ", %s", create_dst_mod_str(dstMod)); - - fprintf(stderr, ", %s, %s, %d", _mesa_lookup_enum_by_nr(arg1), - _mesa_lookup_enum_by_nr(arg1Rep), arg1Mod); - if (arg_count>1) - fprintf(stderr, ", %s, %s, %d", _mesa_lookup_enum_by_nr(arg2), - _mesa_lookup_enum_by_nr(arg2Rep), arg2Mod); - if (arg_count>2) - fprintf(stderr, ", %s, %s, %d", _mesa_lookup_enum_by_nr(arg3), - _mesa_lookup_enum_by_nr(arg3Rep), arg3Mod); - - fprintf(stderr,")\n"); - -} -#endif - -static int check_arith_arg(struct ati_fragment_shader *curProg, - GLuint optype, GLuint arg, GLuint argRep) -{ - GET_CURRENT_CONTEXT(ctx); - - if (((arg < GL_CON_0_ATI) || (arg > GL_CON_7_ATI)) && - ((arg < GL_REG_0_ATI) || (arg > GL_REG_5_ATI)) && - (arg != GL_ZERO) && (arg != GL_ONE) && - (arg != GL_PRIMARY_COLOR_ARB) && (arg != GL_SECONDARY_INTERPOLATOR_ATI)) { - _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(arg)"); - return 0; - } - if ((arg == GL_SECONDARY_INTERPOLATOR_ATI) && (((optype == 0) && (argRep == GL_ALPHA)) || - ((optype == 1) && ((arg == GL_ALPHA) || (argRep == GL_NONE))))) { - _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(sec_interp)"); - return 0; - } - if ((arg == GL_SECONDARY_INTERPOLATOR_ATI) && (((optype == 0) && (argRep == GL_ALPHA)) || - ((optype == 1) && ((arg == GL_ALPHA) || (argRep == GL_NONE))))) { - _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(sec_interp)"); - return 0; - } - if ((curProg->cur_pass == 1) && - ((arg == GL_PRIMARY_COLOR_ARB) || (arg == GL_SECONDARY_INTERPOLATOR_ATI))) { - curProg->interpinp1 = GL_TRUE; - } - return 1; -} - -GLuint GLAPIENTRY -_mesa_GenFragmentShadersATI(GLuint range) -{ - GLuint first; - GLuint i; - GET_CURRENT_CONTEXT(ctx); - - if (range == 0) { - _mesa_error(ctx, GL_INVALID_VALUE, "glGenFragmentShadersATI(range)"); - return 0; - } - - if (ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glGenFragmentShadersATI(insideShader)"); - return 0; - } - - first = _mesa_HashFindFreeKeyBlock(ctx->Shared->ATIShaders, range); - for (i = 0; i < range; i++) { - _mesa_HashInsert(ctx->Shared->ATIShaders, first + i, &DummyShader); - } - - return first; -} - -void GLAPIENTRY -_mesa_BindFragmentShaderATI(GLuint id) -{ - GET_CURRENT_CONTEXT(ctx); - struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; - struct ati_fragment_shader *newProg; - - if (ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glBindFragmentShaderATI(insideShader)"); - return; - } - - FLUSH_VERTICES(ctx, _NEW_PROGRAM); - - if (curProg->Id == id) { - return; - } - - /* unbind current */ - if (curProg->Id != 0) { - curProg->RefCount--; - if (curProg->RefCount <= 0) { - _mesa_HashRemove(ctx->Shared->ATIShaders, id); - } - } - - /* find new shader */ - if (id == 0) { - newProg = ctx->Shared->DefaultFragmentShader; - } - else { - newProg = (struct ati_fragment_shader *) - _mesa_HashLookup(ctx->Shared->ATIShaders, id); - if (!newProg || newProg == &DummyShader) { - /* allocate a new program now */ - newProg = _mesa_new_ati_fragment_shader(ctx, id); - if (!newProg) { - _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBindFragmentShaderATI"); - return; - } - _mesa_HashInsert(ctx->Shared->ATIShaders, id, newProg); - } - - } - - /* do actual bind */ - ctx->ATIFragmentShader.Current = newProg; - - ASSERT(ctx->ATIFragmentShader.Current); - if (newProg) - newProg->RefCount++; - - /*if (ctx->Driver.BindProgram) - ctx->Driver.BindProgram(ctx, target, prog); */ -} - -void GLAPIENTRY -_mesa_DeleteFragmentShaderATI(GLuint id) -{ - GET_CURRENT_CONTEXT(ctx); - - if (ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glDeleteFragmentShaderATI(insideShader)"); - return; - } - - if (id != 0) { - struct ati_fragment_shader *prog = (struct ati_fragment_shader *) - _mesa_HashLookup(ctx->Shared->ATIShaders, id); - if (prog == &DummyShader) { - _mesa_HashRemove(ctx->Shared->ATIShaders, id); - } - else if (prog) { - if (ctx->ATIFragmentShader.Current && - ctx->ATIFragmentShader.Current->Id == id) { - FLUSH_VERTICES(ctx, _NEW_PROGRAM); - _mesa_BindFragmentShaderATI(0); - } - } - - /* The ID is immediately available for re-use now */ - _mesa_HashRemove(ctx->Shared->ATIShaders, id); - if (prog) { - prog->RefCount--; - if (prog->RefCount <= 0) { - assert(prog != &DummyShader); - free(prog); - } - } - } -} - - -void GLAPIENTRY -_mesa_BeginFragmentShaderATI(void) -{ - GLint i; - GET_CURRENT_CONTEXT(ctx); - - if (ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glBeginFragmentShaderATI(insideShader)"); - return; - } - - FLUSH_VERTICES(ctx, _NEW_PROGRAM); - - /* if the shader was already defined free instructions and get new ones - (or, could use the same mem but would need to reinitialize) */ - /* no idea if it's allowed to redefine a shader */ - for (i = 0; i < MAX_NUM_PASSES_ATI; i++) { - if (ctx->ATIFragmentShader.Current->Instructions[i]) - free(ctx->ATIFragmentShader.Current->Instructions[i]); - if (ctx->ATIFragmentShader.Current->SetupInst[i]) - free(ctx->ATIFragmentShader.Current->SetupInst[i]); - } - - /* malloc the instructions here - not sure if the best place but its - a start */ - for (i = 0; i < MAX_NUM_PASSES_ATI; i++) { - ctx->ATIFragmentShader.Current->Instructions[i] = - (struct atifs_instruction *) - calloc(1, sizeof(struct atifs_instruction) * - (MAX_NUM_INSTRUCTIONS_PER_PASS_ATI)); - ctx->ATIFragmentShader.Current->SetupInst[i] = - (struct atifs_setupinst *) - calloc(1, sizeof(struct atifs_setupinst) * - (MAX_NUM_FRAGMENT_REGISTERS_ATI)); - } - -/* can't rely on calloc for initialization as it's possible to redefine a shader (?) */ - ctx->ATIFragmentShader.Current->LocalConstDef = 0; - ctx->ATIFragmentShader.Current->numArithInstr[0] = 0; - ctx->ATIFragmentShader.Current->numArithInstr[1] = 0; - ctx->ATIFragmentShader.Current->regsAssigned[0] = 0; - ctx->ATIFragmentShader.Current->regsAssigned[1] = 0; - ctx->ATIFragmentShader.Current->NumPasses = 0; - ctx->ATIFragmentShader.Current->cur_pass = 0; - ctx->ATIFragmentShader.Current->last_optype = 0; - ctx->ATIFragmentShader.Current->interpinp1 = GL_FALSE; - ctx->ATIFragmentShader.Current->isValid = GL_FALSE; - ctx->ATIFragmentShader.Current->swizzlerq = 0; - ctx->ATIFragmentShader.Compiling = 1; -} - -void GLAPIENTRY -_mesa_EndFragmentShaderATI(void) -{ - GET_CURRENT_CONTEXT(ctx); - struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; -#if MESA_DEBUG_ATI_FS - GLint i, j; -#endif - - if (!ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glEndFragmentShaderATI(outsideShader)"); - return; - } - if (curProg->interpinp1 && (ctx->ATIFragmentShader.Current->cur_pass > 1)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glEndFragmentShaderATI(interpinfirstpass)"); - /* according to spec, DON'T return here */ - } - - match_pair_inst(curProg, 0); - ctx->ATIFragmentShader.Compiling = 0; - ctx->ATIFragmentShader.Current->isValid = GL_TRUE; - if ((ctx->ATIFragmentShader.Current->cur_pass == 0) || - (ctx->ATIFragmentShader.Current->cur_pass == 2)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glEndFragmentShaderATI(noarithinst)"); - } - if (ctx->ATIFragmentShader.Current->cur_pass > 1) - ctx->ATIFragmentShader.Current->NumPasses = 2; - else - ctx->ATIFragmentShader.Current->NumPasses = 1; - - ctx->ATIFragmentShader.Current->cur_pass = 0; - -#if MESA_DEBUG_ATI_FS - for (j = 0; j < MAX_NUM_PASSES_ATI; j++) { - for (i = 0; i < MAX_NUM_FRAGMENT_REGISTERS_ATI; i++) { - GLuint op = curProg->SetupInst[j][i].Opcode; - const char *op_enum = op > 5 ? _mesa_lookup_enum_by_nr(op) : "0"; - GLuint src = curProg->SetupInst[j][i].src; - GLuint swizzle = curProg->SetupInst[j][i].swizzle; - fprintf(stderr, "%2d %04X %s %d %04X\n", i, op, op_enum, src, - swizzle); - } - for (i = 0; i < curProg->numArithInstr[j]; i++) { - GLuint op0 = curProg->Instructions[j][i].Opcode[0]; - GLuint op1 = curProg->Instructions[j][i].Opcode[1]; - const char *op0_enum = op0 > 5 ? _mesa_lookup_enum_by_nr(op0) : "0"; - const char *op1_enum = op1 > 5 ? _mesa_lookup_enum_by_nr(op1) : "0"; - GLuint count0 = curProg->Instructions[j][i].ArgCount[0]; - GLuint count1 = curProg->Instructions[j][i].ArgCount[1]; - fprintf(stderr, "%2d %04X %s %d %04X %s %d\n", i, op0, op0_enum, count0, - op1, op1_enum, count1); - } - } -#endif - - if (!ctx->Driver.ProgramStringNotify(ctx, GL_FRAGMENT_SHADER_ATI, NULL)) { - ctx->ATIFragmentShader.Current->isValid = GL_FALSE; - /* XXX is this the right error? */ - _mesa_error(ctx, GL_INVALID_OPERATION, - "glEndFragmentShaderATI(driver rejected shader)"); - } -} - -void GLAPIENTRY -_mesa_PassTexCoordATI(GLuint dst, GLuint coord, GLenum swizzle) -{ - GET_CURRENT_CONTEXT(ctx); - struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; - struct atifs_setupinst *curI; - - if (!ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(outsideShader)"); - return; - } - - if (curProg->cur_pass == 1) { - match_pair_inst(curProg, 0); - curProg->cur_pass = 2; - } - if ((curProg->cur_pass > 2) || - ((1 << (dst - GL_REG_0_ATI)) & curProg->regsAssigned[curProg->cur_pass >> 1])) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoord(pass)"); - return; - } - if ((dst < GL_REG_0_ATI) || (dst > GL_REG_5_ATI) || - ((dst - GL_REG_0_ATI) >= ctx->Const.MaxTextureUnits)) { - _mesa_error(ctx, GL_INVALID_ENUM, "glPassTexCoordATI(dst)"); - return; - } - if (((coord < GL_REG_0_ATI) || (coord > GL_REG_5_ATI)) && - ((coord < GL_TEXTURE0_ARB) || (coord > GL_TEXTURE7_ARB) || - ((coord - GL_TEXTURE0_ARB) >= ctx->Const.MaxTextureUnits))) { - _mesa_error(ctx, GL_INVALID_ENUM, "glPassTexCoordATI(coord)"); - return; - } - if ((curProg->cur_pass == 0) && (coord >= GL_REG_0_ATI)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(coord)"); - return; - } - if (!(swizzle >= GL_SWIZZLE_STR_ATI) && (swizzle <= GL_SWIZZLE_STQ_DQ_ATI)) { - _mesa_error(ctx, GL_INVALID_ENUM, "glPassTexCoordATI(swizzle)"); - return; - } - if ((swizzle & 1) && (coord >= GL_REG_0_ATI)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(swizzle)"); - return; - } - if (coord <= GL_TEXTURE7_ARB) { - GLuint tmp = coord - GL_TEXTURE0_ARB; - if ((((curProg->swizzlerq >> (tmp * 2)) & 3) != 0) && - (((swizzle & 1) + 1) != ((curProg->swizzlerq >> (tmp * 2)) & 3))) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(swizzle)"); - return; - } else { - curProg->swizzlerq |= (((swizzle & 1) + 1) << (tmp * 2)); - } - } - - curProg->regsAssigned[curProg->cur_pass >> 1] |= 1 << (dst - GL_REG_0_ATI); - new_tex_inst(curProg); - - /* add the instructions */ - curI = &curProg->SetupInst[curProg->cur_pass >> 1][dst - GL_REG_0_ATI]; - - curI->Opcode = ATI_FRAGMENT_SHADER_PASS_OP; - curI->src = coord; - curI->swizzle = swizzle; - -#if MESA_DEBUG_ATI_FS - _mesa_debug(ctx, "%s(%s, %s, %s)\n", __FUNCTION__, - _mesa_lookup_enum_by_nr(dst), _mesa_lookup_enum_by_nr(coord), - _mesa_lookup_enum_by_nr(swizzle)); -#endif -} - -void GLAPIENTRY -_mesa_SampleMapATI(GLuint dst, GLuint interp, GLenum swizzle) -{ - GET_CURRENT_CONTEXT(ctx); - struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; - struct atifs_setupinst *curI; - - if (!ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(outsideShader)"); - return; - } - - if (curProg->cur_pass == 1) { - match_pair_inst(curProg, 0); - curProg->cur_pass = 2; - } - if ((curProg->cur_pass > 2) || - ((1 << (dst - GL_REG_0_ATI)) & curProg->regsAssigned[curProg->cur_pass >> 1])) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(pass)"); - return; - } - if ((dst < GL_REG_0_ATI) || (dst > GL_REG_5_ATI) || - ((dst - GL_REG_0_ATI) >= ctx->Const.MaxTextureUnits)) { - _mesa_error(ctx, GL_INVALID_ENUM, "glSampleMapATI(dst)"); - return; - } - if (((interp < GL_REG_0_ATI) || (interp > GL_REG_5_ATI)) && - ((interp < GL_TEXTURE0_ARB) || (interp > GL_TEXTURE7_ARB) || - ((interp - GL_TEXTURE0_ARB) >= ctx->Const.MaxTextureUnits))) { - /* is this texture5 or texture7? spec is a bit unclear there */ - _mesa_error(ctx, GL_INVALID_ENUM, "glSampleMapATI(interp)"); - return; - } - if ((curProg->cur_pass == 0) && (interp >= GL_REG_0_ATI)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(interp)"); - return; - } - if (!(swizzle >= GL_SWIZZLE_STR_ATI) && (swizzle <= GL_SWIZZLE_STQ_DQ_ATI)) { - _mesa_error(ctx, GL_INVALID_ENUM, "glSampleMapATI(swizzle)"); - return; - } - if ((swizzle & 1) && (interp >= GL_REG_0_ATI)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(swizzle)"); - return; - } - if (interp <= GL_TEXTURE7_ARB) { - GLuint tmp = interp - GL_TEXTURE0_ARB; - if ((((curProg->swizzlerq >> (tmp * 2)) & 3) != 0) && - (((swizzle & 1) + 1) != ((curProg->swizzlerq >> (tmp * 2)) & 3))) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(swizzle)"); - return; - } else { - curProg->swizzlerq |= (((swizzle & 1) + 1) << (tmp * 2)); - } - } - - curProg->regsAssigned[curProg->cur_pass >> 1] |= 1 << (dst - GL_REG_0_ATI); - new_tex_inst(curProg); - - /* add the instructions */ - curI = &curProg->SetupInst[curProg->cur_pass >> 1][dst - GL_REG_0_ATI]; - - curI->Opcode = ATI_FRAGMENT_SHADER_SAMPLE_OP; - curI->src = interp; - curI->swizzle = swizzle; - -#if MESA_DEBUG_ATI_FS - _mesa_debug(ctx, "%s(%s, %s, %s)\n", __FUNCTION__, - _mesa_lookup_enum_by_nr(dst), _mesa_lookup_enum_by_nr(interp), - _mesa_lookup_enum_by_nr(swizzle)); -#endif -} - -static void -_mesa_FragmentOpXATI(GLint optype, GLuint arg_count, GLenum op, GLuint dst, - GLuint dstMask, GLuint dstMod, GLuint arg1, - GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, - GLuint arg2Rep, GLuint arg2Mod, GLuint arg3, - GLuint arg3Rep, GLuint arg3Mod) -{ - GET_CURRENT_CONTEXT(ctx); - struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; - GLint ci; - struct atifs_instruction *curI; - GLuint modtemp = dstMod & ~GL_SATURATE_BIT_ATI; - - if (!ctx->ATIFragmentShader.Compiling) { - _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(outsideShader)"); - return; - } - - if (curProg->cur_pass==0) - curProg->cur_pass=1; - - else if (curProg->cur_pass==2) - curProg->cur_pass=3; - - /* decide whether this is a new instruction or not ... all color instructions are new, - and alpha instructions might also be new if there was no preceding color inst */ - if ((optype == 0) || (curProg->last_optype == optype)) { - if (curProg->numArithInstr[curProg->cur_pass >> 1] > 7) { - _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(instrCount)"); - return; - } - /* easier to do that here slight side effect invalid instr will still be inserted as nops */ - match_pair_inst(curProg, optype); - new_arith_inst(curProg); - } - curProg->last_optype = optype; - ci = curProg->numArithInstr[curProg->cur_pass >> 1] - 1; - - /* add the instructions */ - curI = &curProg->Instructions[curProg->cur_pass >> 1][ci]; - - /* error checking */ - if ((dst < GL_REG_0_ATI) || (dst > GL_REG_5_ATI)) { - _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(dst)"); - return; - } - if ((modtemp != GL_NONE) && (modtemp != GL_2X_BIT_ATI) && - (modtemp != GL_4X_BIT_ATI) && (modtemp != GL_8X_BIT_ATI) && - (modtemp != GL_HALF_BIT_ATI) && !(modtemp != GL_QUARTER_BIT_ATI) && - (modtemp != GL_EIGHTH_BIT_ATI)) { - _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(dstMod)%x", modtemp); - return; - } - /* op checking? Actually looks like that's missing in the spec but we'll do it anyway */ - if (((op < GL_ADD_ATI) || (op > GL_DOT2_ADD_ATI)) && !(op == GL_MOV_ATI)) { - _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(op)"); - return; - } - if (optype == 1) { - if (((op == GL_DOT2_ADD_ATI) && (curI->Opcode[0] != GL_DOT2_ADD_ATI)) || - ((op == GL_DOT3_ATI) && (curI->Opcode[0] != GL_DOT3_ATI)) || - ((op == GL_DOT4_ATI) && (curI->Opcode[0] != GL_DOT4_ATI)) || - ((op != GL_DOT4_ATI) && (curI->Opcode[0] == GL_DOT4_ATI))) { - _mesa_error(ctx, GL_INVALID_OPERATION, "AFragmentOpATI(op)"); - return; - } - } - if ((op == GL_DOT4_ATI) && - (((arg1 == GL_SECONDARY_INTERPOLATOR_ATI) && ((arg1Rep == GL_ALPHA) || (arg1Rep == GL_NONE))) || - (((arg2 == GL_SECONDARY_INTERPOLATOR_ATI) && ((arg2Rep == GL_ALPHA) || (arg2Rep == GL_NONE)))))) { - _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(sec_interp)"); - } - - if (!check_arith_arg(curProg, optype, arg1, arg1Rep)) { - return; - } - if (arg2) { - if (!check_arith_arg(curProg, optype, arg2, arg2Rep)) { - return; - } - } - if (arg3) { - if (!check_arith_arg(curProg, optype, arg3, arg3Rep)) { - return; - } - if ((arg1 >= GL_CON_0_ATI) && (arg1 <= GL_CON_7_ATI) && - (arg2 >= GL_CON_0_ATI) && (arg2 <= GL_CON_7_ATI) && - (arg3 >= GL_CON_0_ATI) && (arg3 <= GL_CON_7_ATI) && - (arg1 != arg2) && (arg1 != arg3) && (arg2 != arg3)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(3Consts)"); - return; - } - } - - /* all ok - not all fully validated though (e.g. argNMod - spec doesn't say anything) */ - - curI->Opcode[optype] = op; - curI->SrcReg[optype][0].Index = arg1; - curI->SrcReg[optype][0].argRep = arg1Rep; - curI->SrcReg[optype][0].argMod = arg1Mod; - curI->ArgCount[optype] = arg_count; - - if (arg2) { - curI->SrcReg[optype][1].Index = arg2; - curI->SrcReg[optype][1].argRep = arg2Rep; - curI->SrcReg[optype][1].argMod = arg2Mod; - } - - if (arg3) { - curI->SrcReg[optype][2].Index = arg3; - curI->SrcReg[optype][2].argRep = arg3Rep; - curI->SrcReg[optype][2].argMod = arg3Mod; - } - - curI->DstReg[optype].Index = dst; - curI->DstReg[optype].dstMod = dstMod; - curI->DstReg[optype].dstMask = dstMask; - -#if MESA_DEBUG_ATI_FS - debug_op(optype, arg_count, op, dst, dstMask, dstMod, arg1, arg1Rep, arg1Mod, arg2, arg2Rep, arg2Mod, arg3, arg3Rep, arg3Mod); -#endif - -} - -void GLAPIENTRY -_mesa_ColorFragmentOp1ATI(GLenum op, GLuint dst, GLuint dstMask, - GLuint dstMod, GLuint arg1, GLuint arg1Rep, - GLuint arg1Mod) -{ - _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_COLOR_OP, 1, op, dst, dstMask, - dstMod, arg1, arg1Rep, arg1Mod, 0, 0, 0, 0, 0, 0); -} - -void GLAPIENTRY -_mesa_ColorFragmentOp2ATI(GLenum op, GLuint dst, GLuint dstMask, - GLuint dstMod, GLuint arg1, GLuint arg1Rep, - GLuint arg1Mod, GLuint arg2, GLuint arg2Rep, - GLuint arg2Mod) -{ - _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_COLOR_OP, 2, op, dst, dstMask, - dstMod, arg1, arg1Rep, arg1Mod, arg2, arg2Rep, - arg2Mod, 0, 0, 0); -} - -void GLAPIENTRY -_mesa_ColorFragmentOp3ATI(GLenum op, GLuint dst, GLuint dstMask, - GLuint dstMod, GLuint arg1, GLuint arg1Rep, - GLuint arg1Mod, GLuint arg2, GLuint arg2Rep, - GLuint arg2Mod, GLuint arg3, GLuint arg3Rep, - GLuint arg3Mod) -{ - _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_COLOR_OP, 3, op, dst, dstMask, - dstMod, arg1, arg1Rep, arg1Mod, arg2, arg2Rep, - arg2Mod, arg3, arg3Rep, arg3Mod); -} - -void GLAPIENTRY -_mesa_AlphaFragmentOp1ATI(GLenum op, GLuint dst, GLuint dstMod, GLuint arg1, - GLuint arg1Rep, GLuint arg1Mod) -{ - _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_ALPHA_OP, 1, op, dst, 0, dstMod, - arg1, arg1Rep, arg1Mod, 0, 0, 0, 0, 0, 0); -} - -void GLAPIENTRY -_mesa_AlphaFragmentOp2ATI(GLenum op, GLuint dst, GLuint dstMod, GLuint arg1, - GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, - GLuint arg2Rep, GLuint arg2Mod) -{ - _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_ALPHA_OP, 2, op, dst, 0, dstMod, - arg1, arg1Rep, arg1Mod, arg2, arg2Rep, arg2Mod, 0, 0, - 0); -} - -void GLAPIENTRY -_mesa_AlphaFragmentOp3ATI(GLenum op, GLuint dst, GLuint dstMod, GLuint arg1, - GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, - GLuint arg2Rep, GLuint arg2Mod, GLuint arg3, - GLuint arg3Rep, GLuint arg3Mod) -{ - _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_ALPHA_OP, 3, op, dst, 0, dstMod, - arg1, arg1Rep, arg1Mod, arg2, arg2Rep, arg2Mod, arg3, - arg3Rep, arg3Mod); -} - -void GLAPIENTRY -_mesa_SetFragmentShaderConstantATI(GLuint dst, const GLfloat * value) -{ - GLuint dstindex; - GET_CURRENT_CONTEXT(ctx); - - if ((dst < GL_CON_0_ATI) || (dst > GL_CON_7_ATI)) { - /* spec says nothing about what should happen here but we can't just segfault...*/ - _mesa_error(ctx, GL_INVALID_ENUM, "glSetFragmentShaderConstantATI(dst)"); - return; - } - - dstindex = dst - GL_CON_0_ATI; - if (ctx->ATIFragmentShader.Compiling) { - struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; - COPY_4V(curProg->Constants[dstindex], value); - curProg->LocalConstDef |= 1 << dstindex; - } - else { - FLUSH_VERTICES(ctx, _NEW_PROGRAM); - COPY_4V(ctx->ATIFragmentShader.GlobalConstants[dstindex], value); - } -} - -#endif /* FEATURE_ATI_fragment_shader */ +/** + * \file atifragshader.c + * \author David Airlie + * Copyright (C) 2004 David Airlie All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * DAVID AIRLIE BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "main/hash.h" +#include "main/imports.h" +#include "main/macros.h" +#include "main/mfeatures.h" +#include "main/enums.h" +#include "main/mtypes.h" +#include "main/dispatch.h" +#include "main/atifragshader.h" + +#if FEATURE_ATI_fragment_shader + +#define MESA_DEBUG_ATI_FS 0 + +static struct ati_fragment_shader DummyShader; + + +void +_mesa_init_ati_fragment_shader_dispatch(struct _glapi_table *disp) +{ + SET_GenFragmentShadersATI(disp, _mesa_GenFragmentShadersATI); + SET_BindFragmentShaderATI(disp, _mesa_BindFragmentShaderATI); + SET_DeleteFragmentShaderATI(disp, _mesa_DeleteFragmentShaderATI); + SET_BeginFragmentShaderATI(disp, _mesa_BeginFragmentShaderATI); + SET_EndFragmentShaderATI(disp, _mesa_EndFragmentShaderATI); + SET_PassTexCoordATI(disp, _mesa_PassTexCoordATI); + SET_SampleMapATI(disp, _mesa_SampleMapATI); + SET_ColorFragmentOp1ATI(disp, _mesa_ColorFragmentOp1ATI); + SET_ColorFragmentOp2ATI(disp, _mesa_ColorFragmentOp2ATI); + SET_ColorFragmentOp3ATI(disp, _mesa_ColorFragmentOp3ATI); + SET_AlphaFragmentOp1ATI(disp, _mesa_AlphaFragmentOp1ATI); + SET_AlphaFragmentOp2ATI(disp, _mesa_AlphaFragmentOp2ATI); + SET_AlphaFragmentOp3ATI(disp, _mesa_AlphaFragmentOp3ATI); + SET_SetFragmentShaderConstantATI(disp, _mesa_SetFragmentShaderConstantATI); +} + + +/** + * Allocate and initialize a new ATI fragment shader object. + */ +struct ati_fragment_shader * +_mesa_new_ati_fragment_shader(struct gl_context *ctx, GLuint id) +{ + struct ati_fragment_shader *s = CALLOC_STRUCT(ati_fragment_shader); + (void) ctx; + if (s) { + s->Id = id; + s->RefCount = 1; + } + return s; +} + + +/** + * Delete the given ati fragment shader + */ +void +_mesa_delete_ati_fragment_shader(struct gl_context *ctx, struct ati_fragment_shader *s) +{ + GLuint i; + for (i = 0; i < MAX_NUM_PASSES_ATI; i++) { + if (s->Instructions[i]) + free(s->Instructions[i]); + if (s->SetupInst[i]) + free(s->SetupInst[i]); + } + free(s); +} + + + +static void +new_arith_inst(struct ati_fragment_shader *prog) +{ +/* set "default" instruction as not all may get defined. + there is no specified way to express a nop with ati fragment shaders we use + GL_NONE as the op enum and just set some params to 0 - so nothing to do here */ + prog->numArithInstr[prog->cur_pass >> 1]++; +} + +static void +new_tex_inst(struct ati_fragment_shader *prog) +{ +} + +static void match_pair_inst(struct ati_fragment_shader *curProg, GLuint optype) +{ + if (optype == curProg->last_optype) { + curProg->last_optype = 1; + } +} + +#if MESA_DEBUG_ATI_FS +static char * +create_dst_mod_str(GLuint mod) +{ + static char ret_str[1024]; + + memset(ret_str, 0, 1024); + if (mod & GL_2X_BIT_ATI) + strncat(ret_str, "|2X", 1024); + + if (mod & GL_4X_BIT_ATI) + strncat(ret_str, "|4X", 1024); + + if (mod & GL_8X_BIT_ATI) + strncat(ret_str, "|8X", 1024); + if (mod & GL_HALF_BIT_ATI) + strncat(ret_str, "|HA", 1024); + if (mod & GL_QUARTER_BIT_ATI) + strncat(ret_str, "|QU", 1024); + if (mod & GL_EIGHTH_BIT_ATI) + strncat(ret_str, "|EI", 1024); + + if (mod & GL_SATURATE_BIT_ATI) + strncat(ret_str, "|SAT", 1024); + + if (strlen(ret_str) == 0) + strncat(ret_str, "NONE", 1024); + return ret_str; +} + +static char *atifs_ops[] = {"ColorFragmentOp1ATI", "ColorFragmentOp2ATI", "ColorFragmentOp3ATI", + "AlphaFragmentOp1ATI", "AlphaFragmentOp2ATI", "AlphaFragmentOp3ATI" }; + +static void debug_op(GLint optype, GLuint arg_count, GLenum op, GLuint dst, + GLuint dstMask, GLuint dstMod, GLuint arg1, + GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, + GLuint arg2Rep, GLuint arg2Mod, GLuint arg3, + GLuint arg3Rep, GLuint arg3Mod) +{ + char *op_name; + + op_name = atifs_ops[(arg_count-1)+(optype?3:0)]; + + fprintf(stderr, "%s(%s, %s", op_name, _mesa_lookup_enum_by_nr(op), + _mesa_lookup_enum_by_nr(dst)); + if (!optype) + fprintf(stderr, ", %d", dstMask); + + fprintf(stderr, ", %s", create_dst_mod_str(dstMod)); + + fprintf(stderr, ", %s, %s, %d", _mesa_lookup_enum_by_nr(arg1), + _mesa_lookup_enum_by_nr(arg1Rep), arg1Mod); + if (arg_count>1) + fprintf(stderr, ", %s, %s, %d", _mesa_lookup_enum_by_nr(arg2), + _mesa_lookup_enum_by_nr(arg2Rep), arg2Mod); + if (arg_count>2) + fprintf(stderr, ", %s, %s, %d", _mesa_lookup_enum_by_nr(arg3), + _mesa_lookup_enum_by_nr(arg3Rep), arg3Mod); + + fprintf(stderr,")\n"); + +} +#endif + +static int check_arith_arg(struct ati_fragment_shader *curProg, + GLuint optype, GLuint arg, GLuint argRep) +{ + GET_CURRENT_CONTEXT(ctx); + + if (((arg < GL_CON_0_ATI) || (arg > GL_CON_7_ATI)) && + ((arg < GL_REG_0_ATI) || (arg > GL_REG_5_ATI)) && + (arg != GL_ZERO) && (arg != GL_ONE) && + (arg != GL_PRIMARY_COLOR_ARB) && (arg != GL_SECONDARY_INTERPOLATOR_ATI)) { + _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(arg)"); + return 0; + } + if ((arg == GL_SECONDARY_INTERPOLATOR_ATI) && (((optype == 0) && (argRep == GL_ALPHA)) || + ((optype == 1) && ((arg == GL_ALPHA) || (argRep == GL_NONE))))) { + _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(sec_interp)"); + return 0; + } + if ((arg == GL_SECONDARY_INTERPOLATOR_ATI) && (((optype == 0) && (argRep == GL_ALPHA)) || + ((optype == 1) && ((arg == GL_ALPHA) || (argRep == GL_NONE))))) { + _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(sec_interp)"); + return 0; + } + if ((curProg->cur_pass == 1) && + ((arg == GL_PRIMARY_COLOR_ARB) || (arg == GL_SECONDARY_INTERPOLATOR_ATI))) { + curProg->interpinp1 = GL_TRUE; + } + return 1; +} + +GLuint GLAPIENTRY +_mesa_GenFragmentShadersATI(GLuint range) +{ + GLuint first; + GLuint i; + GET_CURRENT_CONTEXT(ctx); + + if (range == 0) { + _mesa_error(ctx, GL_INVALID_VALUE, "glGenFragmentShadersATI(range)"); + return 0; + } + + if (ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glGenFragmentShadersATI(insideShader)"); + return 0; + } + + first = _mesa_HashFindFreeKeyBlock(ctx->Shared->ATIShaders, range); + for (i = 0; i < range; i++) { + _mesa_HashInsert(ctx->Shared->ATIShaders, first + i, &DummyShader); + } + + return first; +} + +void GLAPIENTRY +_mesa_BindFragmentShaderATI(GLuint id) +{ + GET_CURRENT_CONTEXT(ctx); + struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; + struct ati_fragment_shader *newProg; + + if (ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glBindFragmentShaderATI(insideShader)"); + return; + } + + FLUSH_VERTICES(ctx, _NEW_PROGRAM); + + if (curProg->Id == id) { + return; + } + + /* unbind current */ + if (curProg->Id != 0) { + curProg->RefCount--; + if (curProg->RefCount <= 0) { + _mesa_HashRemove(ctx->Shared->ATIShaders, id); + } + } + + /* find new shader */ + if (id == 0) { + newProg = ctx->Shared->DefaultFragmentShader; + } + else { + newProg = (struct ati_fragment_shader *) + _mesa_HashLookup(ctx->Shared->ATIShaders, id); + if (!newProg || newProg == &DummyShader) { + /* allocate a new program now */ + newProg = _mesa_new_ati_fragment_shader(ctx, id); + if (!newProg) { + _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBindFragmentShaderATI"); + return; + } + _mesa_HashInsert(ctx->Shared->ATIShaders, id, newProg); + } + + } + + /* do actual bind */ + ctx->ATIFragmentShader.Current = newProg; + + ASSERT(ctx->ATIFragmentShader.Current); + if (newProg) + newProg->RefCount++; + + /*if (ctx->Driver.BindProgram) + ctx->Driver.BindProgram(ctx, target, prog); */ +} + +void GLAPIENTRY +_mesa_DeleteFragmentShaderATI(GLuint id) +{ + GET_CURRENT_CONTEXT(ctx); + + if (ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glDeleteFragmentShaderATI(insideShader)"); + return; + } + + if (id != 0) { + struct ati_fragment_shader *prog = (struct ati_fragment_shader *) + _mesa_HashLookup(ctx->Shared->ATIShaders, id); + if (prog == &DummyShader) { + _mesa_HashRemove(ctx->Shared->ATIShaders, id); + } + else if (prog) { + if (ctx->ATIFragmentShader.Current && + ctx->ATIFragmentShader.Current->Id == id) { + FLUSH_VERTICES(ctx, _NEW_PROGRAM); + _mesa_BindFragmentShaderATI(0); + } + } + + /* The ID is immediately available for re-use now */ + _mesa_HashRemove(ctx->Shared->ATIShaders, id); + if (prog) { + prog->RefCount--; + if (prog->RefCount <= 0) { + assert(prog != &DummyShader); + free(prog); + } + } + } +} + + +void GLAPIENTRY +_mesa_BeginFragmentShaderATI(void) +{ + GLint i; + GET_CURRENT_CONTEXT(ctx); + + if (ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glBeginFragmentShaderATI(insideShader)"); + return; + } + + FLUSH_VERTICES(ctx, _NEW_PROGRAM); + + /* if the shader was already defined free instructions and get new ones + (or, could use the same mem but would need to reinitialize) */ + /* no idea if it's allowed to redefine a shader */ + for (i = 0; i < MAX_NUM_PASSES_ATI; i++) { + if (ctx->ATIFragmentShader.Current->Instructions[i]) + free(ctx->ATIFragmentShader.Current->Instructions[i]); + if (ctx->ATIFragmentShader.Current->SetupInst[i]) + free(ctx->ATIFragmentShader.Current->SetupInst[i]); + } + + /* malloc the instructions here - not sure if the best place but its + a start */ + for (i = 0; i < MAX_NUM_PASSES_ATI; i++) { + ctx->ATIFragmentShader.Current->Instructions[i] = + (struct atifs_instruction *) + calloc(1, sizeof(struct atifs_instruction) * + (MAX_NUM_INSTRUCTIONS_PER_PASS_ATI)); + ctx->ATIFragmentShader.Current->SetupInst[i] = + (struct atifs_setupinst *) + calloc(1, sizeof(struct atifs_setupinst) * + (MAX_NUM_FRAGMENT_REGISTERS_ATI)); + } + +/* can't rely on calloc for initialization as it's possible to redefine a shader (?) */ + ctx->ATIFragmentShader.Current->LocalConstDef = 0; + ctx->ATIFragmentShader.Current->numArithInstr[0] = 0; + ctx->ATIFragmentShader.Current->numArithInstr[1] = 0; + ctx->ATIFragmentShader.Current->regsAssigned[0] = 0; + ctx->ATIFragmentShader.Current->regsAssigned[1] = 0; + ctx->ATIFragmentShader.Current->NumPasses = 0; + ctx->ATIFragmentShader.Current->cur_pass = 0; + ctx->ATIFragmentShader.Current->last_optype = 0; + ctx->ATIFragmentShader.Current->interpinp1 = GL_FALSE; + ctx->ATIFragmentShader.Current->isValid = GL_FALSE; + ctx->ATIFragmentShader.Current->swizzlerq = 0; + ctx->ATIFragmentShader.Compiling = 1; +} + +void GLAPIENTRY +_mesa_EndFragmentShaderATI(void) +{ + GET_CURRENT_CONTEXT(ctx); + struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; +#if MESA_DEBUG_ATI_FS + GLint i, j; +#endif + + if (!ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glEndFragmentShaderATI(outsideShader)"); + return; + } + if (curProg->interpinp1 && (ctx->ATIFragmentShader.Current->cur_pass > 1)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glEndFragmentShaderATI(interpinfirstpass)"); + /* according to spec, DON'T return here */ + } + + match_pair_inst(curProg, 0); + ctx->ATIFragmentShader.Compiling = 0; + ctx->ATIFragmentShader.Current->isValid = GL_TRUE; + if ((ctx->ATIFragmentShader.Current->cur_pass == 0) || + (ctx->ATIFragmentShader.Current->cur_pass == 2)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glEndFragmentShaderATI(noarithinst)"); + } + if (ctx->ATIFragmentShader.Current->cur_pass > 1) + ctx->ATIFragmentShader.Current->NumPasses = 2; + else + ctx->ATIFragmentShader.Current->NumPasses = 1; + + ctx->ATIFragmentShader.Current->cur_pass = 0; + +#if MESA_DEBUG_ATI_FS + for (j = 0; j < MAX_NUM_PASSES_ATI; j++) { + for (i = 0; i < MAX_NUM_FRAGMENT_REGISTERS_ATI; i++) { + GLuint op = curProg->SetupInst[j][i].Opcode; + const char *op_enum = op > 5 ? _mesa_lookup_enum_by_nr(op) : "0"; + GLuint src = curProg->SetupInst[j][i].src; + GLuint swizzle = curProg->SetupInst[j][i].swizzle; + fprintf(stderr, "%2d %04X %s %d %04X\n", i, op, op_enum, src, + swizzle); + } + for (i = 0; i < curProg->numArithInstr[j]; i++) { + GLuint op0 = curProg->Instructions[j][i].Opcode[0]; + GLuint op1 = curProg->Instructions[j][i].Opcode[1]; + const char *op0_enum = op0 > 5 ? _mesa_lookup_enum_by_nr(op0) : "0"; + const char *op1_enum = op1 > 5 ? _mesa_lookup_enum_by_nr(op1) : "0"; + GLuint count0 = curProg->Instructions[j][i].ArgCount[0]; + GLuint count1 = curProg->Instructions[j][i].ArgCount[1]; + fprintf(stderr, "%2d %04X %s %d %04X %s %d\n", i, op0, op0_enum, count0, + op1, op1_enum, count1); + } + } +#endif + + if (!ctx->Driver.ProgramStringNotify(ctx, GL_FRAGMENT_SHADER_ATI, NULL)) { + ctx->ATIFragmentShader.Current->isValid = GL_FALSE; + /* XXX is this the right error? */ + _mesa_error(ctx, GL_INVALID_OPERATION, + "glEndFragmentShaderATI(driver rejected shader)"); + } +} + +void GLAPIENTRY +_mesa_PassTexCoordATI(GLuint dst, GLuint coord, GLenum swizzle) +{ + GET_CURRENT_CONTEXT(ctx); + struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; + struct atifs_setupinst *curI; + + if (!ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(outsideShader)"); + return; + } + + if (curProg->cur_pass == 1) { + match_pair_inst(curProg, 0); + curProg->cur_pass = 2; + } + if ((curProg->cur_pass > 2) || + ((1 << (dst - GL_REG_0_ATI)) & curProg->regsAssigned[curProg->cur_pass >> 1])) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoord(pass)"); + return; + } + if ((dst < GL_REG_0_ATI) || (dst > GL_REG_5_ATI) || + ((dst - GL_REG_0_ATI) >= ctx->Const.MaxTextureUnits)) { + _mesa_error(ctx, GL_INVALID_ENUM, "glPassTexCoordATI(dst)"); + return; + } + if (((coord < GL_REG_0_ATI) || (coord > GL_REG_5_ATI)) && + ((coord < GL_TEXTURE0_ARB) || (coord > GL_TEXTURE7_ARB) || + ((coord - GL_TEXTURE0_ARB) >= ctx->Const.MaxTextureUnits))) { + _mesa_error(ctx, GL_INVALID_ENUM, "glPassTexCoordATI(coord)"); + return; + } + if ((curProg->cur_pass == 0) && (coord >= GL_REG_0_ATI)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(coord)"); + return; + } + if (!(swizzle >= GL_SWIZZLE_STR_ATI) && (swizzle <= GL_SWIZZLE_STQ_DQ_ATI)) { + _mesa_error(ctx, GL_INVALID_ENUM, "glPassTexCoordATI(swizzle)"); + return; + } + if ((swizzle & 1) && (coord >= GL_REG_0_ATI)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(swizzle)"); + return; + } + if (coord <= GL_TEXTURE7_ARB) { + GLuint tmp = coord - GL_TEXTURE0_ARB; + if ((((curProg->swizzlerq >> (tmp * 2)) & 3) != 0) && + (((swizzle & 1) + 1) != ((curProg->swizzlerq >> (tmp * 2)) & 3))) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glPassTexCoordATI(swizzle)"); + return; + } else { + curProg->swizzlerq |= (((swizzle & 1) + 1) << (tmp * 2)); + } + } + + curProg->regsAssigned[curProg->cur_pass >> 1] |= 1 << (dst - GL_REG_0_ATI); + new_tex_inst(curProg); + + /* add the instructions */ + curI = &curProg->SetupInst[curProg->cur_pass >> 1][dst - GL_REG_0_ATI]; + + curI->Opcode = ATI_FRAGMENT_SHADER_PASS_OP; + curI->src = coord; + curI->swizzle = swizzle; + +#if MESA_DEBUG_ATI_FS + _mesa_debug(ctx, "%s(%s, %s, %s)\n", __FUNCTION__, + _mesa_lookup_enum_by_nr(dst), _mesa_lookup_enum_by_nr(coord), + _mesa_lookup_enum_by_nr(swizzle)); +#endif +} + +void GLAPIENTRY +_mesa_SampleMapATI(GLuint dst, GLuint interp, GLenum swizzle) +{ + GET_CURRENT_CONTEXT(ctx); + struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; + struct atifs_setupinst *curI; + + if (!ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(outsideShader)"); + return; + } + + if (curProg->cur_pass == 1) { + match_pair_inst(curProg, 0); + curProg->cur_pass = 2; + } + if ((curProg->cur_pass > 2) || + ((1 << (dst - GL_REG_0_ATI)) & curProg->regsAssigned[curProg->cur_pass >> 1])) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(pass)"); + return; + } + if ((dst < GL_REG_0_ATI) || (dst > GL_REG_5_ATI) || + ((dst - GL_REG_0_ATI) >= ctx->Const.MaxTextureUnits)) { + _mesa_error(ctx, GL_INVALID_ENUM, "glSampleMapATI(dst)"); + return; + } + if (((interp < GL_REG_0_ATI) || (interp > GL_REG_5_ATI)) && + ((interp < GL_TEXTURE0_ARB) || (interp > GL_TEXTURE7_ARB) || + ((interp - GL_TEXTURE0_ARB) >= ctx->Const.MaxTextureUnits))) { + /* is this texture5 or texture7? spec is a bit unclear there */ + _mesa_error(ctx, GL_INVALID_ENUM, "glSampleMapATI(interp)"); + return; + } + if ((curProg->cur_pass == 0) && (interp >= GL_REG_0_ATI)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(interp)"); + return; + } + if (!(swizzle >= GL_SWIZZLE_STR_ATI) && (swizzle <= GL_SWIZZLE_STQ_DQ_ATI)) { + _mesa_error(ctx, GL_INVALID_ENUM, "glSampleMapATI(swizzle)"); + return; + } + if ((swizzle & 1) && (interp >= GL_REG_0_ATI)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(swizzle)"); + return; + } + if (interp <= GL_TEXTURE7_ARB) { + GLuint tmp = interp - GL_TEXTURE0_ARB; + if ((((curProg->swizzlerq >> (tmp * 2)) & 3) != 0) && + (((swizzle & 1) + 1) != ((curProg->swizzlerq >> (tmp * 2)) & 3))) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glSampleMapATI(swizzle)"); + return; + } else { + curProg->swizzlerq |= (((swizzle & 1) + 1) << (tmp * 2)); + } + } + + curProg->regsAssigned[curProg->cur_pass >> 1] |= 1 << (dst - GL_REG_0_ATI); + new_tex_inst(curProg); + + /* add the instructions */ + curI = &curProg->SetupInst[curProg->cur_pass >> 1][dst - GL_REG_0_ATI]; + + curI->Opcode = ATI_FRAGMENT_SHADER_SAMPLE_OP; + curI->src = interp; + curI->swizzle = swizzle; + +#if MESA_DEBUG_ATI_FS + _mesa_debug(ctx, "%s(%s, %s, %s)\n", __FUNCTION__, + _mesa_lookup_enum_by_nr(dst), _mesa_lookup_enum_by_nr(interp), + _mesa_lookup_enum_by_nr(swizzle)); +#endif +} + +static void +_mesa_FragmentOpXATI(GLint optype, GLuint arg_count, GLenum op, GLuint dst, + GLuint dstMask, GLuint dstMod, GLuint arg1, + GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, + GLuint arg2Rep, GLuint arg2Mod, GLuint arg3, + GLuint arg3Rep, GLuint arg3Mod) +{ + GET_CURRENT_CONTEXT(ctx); + struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; + GLint ci; + struct atifs_instruction *curI; + GLuint modtemp = dstMod & ~GL_SATURATE_BIT_ATI; + + if (!ctx->ATIFragmentShader.Compiling) { + _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(outsideShader)"); + return; + } + + if (curProg->cur_pass==0) + curProg->cur_pass=1; + + else if (curProg->cur_pass==2) + curProg->cur_pass=3; + + /* decide whether this is a new instruction or not ... all color instructions are new, + and alpha instructions might also be new if there was no preceding color inst */ + if ((optype == 0) || (curProg->last_optype == optype)) { + if (curProg->numArithInstr[curProg->cur_pass >> 1] > 7) { + _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(instrCount)"); + return; + } + /* easier to do that here slight side effect invalid instr will still be inserted as nops */ + match_pair_inst(curProg, optype); + new_arith_inst(curProg); + } + curProg->last_optype = optype; + ci = curProg->numArithInstr[curProg->cur_pass >> 1] - 1; + + /* add the instructions */ + curI = &curProg->Instructions[curProg->cur_pass >> 1][ci]; + + /* error checking */ + if ((dst < GL_REG_0_ATI) || (dst > GL_REG_5_ATI)) { + _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(dst)"); + return; + } + if ((modtemp != GL_NONE) && (modtemp != GL_2X_BIT_ATI) && + (modtemp != GL_4X_BIT_ATI) && (modtemp != GL_8X_BIT_ATI) && + (modtemp != GL_HALF_BIT_ATI) && !(modtemp != GL_QUARTER_BIT_ATI) && + (modtemp != GL_EIGHTH_BIT_ATI)) { + _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(dstMod)%x", modtemp); + return; + } + /* op checking? Actually looks like that's missing in the spec but we'll do it anyway */ + if (((op < GL_ADD_ATI) || (op > GL_DOT2_ADD_ATI)) && !(op == GL_MOV_ATI)) { + _mesa_error(ctx, GL_INVALID_ENUM, "C/AFragmentOpATI(op)"); + return; + } + if (optype == 1) { + if (((op == GL_DOT2_ADD_ATI) && (curI->Opcode[0] != GL_DOT2_ADD_ATI)) || + ((op == GL_DOT3_ATI) && (curI->Opcode[0] != GL_DOT3_ATI)) || + ((op == GL_DOT4_ATI) && (curI->Opcode[0] != GL_DOT4_ATI)) || + ((op != GL_DOT4_ATI) && (curI->Opcode[0] == GL_DOT4_ATI))) { + _mesa_error(ctx, GL_INVALID_OPERATION, "AFragmentOpATI(op)"); + return; + } + } + if ((op == GL_DOT4_ATI) && + (((arg1 == GL_SECONDARY_INTERPOLATOR_ATI) && ((arg1Rep == GL_ALPHA) || (arg1Rep == GL_NONE))) || + (((arg2 == GL_SECONDARY_INTERPOLATOR_ATI) && ((arg2Rep == GL_ALPHA) || (arg2Rep == GL_NONE)))))) { + _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(sec_interp)"); + } + + if (!check_arith_arg(curProg, optype, arg1, arg1Rep)) { + return; + } + if (arg2) { + if (!check_arith_arg(curProg, optype, arg2, arg2Rep)) { + return; + } + } + if (arg3) { + if (!check_arith_arg(curProg, optype, arg3, arg3Rep)) { + return; + } + if ((arg1 >= GL_CON_0_ATI) && (arg1 <= GL_CON_7_ATI) && + (arg2 >= GL_CON_0_ATI) && (arg2 <= GL_CON_7_ATI) && + (arg3 >= GL_CON_0_ATI) && (arg3 <= GL_CON_7_ATI) && + (arg1 != arg2) && (arg1 != arg3) && (arg2 != arg3)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "C/AFragmentOpATI(3Consts)"); + return; + } + } + + /* all ok - not all fully validated though (e.g. argNMod - spec doesn't say anything) */ + + curI->Opcode[optype] = op; + curI->SrcReg[optype][0].Index = arg1; + curI->SrcReg[optype][0].argRep = arg1Rep; + curI->SrcReg[optype][0].argMod = arg1Mod; + curI->ArgCount[optype] = arg_count; + + if (arg2) { + curI->SrcReg[optype][1].Index = arg2; + curI->SrcReg[optype][1].argRep = arg2Rep; + curI->SrcReg[optype][1].argMod = arg2Mod; + } + + if (arg3) { + curI->SrcReg[optype][2].Index = arg3; + curI->SrcReg[optype][2].argRep = arg3Rep; + curI->SrcReg[optype][2].argMod = arg3Mod; + } + + curI->DstReg[optype].Index = dst; + curI->DstReg[optype].dstMod = dstMod; + curI->DstReg[optype].dstMask = dstMask; + +#if MESA_DEBUG_ATI_FS + debug_op(optype, arg_count, op, dst, dstMask, dstMod, arg1, arg1Rep, arg1Mod, arg2, arg2Rep, arg2Mod, arg3, arg3Rep, arg3Mod); +#endif + +} + +void GLAPIENTRY +_mesa_ColorFragmentOp1ATI(GLenum op, GLuint dst, GLuint dstMask, + GLuint dstMod, GLuint arg1, GLuint arg1Rep, + GLuint arg1Mod) +{ + _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_COLOR_OP, 1, op, dst, dstMask, + dstMod, arg1, arg1Rep, arg1Mod, 0, 0, 0, 0, 0, 0); +} + +void GLAPIENTRY +_mesa_ColorFragmentOp2ATI(GLenum op, GLuint dst, GLuint dstMask, + GLuint dstMod, GLuint arg1, GLuint arg1Rep, + GLuint arg1Mod, GLuint arg2, GLuint arg2Rep, + GLuint arg2Mod) +{ + _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_COLOR_OP, 2, op, dst, dstMask, + dstMod, arg1, arg1Rep, arg1Mod, arg2, arg2Rep, + arg2Mod, 0, 0, 0); +} + +void GLAPIENTRY +_mesa_ColorFragmentOp3ATI(GLenum op, GLuint dst, GLuint dstMask, + GLuint dstMod, GLuint arg1, GLuint arg1Rep, + GLuint arg1Mod, GLuint arg2, GLuint arg2Rep, + GLuint arg2Mod, GLuint arg3, GLuint arg3Rep, + GLuint arg3Mod) +{ + _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_COLOR_OP, 3, op, dst, dstMask, + dstMod, arg1, arg1Rep, arg1Mod, arg2, arg2Rep, + arg2Mod, arg3, arg3Rep, arg3Mod); +} + +void GLAPIENTRY +_mesa_AlphaFragmentOp1ATI(GLenum op, GLuint dst, GLuint dstMod, GLuint arg1, + GLuint arg1Rep, GLuint arg1Mod) +{ + _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_ALPHA_OP, 1, op, dst, 0, dstMod, + arg1, arg1Rep, arg1Mod, 0, 0, 0, 0, 0, 0); +} + +void GLAPIENTRY +_mesa_AlphaFragmentOp2ATI(GLenum op, GLuint dst, GLuint dstMod, GLuint arg1, + GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, + GLuint arg2Rep, GLuint arg2Mod) +{ + _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_ALPHA_OP, 2, op, dst, 0, dstMod, + arg1, arg1Rep, arg1Mod, arg2, arg2Rep, arg2Mod, 0, 0, + 0); +} + +void GLAPIENTRY +_mesa_AlphaFragmentOp3ATI(GLenum op, GLuint dst, GLuint dstMod, GLuint arg1, + GLuint arg1Rep, GLuint arg1Mod, GLuint arg2, + GLuint arg2Rep, GLuint arg2Mod, GLuint arg3, + GLuint arg3Rep, GLuint arg3Mod) +{ + _mesa_FragmentOpXATI(ATI_FRAGMENT_SHADER_ALPHA_OP, 3, op, dst, 0, dstMod, + arg1, arg1Rep, arg1Mod, arg2, arg2Rep, arg2Mod, arg3, + arg3Rep, arg3Mod); +} + +void GLAPIENTRY +_mesa_SetFragmentShaderConstantATI(GLuint dst, const GLfloat * value) +{ + GLuint dstindex; + GET_CURRENT_CONTEXT(ctx); + + if ((dst < GL_CON_0_ATI) || (dst > GL_CON_7_ATI)) { + /* spec says nothing about what should happen here but we can't just segfault...*/ + _mesa_error(ctx, GL_INVALID_ENUM, "glSetFragmentShaderConstantATI(dst)"); + return; + } + + dstindex = dst - GL_CON_0_ATI; + if (ctx->ATIFragmentShader.Compiling) { + struct ati_fragment_shader *curProg = ctx->ATIFragmentShader.Current; + COPY_4V(curProg->Constants[dstindex], value); + curProg->LocalConstDef |= 1 << dstindex; + } + else { + FLUSH_VERTICES(ctx, _NEW_PROGRAM); + COPY_4V(ctx->ATIFragmentShader.GlobalConstants[dstindex], value); + } +} + +#endif /* FEATURE_ATI_fragment_shader */ diff --git a/mesalib/src/mesa/main/blend.h b/mesalib/src/mesa/main/blend.h index b0a6b0135..d74f17ef1 100644 --- a/mesalib/src/mesa/main/blend.h +++ b/mesalib/src/mesa/main/blend.h @@ -1,106 +1,106 @@ -/** - * \file blend.h - * Blending functions operations. - */ - -/* - * Mesa 3-D graphics library - * Version: 6.5.2 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - - -#ifndef BLEND_H -#define BLEND_H - - -#include "glheader.h" - -struct gl_context; - - -extern void GLAPIENTRY -_mesa_BlendFunc( GLenum sfactor, GLenum dfactor ); - - -extern void GLAPIENTRY -_mesa_BlendFuncSeparateEXT( GLenum sfactorRGB, GLenum dfactorRGB, - GLenum sfactorA, GLenum dfactorA ); - - -extern void GLAPIENTRY -_mesa_BlendFunci(GLuint buf, GLenum sfactor, GLenum dfactor); - - -extern void GLAPIENTRY -_mesa_BlendFuncSeparatei(GLuint buf, GLenum sfactorRGB, GLenum dfactorRGB, - GLenum sfactorA, GLenum dfactorA); - - -extern void GLAPIENTRY -_mesa_BlendEquation( GLenum mode ); - - -extern void GLAPIENTRY -_mesa_BlendEquationi(GLuint buf, GLenum mode); - - -extern void GLAPIENTRY -_mesa_BlendEquationSeparateEXT( GLenum modeRGB, GLenum modeA ); - - -extern void GLAPIENTRY -_mesa_BlendEquationSeparatei(GLuint buf, GLenum modeRGB, GLenum modeA); - - -extern void GLAPIENTRY -_mesa_BlendColor(GLclampf red, GLclampf green, GLclampf blue, GLclampf alpha); - - -extern void GLAPIENTRY -_mesa_AlphaFunc( GLenum func, GLclampf ref ); - - -extern void GLAPIENTRY -_mesa_LogicOp( GLenum opcode ); - - -extern void GLAPIENTRY -_mesa_IndexMask( GLuint mask ); - -extern void GLAPIENTRY -_mesa_ColorMask( GLboolean red, GLboolean green, - GLboolean blue, GLboolean alpha ); - -extern void GLAPIENTRY -_mesa_ColorMaskIndexed( GLuint buf, GLboolean red, GLboolean green, - GLboolean blue, GLboolean alpha ); - - -extern void GLAPIENTRY -_mesa_ClampColorARB(GLenum target, GLenum clamp); - - -extern void -_mesa_init_color( struct gl_context * ctx ); - -#endif +/** + * \file blend.h + * Blending functions operations. + */ + +/* + * Mesa 3-D graphics library + * Version: 6.5.2 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + + +#ifndef BLEND_H +#define BLEND_H + + +#include "glheader.h" + +struct gl_context; + + +extern void GLAPIENTRY +_mesa_BlendFunc( GLenum sfactor, GLenum dfactor ); + + +extern void GLAPIENTRY +_mesa_BlendFuncSeparateEXT( GLenum sfactorRGB, GLenum dfactorRGB, + GLenum sfactorA, GLenum dfactorA ); + + +extern void GLAPIENTRY +_mesa_BlendFunci(GLuint buf, GLenum sfactor, GLenum dfactor); + + +extern void GLAPIENTRY +_mesa_BlendFuncSeparatei(GLuint buf, GLenum sfactorRGB, GLenum dfactorRGB, + GLenum sfactorA, GLenum dfactorA); + + +extern void GLAPIENTRY +_mesa_BlendEquation( GLenum mode ); + + +extern void GLAPIENTRY +_mesa_BlendEquationi(GLuint buf, GLenum mode); + + +extern void GLAPIENTRY +_mesa_BlendEquationSeparateEXT( GLenum modeRGB, GLenum modeA ); + + +extern void GLAPIENTRY +_mesa_BlendEquationSeparatei(GLuint buf, GLenum modeRGB, GLenum modeA); + + +extern void GLAPIENTRY +_mesa_BlendColor(GLclampf red, GLclampf green, GLclampf blue, GLclampf alpha); + + +extern void GLAPIENTRY +_mesa_AlphaFunc( GLenum func, GLclampf ref ); + + +extern void GLAPIENTRY +_mesa_LogicOp( GLenum opcode ); + + +extern void GLAPIENTRY +_mesa_IndexMask( GLuint mask ); + +extern void GLAPIENTRY +_mesa_ColorMask( GLboolean red, GLboolean green, + GLboolean blue, GLboolean alpha ); + +extern void GLAPIENTRY +_mesa_ColorMaskIndexed( GLuint buf, GLboolean red, GLboolean green, + GLboolean blue, GLboolean alpha ); + + +extern void GLAPIENTRY +_mesa_ClampColorARB(GLenum target, GLenum clamp); + + +extern void +_mesa_init_color( struct gl_context * ctx ); + +#endif diff --git a/mesalib/src/mesa/main/bufferobj.c b/mesalib/src/mesa/main/bufferobj.c index ae7bac14d..36a761986 100644 --- a/mesalib/src/mesa/main/bufferobj.c +++ b/mesalib/src/mesa/main/bufferobj.c @@ -42,6 +42,7 @@ #include "mfeatures.h" #include "mtypes.h" #include "texobj.h" +#include "transformfeedback.h" /* Debug flags */ @@ -821,6 +822,24 @@ _mesa_DeleteBuffersARB(GLsizei n, const GLuint *ids) _mesa_BindBufferARB( GL_ELEMENT_ARRAY_BUFFER_ARB, 0 ); } + /* unbind ARB_copy_buffer binding points */ + if (ctx->CopyReadBuffer == bufObj) { + _mesa_BindBufferARB( GL_COPY_READ_BUFFER, 0 ); + } + if (ctx->CopyWriteBuffer == bufObj) { + _mesa_BindBufferARB( GL_COPY_WRITE_BUFFER, 0 ); + } + + /* unbind transform feedback binding points */ + if (ctx->TransformFeedback.CurrentBuffer == bufObj) { + _mesa_BindBufferARB( GL_TRANSFORM_FEEDBACK_BUFFER, 0 ); + } + for (j = 0; j < MAX_FEEDBACK_ATTRIBS; j++) { + if (ctx->TransformFeedback.CurrentObject->Buffers[j] == bufObj) { + _mesa_BindBufferBase( GL_TRANSFORM_FEEDBACK_BUFFER, j, 0 ); + } + } + /* unbind any pixel pack/unpack pointers bound to this buffer */ if (ctx->Pack.BufferObj == bufObj) { _mesa_BindBufferARB( GL_PIXEL_PACK_BUFFER_EXT, 0 ); diff --git a/mesalib/src/mesa/main/clear.h b/mesalib/src/mesa/main/clear.h index d3d64fd62..783271920 100644 --- a/mesalib/src/mesa/main/clear.h +++ b/mesalib/src/mesa/main/clear.h @@ -1,64 +1,64 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef CLEAR_H -#define CLEAR_H - - -#include "glheader.h" - - -extern void GLAPIENTRY -_mesa_ClearIndex( GLfloat c ); - -extern void GLAPIENTRY -_mesa_ClearColor( GLclampf red, GLclampf green, - GLclampf blue, GLclampf alpha ); - -extern void GLAPIENTRY -_mesa_ClearColorIiEXT(GLint r, GLint g, GLint b, GLint a); - -extern void GLAPIENTRY -_mesa_ClearColorIuiEXT(GLuint r, GLuint g, GLuint b, GLuint a); - - -extern void GLAPIENTRY -_mesa_Clear( GLbitfield mask ); - - -extern void GLAPIENTRY -_mesa_ClearBufferiv(GLenum buffer, GLint drawbuffer, const GLint *value); - -extern void GLAPIENTRY -_mesa_ClearBufferuiv(GLenum buffer, GLint drawbuffer, const GLuint *value); - -extern void GLAPIENTRY -_mesa_ClearBufferfv(GLenum buffer, GLint drawbuffer, const GLfloat *value); - -extern void GLAPIENTRY -_mesa_ClearBufferfi(GLenum buffer, GLint drawbuffer, - GLfloat depth, GLint stencil); - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef CLEAR_H +#define CLEAR_H + + +#include "glheader.h" + + +extern void GLAPIENTRY +_mesa_ClearIndex( GLfloat c ); + +extern void GLAPIENTRY +_mesa_ClearColor( GLclampf red, GLclampf green, + GLclampf blue, GLclampf alpha ); + +extern void GLAPIENTRY +_mesa_ClearColorIiEXT(GLint r, GLint g, GLint b, GLint a); + +extern void GLAPIENTRY +_mesa_ClearColorIuiEXT(GLuint r, GLuint g, GLuint b, GLuint a); + + +extern void GLAPIENTRY +_mesa_Clear( GLbitfield mask ); + + +extern void GLAPIENTRY +_mesa_ClearBufferiv(GLenum buffer, GLint drawbuffer, const GLint *value); + +extern void GLAPIENTRY +_mesa_ClearBufferuiv(GLenum buffer, GLint drawbuffer, const GLuint *value); + +extern void GLAPIENTRY +_mesa_ClearBufferfv(GLenum buffer, GLint drawbuffer, const GLfloat *value); + +extern void GLAPIENTRY +_mesa_ClearBufferfi(GLenum buffer, GLint drawbuffer, + GLfloat depth, GLint stencil); + +#endif diff --git a/mesalib/src/mesa/main/clip.c b/mesalib/src/mesa/main/clip.c index 6b9b3b58c..ba2028cf2 100644 --- a/mesalib/src/mesa/main/clip.c +++ b/mesalib/src/mesa/main/clip.c @@ -1,116 +1,116 @@ -/* - * Mesa 3-D graphics library - * Version: 6.3 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#include "glheader.h" -#include "clip.h" -#include "context.h" -#include "macros.h" -#include "mtypes.h" - -#include "math/m_matrix.h" - - -/** - * Update derived clip plane state. - */ -void -_mesa_update_clip_plane(struct gl_context *ctx, GLuint plane) -{ - if (_math_matrix_is_dirty(ctx->ProjectionMatrixStack.Top)) - _math_matrix_analyse( ctx->ProjectionMatrixStack.Top ); - - /* Clip-Space Plane = Eye-Space Plane * Projection Matrix */ - _mesa_transform_vector(ctx->Transform._ClipUserPlane[plane], - ctx->Transform.EyeUserPlane[plane], - ctx->ProjectionMatrixStack.Top->inv); -} - - -void GLAPIENTRY -_mesa_ClipPlane( GLenum plane, const GLdouble *eq ) -{ - GET_CURRENT_CONTEXT(ctx); - GLint p; - GLfloat equation[4]; - ASSERT_OUTSIDE_BEGIN_END(ctx); - - p = (GLint) plane - (GLint) GL_CLIP_PLANE0; - if (p < 0 || p >= (GLint) ctx->Const.MaxClipPlanes) { - _mesa_error( ctx, GL_INVALID_ENUM, "glClipPlane" ); - return; - } - - equation[0] = (GLfloat) eq[0]; - equation[1] = (GLfloat) eq[1]; - equation[2] = (GLfloat) eq[2]; - equation[3] = (GLfloat) eq[3]; - - /* - * The equation is transformed by the transpose of the inverse of the - * current modelview matrix and stored in the resulting eye coordinates. - * - * KW: Eqn is then transformed to the current clip space, where user - * clipping now takes place. The clip-space equations are recalculated - * whenever the projection matrix changes. - */ - if (_math_matrix_is_dirty(ctx->ModelviewMatrixStack.Top)) - _math_matrix_analyse( ctx->ModelviewMatrixStack.Top ); - - _mesa_transform_vector( equation, equation, - ctx->ModelviewMatrixStack.Top->inv ); - - if (TEST_EQ_4V(ctx->Transform.EyeUserPlane[p], equation)) - return; - - FLUSH_VERTICES(ctx, _NEW_TRANSFORM); - COPY_4FV(ctx->Transform.EyeUserPlane[p], equation); - - if (ctx->Transform.ClipPlanesEnabled & (1 << p)) { - _mesa_update_clip_plane(ctx, p); - } - - if (ctx->Driver.ClipPlane) - ctx->Driver.ClipPlane( ctx, plane, equation ); -} - - -void GLAPIENTRY -_mesa_GetClipPlane( GLenum plane, GLdouble *equation ) -{ - GET_CURRENT_CONTEXT(ctx); - GLint p; - ASSERT_OUTSIDE_BEGIN_END(ctx); - - p = (GLint) (plane - GL_CLIP_PLANE0); - if (p < 0 || p >= (GLint) ctx->Const.MaxClipPlanes) { - _mesa_error( ctx, GL_INVALID_ENUM, "glGetClipPlane" ); - return; - } - - equation[0] = (GLdouble) ctx->Transform.EyeUserPlane[p][0]; - equation[1] = (GLdouble) ctx->Transform.EyeUserPlane[p][1]; - equation[2] = (GLdouble) ctx->Transform.EyeUserPlane[p][2]; - equation[3] = (GLdouble) ctx->Transform.EyeUserPlane[p][3]; -} +/* + * Mesa 3-D graphics library + * Version: 6.3 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#include "glheader.h" +#include "clip.h" +#include "context.h" +#include "macros.h" +#include "mtypes.h" + +#include "math/m_matrix.h" + + +/** + * Update derived clip plane state. + */ +void +_mesa_update_clip_plane(struct gl_context *ctx, GLuint plane) +{ + if (_math_matrix_is_dirty(ctx->ProjectionMatrixStack.Top)) + _math_matrix_analyse( ctx->ProjectionMatrixStack.Top ); + + /* Clip-Space Plane = Eye-Space Plane * Projection Matrix */ + _mesa_transform_vector(ctx->Transform._ClipUserPlane[plane], + ctx->Transform.EyeUserPlane[plane], + ctx->ProjectionMatrixStack.Top->inv); +} + + +void GLAPIENTRY +_mesa_ClipPlane( GLenum plane, const GLdouble *eq ) +{ + GET_CURRENT_CONTEXT(ctx); + GLint p; + GLfloat equation[4]; + ASSERT_OUTSIDE_BEGIN_END(ctx); + + p = (GLint) plane - (GLint) GL_CLIP_PLANE0; + if (p < 0 || p >= (GLint) ctx->Const.MaxClipPlanes) { + _mesa_error( ctx, GL_INVALID_ENUM, "glClipPlane" ); + return; + } + + equation[0] = (GLfloat) eq[0]; + equation[1] = (GLfloat) eq[1]; + equation[2] = (GLfloat) eq[2]; + equation[3] = (GLfloat) eq[3]; + + /* + * The equation is transformed by the transpose of the inverse of the + * current modelview matrix and stored in the resulting eye coordinates. + * + * KW: Eqn is then transformed to the current clip space, where user + * clipping now takes place. The clip-space equations are recalculated + * whenever the projection matrix changes. + */ + if (_math_matrix_is_dirty(ctx->ModelviewMatrixStack.Top)) + _math_matrix_analyse( ctx->ModelviewMatrixStack.Top ); + + _mesa_transform_vector( equation, equation, + ctx->ModelviewMatrixStack.Top->inv ); + + if (TEST_EQ_4V(ctx->Transform.EyeUserPlane[p], equation)) + return; + + FLUSH_VERTICES(ctx, _NEW_TRANSFORM); + COPY_4FV(ctx->Transform.EyeUserPlane[p], equation); + + if (ctx->Transform.ClipPlanesEnabled & (1 << p)) { + _mesa_update_clip_plane(ctx, p); + } + + if (ctx->Driver.ClipPlane) + ctx->Driver.ClipPlane( ctx, plane, equation ); +} + + +void GLAPIENTRY +_mesa_GetClipPlane( GLenum plane, GLdouble *equation ) +{ + GET_CURRENT_CONTEXT(ctx); + GLint p; + ASSERT_OUTSIDE_BEGIN_END(ctx); + + p = (GLint) (plane - GL_CLIP_PLANE0); + if (p < 0 || p >= (GLint) ctx->Const.MaxClipPlanes) { + _mesa_error( ctx, GL_INVALID_ENUM, "glGetClipPlane" ); + return; + } + + equation[0] = (GLdouble) ctx->Transform.EyeUserPlane[p][0]; + equation[1] = (GLdouble) ctx->Transform.EyeUserPlane[p][1]; + equation[2] = (GLdouble) ctx->Transform.EyeUserPlane[p][2]; + equation[3] = (GLdouble) ctx->Transform.EyeUserPlane[p][3]; +} diff --git a/mesalib/src/mesa/main/clip.h b/mesalib/src/mesa/main/clip.h index 9a94542ec..a8e6d7687 100644 --- a/mesalib/src/mesa/main/clip.h +++ b/mesalib/src/mesa/main/clip.h @@ -1,47 +1,47 @@ -/** - * \file clip.h - */ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - - -#ifndef CLIP_H -#define CLIP_H - -#include "glheader.h" - -struct gl_context; - -extern void -_mesa_update_clip_plane(struct gl_context *ctx, GLuint plane); - -extern void GLAPIENTRY -_mesa_ClipPlane( GLenum plane, const GLdouble *equation ); - -extern void GLAPIENTRY -_mesa_GetClipPlane( GLenum plane, GLdouble *equation ); - -#endif +/** + * \file clip.h + */ + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + + +#ifndef CLIP_H +#define CLIP_H + +#include "glheader.h" + +struct gl_context; + +extern void +_mesa_update_clip_plane(struct gl_context *ctx, GLuint plane); + +extern void GLAPIENTRY +_mesa_ClipPlane( GLenum plane, const GLdouble *equation ); + +extern void GLAPIENTRY +_mesa_GetClipPlane( GLenum plane, GLdouble *equation ); + +#endif diff --git a/mesalib/src/mesa/main/condrender.h b/mesalib/src/mesa/main/condrender.h index 49e9953be..cf6d4ca28 100644 --- a/mesalib/src/mesa/main/condrender.h +++ b/mesalib/src/mesa/main/condrender.h @@ -1,45 +1,45 @@ -/* - * Mesa 3-D graphics library - * Version: 7.8 - * - * Copyright (C) 2009 VMware, Inc. All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#ifndef CONDRENDER_H -#define CONDRENDER_H - - -#include "glheader.h" -#include "context.h" - - -extern void GLAPIENTRY -_mesa_BeginConditionalRender(GLuint queryId, GLenum mode); - -extern void APIENTRY -_mesa_EndConditionalRender(void); - -extern GLboolean -_mesa_check_conditional_render(struct gl_context *ctx); - - -#endif /* CONDRENDER_H */ +/* + * Mesa 3-D graphics library + * Version: 7.8 + * + * Copyright (C) 2009 VMware, Inc. All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ + +#ifndef CONDRENDER_H +#define CONDRENDER_H + + +#include "glheader.h" +#include "context.h" + + +extern void GLAPIENTRY +_mesa_BeginConditionalRender(GLuint queryId, GLenum mode); + +extern void APIENTRY +_mesa_EndConditionalRender(void); + +extern GLboolean +_mesa_check_conditional_render(struct gl_context *ctx); + + +#endif /* CONDRENDER_H */ diff --git a/mesalib/src/mesa/main/dd.h b/mesalib/src/mesa/main/dd.h index 582eb5d49..1582a8c81 100644 --- a/mesalib/src/mesa/main/dd.h +++ b/mesalib/src/mesa/main/dd.h @@ -198,81 +198,32 @@ struct dd_function_table { GLenum srcFormat, GLenum srcType ); /** - * Called by glTexImage1D(). Simply copy the source texture data into the - * destination texture memory. The gl_texture_image fields, etc. will be - * fully initialized. - * The parameters are the same as glTexImage1D(), plus: + * Called by glTexImage[123]D() and glCopyTexImage[12]D() + * Allocate texture memory and copy the user's image to the buffer. + * The gl_texture_image fields, etc. will be fully initialized. + * The parameters are the same as glTexImage3D(), plus: + * \param dims 1, 2, or 3 indicating glTexImage1/2/3D() * \param packing describes how to unpack the source data. * \param texImage is the destination texture image. */ - void (*TexImage1D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint border, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); + void (*TexImage)(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLint width, GLint height, GLint depth, GLint border, + GLenum format, GLenum type, const GLvoid *pixels, + const struct gl_pixelstore_attrib *packing); /** - * Called by glTexImage2D(). - * - * \sa dd_function_table::TexImage1D. + * Called by glTexSubImage[123]D(). + * Replace a subset of the target texture with new texel data. */ - void (*TexImage2D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint border, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); - - /** - * Called by glTexImage3D(). - * - * \sa dd_function_table::TexImage1D. - */ - void (*TexImage3D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint depth, GLint border, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); - - /** - * Called by glTexSubImage1D(). Replace a subset of the target texture - * with new texel data. - * \sa dd_function_table::TexImage1D. - */ - void (*TexSubImage1D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLsizei width, - GLenum format, GLenum type, - const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); - - /** - * Called by glTexSubImage2D(). - * - * \sa dd_function_table::TexSubImage1D. - */ - void (*TexSubImage2D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - GLsizei width, GLsizei height, - GLenum format, GLenum type, - const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); - - /** - * Called by glTexSubImage3D(). - * - * \sa dd_function_table::TexSubImage1D. - */ - void (*TexSubImage3D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - GLsizei width, GLsizei height, GLint depth, - GLenum format, GLenum type, - const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); + void (*TexSubImage)(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + GLsizei width, GLsizei height, GLint depth, + GLenum format, GLenum type, + const GLvoid *pixels, + const struct gl_pixelstore_attrib *packing); /** @@ -283,33 +234,14 @@ struct dd_function_table { struct gl_texture_image *texImage ); /** - * Called by glCopyTexSubImage1D() and glCopyTexImage1D(). + * Called by glCopyTex[Sub]Image[123]D(). */ - void (*CopyTexSubImage1D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, GLsizei width); - - /** - * Called by glCopyTexSubImage2D() and glCopyTexImage2D(). - */ - void (*CopyTexSubImage2D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, - GLsizei width, GLsizei height); - - /** - * Called by glCopyTexSubImage3D() and glCopyTexImage3D(). - */ - void (*CopyTexSubImage3D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, - GLsizei width, GLsizei height); + void (*CopyTexSubImage)(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + struct gl_renderbuffer *rb, + GLint x, GLint y, + GLsizei width, GLsizei height); /** * Called by glGenerateMipmap() or when GL_GENERATE_MIPMAP_SGIS is enabled. @@ -337,66 +269,24 @@ struct dd_function_table { /*@{*/ /** - * Called by glCompressedTexImage1D(). - * The parameters are the same as for glCompressedTexImage1D(), plus a - * pointer to the destination texure image. - */ - void (*CompressedTexImage1D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLsizei width, GLint border, - GLsizei imageSize, const GLvoid *data); - /** - * Called by glCompressedTexImage2D(). - * - * \sa dd_function_table::CompressedTexImage1D. - */ - void (*CompressedTexImage2D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLsizei width, GLsizei height, GLint border, - GLsizei imageSize, const GLvoid *data); - - /** - * Called by glCompressedTexImage3D(). - * - * \sa dd_function_table::CompressedTexImage3D. - */ - void (*CompressedTexImage3D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLsizei width, GLsizei height, GLsizei depth, - GLint border, - GLsizei imageSize, const GLvoid *data); - - /** - * Called by glCompressedTexSubImage1D(). + * Called by glCompressedTexImage[123]D(). */ - void (*CompressedTexSubImage1D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLsizei width, - GLenum format, - GLsizei imageSize, const GLvoid *data); + void (*CompressedTexImage)(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLsizei width, GLsizei height, GLsizei depth, + GLint border, + GLsizei imageSize, const GLvoid *data); /** - * Called by glCompressedTexSubImage2D(). + * Called by glCompressedTexSubImage[123]D(). */ - void (*CompressedTexSubImage2D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - GLsizei width, GLint height, - GLenum format, - GLsizei imageSize, const GLvoid *data); - - /** - * Called by glCompressedTexSubImage3D(). - */ - void (*CompressedTexSubImage3D)(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - GLsizei width, GLint height, GLint depth, - GLenum format, - GLsizei imageSize, const GLvoid *data); + void (*CompressedTexSubImage)(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + GLsizei width, GLint height, GLint depth, + GLenum format, + GLsizei imageSize, const GLvoid *data); /** * Called by glGetCompressedTexImage. diff --git a/mesalib/src/mesa/main/depth.h b/mesalib/src/mesa/main/depth.h index b1ed1f3fd..2f428629f 100644 --- a/mesalib/src/mesa/main/depth.h +++ b/mesalib/src/mesa/main/depth.h @@ -1,68 +1,68 @@ -/** - * \file depth.h - * Depth buffer operations. - */ - -/* - * Mesa 3-D graphics library - * Version: 6.3 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef DEPTH_H -#define DEPTH_H - - -#include "glheader.h" -#include "mfeatures.h" - -struct gl_context; - - -#if _HAVE_FULL_GL - -extern void GLAPIENTRY -_mesa_ClearDepth( GLclampd depth ); - -extern void GLAPIENTRY -_mesa_ClearDepthf( GLclampf depth ); - -extern void GLAPIENTRY -_mesa_DepthFunc( GLenum func ); - -extern void GLAPIENTRY -_mesa_DepthMask( GLboolean flag ); - -extern void GLAPIENTRY -_mesa_DepthBoundsEXT( GLclampd zmin, GLclampd zmax ); - -extern void -_mesa_init_depth( struct gl_context * ctx ); - -#else - -/** No-op */ -#define _mesa_init_depth( c ) ((void)0) - -#endif - -#endif +/** + * \file depth.h + * Depth buffer operations. + */ + +/* + * Mesa 3-D graphics library + * Version: 6.3 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef DEPTH_H +#define DEPTH_H + + +#include "glheader.h" +#include "mfeatures.h" + +struct gl_context; + + +#if _HAVE_FULL_GL + +extern void GLAPIENTRY +_mesa_ClearDepth( GLclampd depth ); + +extern void GLAPIENTRY +_mesa_ClearDepthf( GLclampf depth ); + +extern void GLAPIENTRY +_mesa_DepthFunc( GLenum func ); + +extern void GLAPIENTRY +_mesa_DepthMask( GLboolean flag ); + +extern void GLAPIENTRY +_mesa_DepthBoundsEXT( GLclampd zmin, GLclampd zmax ); + +extern void +_mesa_init_depth( struct gl_context * ctx ); + +#else + +/** No-op */ +#define _mesa_init_depth( c ) ((void)0) + +#endif + +#endif diff --git a/mesalib/src/mesa/main/drawtex.h b/mesalib/src/mesa/main/drawtex.h index 78bbd5b45..13ff6f97e 100644 --- a/mesalib/src/mesa/main/drawtex.h +++ b/mesalib/src/mesa/main/drawtex.h @@ -1,61 +1,61 @@ -/* - * Copyright (C) 2009 Chia-I Wu - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -#ifndef DRAWTEX_H -#define DRAWTEX_H - - -#include "glheader.h" -#include "mfeatures.h" - - -#if FEATURE_OES_draw_texture - -extern void GLAPIENTRY -_mesa_DrawTexf(GLfloat x, GLfloat y, GLfloat z, GLfloat width, GLfloat height); - -extern void GLAPIENTRY -_mesa_DrawTexfv(const GLfloat *coords); - -extern void GLAPIENTRY -_mesa_DrawTexi(GLint x, GLint y, GLint z, GLint width, GLint height); - -extern void GLAPIENTRY -_mesa_DrawTexiv(const GLint *coords); - -extern void GLAPIENTRY -_mesa_DrawTexs(GLshort x, GLshort y, GLshort z, GLshort width, GLshort height); - -extern void GLAPIENTRY -_mesa_DrawTexsv(const GLshort *coords); - -extern void GLAPIENTRY -_mesa_DrawTexx(GLfixed x, GLfixed y, GLfixed z, GLfixed width, GLfixed height); - -extern void GLAPIENTRY -_mesa_DrawTexxv(const GLfixed *coords); - -#endif /* FEATURE_OES_draw_texture */ - - -#endif /* DRAWTEX_H */ +/* + * Copyright (C) 2009 Chia-I Wu + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ + +#ifndef DRAWTEX_H +#define DRAWTEX_H + + +#include "glheader.h" +#include "mfeatures.h" + + +#if FEATURE_OES_draw_texture + +extern void GLAPIENTRY +_mesa_DrawTexf(GLfloat x, GLfloat y, GLfloat z, GLfloat width, GLfloat height); + +extern void GLAPIENTRY +_mesa_DrawTexfv(const GLfloat *coords); + +extern void GLAPIENTRY +_mesa_DrawTexi(GLint x, GLint y, GLint z, GLint width, GLint height); + +extern void GLAPIENTRY +_mesa_DrawTexiv(const GLint *coords); + +extern void GLAPIENTRY +_mesa_DrawTexs(GLshort x, GLshort y, GLshort z, GLshort width, GLshort height); + +extern void GLAPIENTRY +_mesa_DrawTexsv(const GLshort *coords); + +extern void GLAPIENTRY +_mesa_DrawTexx(GLfixed x, GLfixed y, GLfixed z, GLfixed width, GLfixed height); + +extern void GLAPIENTRY +_mesa_DrawTexxv(const GLfixed *coords); + +#endif /* FEATURE_OES_draw_texture */ + + +#endif /* DRAWTEX_H */ diff --git a/mesalib/src/mesa/main/enable.h b/mesalib/src/mesa/main/enable.h index 685e11d47..6d90c170c 100644 --- a/mesalib/src/mesa/main/enable.h +++ b/mesalib/src/mesa/main/enable.h @@ -1,71 +1,71 @@ -/** - * \file enable.h - * Enable/disable/query GL capabilities. - */ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef ENABLE_H -#define ENABLE_H - - -#include "glheader.h" - -struct gl_context; - - -extern void -_mesa_set_enable( struct gl_context* ctx, GLenum cap, GLboolean state ); - -extern void GLAPIENTRY -_mesa_Disable( GLenum cap ); - -extern void GLAPIENTRY -_mesa_Enable( GLenum cap ); - -extern GLboolean GLAPIENTRY -_mesa_IsEnabled( GLenum cap ); - -extern void -_mesa_set_enablei(struct gl_context *ctx, GLenum cap, GLuint index, GLboolean state); - -extern void GLAPIENTRY -_mesa_DisableIndexed( GLenum cap, GLuint index ); - -extern void GLAPIENTRY -_mesa_EnableIndexed( GLenum cap, GLuint index ); - -extern GLboolean GLAPIENTRY -_mesa_IsEnabledIndexed( GLenum cap, GLuint index ); - -extern void GLAPIENTRY -_mesa_EnableClientState( GLenum cap ); - -extern void GLAPIENTRY -_mesa_DisableClientState( GLenum cap ); - - -#endif +/** + * \file enable.h + * Enable/disable/query GL capabilities. + */ + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef ENABLE_H +#define ENABLE_H + + +#include "glheader.h" + +struct gl_context; + + +extern void +_mesa_set_enable( struct gl_context* ctx, GLenum cap, GLboolean state ); + +extern void GLAPIENTRY +_mesa_Disable( GLenum cap ); + +extern void GLAPIENTRY +_mesa_Enable( GLenum cap ); + +extern GLboolean GLAPIENTRY +_mesa_IsEnabled( GLenum cap ); + +extern void +_mesa_set_enablei(struct gl_context *ctx, GLenum cap, GLuint index, GLboolean state); + +extern void GLAPIENTRY +_mesa_DisableIndexed( GLenum cap, GLuint index ); + +extern void GLAPIENTRY +_mesa_EnableIndexed( GLenum cap, GLuint index ); + +extern GLboolean GLAPIENTRY +_mesa_IsEnabledIndexed( GLenum cap, GLuint index ); + +extern void GLAPIENTRY +_mesa_EnableClientState( GLenum cap ); + +extern void GLAPIENTRY +_mesa_DisableClientState( GLenum cap ); + + +#endif diff --git a/mesalib/src/mesa/main/extensions.c b/mesalib/src/mesa/main/extensions.c index cd76eeb3d..2688f7ace 100644 --- a/mesalib/src/mesa/main/extensions.c +++ b/mesalib/src/mesa/main/extensions.c @@ -112,6 +112,7 @@ static const struct extension extension_table[] = { { "GL_ARB_robustness", o(dummy_true), GL, 2010 }, { "GL_ARB_sampler_objects", o(ARB_sampler_objects), GL, 2009 }, { "GL_ARB_seamless_cube_map", o(ARB_seamless_cube_map), GL, 2009 }, + { "GL_ARB_shader_bit_encoding", o(ARB_shader_bit_encoding), GL, 2010 }, { "GL_ARB_shader_objects", o(ARB_shader_objects), GL, 2002 }, { "GL_ARB_shader_stencil_export", o(ARB_shader_stencil_export), GL, 2009 }, { "GL_ARB_shader_texture_lod", o(ARB_shader_texture_lod), GL, 2009 }, diff --git a/mesalib/src/mesa/main/ffvertex_prog.h b/mesalib/src/mesa/main/ffvertex_prog.h index 6d485bb2a..837a15efc 100644 --- a/mesalib/src/mesa/main/ffvertex_prog.h +++ b/mesalib/src/mesa/main/ffvertex_prog.h @@ -1,40 +1,40 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef FFVERTEX_PROG_H -#define FFVERTEX_PROG_H - - -struct gl_context; - -struct gl_vertex_program * -_mesa_get_fixed_func_vertex_program(struct gl_context *ctx); - - - -#endif /* FFVERTEX_PROG_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef FFVERTEX_PROG_H +#define FFVERTEX_PROG_H + + +struct gl_context; + +struct gl_vertex_program * +_mesa_get_fixed_func_vertex_program(struct gl_context *ctx); + + + +#endif /* FFVERTEX_PROG_H */ diff --git a/mesalib/src/mesa/main/fog.h b/mesalib/src/mesa/main/fog.h index 72a0657af..9191a4a54 100644 --- a/mesalib/src/mesa/main/fog.h +++ b/mesalib/src/mesa/main/fog.h @@ -1,69 +1,69 @@ -/** - * \file fog.h - * Fog operations. - * - * \if subset - * (No-op) - * - * \endif - */ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef FOG_H -#define FOG_H - - -#include "glheader.h" -#include "mfeatures.h" - -struct gl_context; - - -#if _HAVE_FULL_GL - -extern void GLAPIENTRY -_mesa_Fogf(GLenum pname, GLfloat param); - -extern void GLAPIENTRY -_mesa_Fogi(GLenum pname, GLint param ); - -extern void GLAPIENTRY -_mesa_Fogfv(GLenum pname, const GLfloat *params ); - -extern void GLAPIENTRY -_mesa_Fogiv(GLenum pname, const GLint *params ); - -extern void _mesa_init_fog( struct gl_context * ctx ); - -#else - -/** No-op */ -#define _mesa_init_fog( c ) ((void)0) - -#endif - -#endif +/** + * \file fog.h + * Fog operations. + * + * \if subset + * (No-op) + * + * \endif + */ + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef FOG_H +#define FOG_H + + +#include "glheader.h" +#include "mfeatures.h" + +struct gl_context; + + +#if _HAVE_FULL_GL + +extern void GLAPIENTRY +_mesa_Fogf(GLenum pname, GLfloat param); + +extern void GLAPIENTRY +_mesa_Fogi(GLenum pname, GLint param ); + +extern void GLAPIENTRY +_mesa_Fogfv(GLenum pname, const GLfloat *params ); + +extern void GLAPIENTRY +_mesa_Fogiv(GLenum pname, const GLint *params ); + +extern void _mesa_init_fog( struct gl_context * ctx ); + +#else + +/** No-op */ +#define _mesa_init_fog( c ) ((void)0) + +#endif + +#endif diff --git a/mesalib/src/mesa/main/formats.c b/mesalib/src/mesa/main/formats.c index cecb70c01..ccc0b1707 100644 --- a/mesalib/src/mesa/main/formats.c +++ b/mesalib/src/mesa/main/formats.c @@ -1319,7 +1319,7 @@ static struct gl_format_info format_info[MESA_FORMAT_COUNT] = "MESA_FORMAT_RED_RGTC1", GL_RED, GL_UNSIGNED_NORMALIZED, - 4, 0, 0, 0, + 8, 0, 0, 0, 0, 0, 0, 0, 0, 4, 4, 8 /* 8 bytes per 4x4 block */ }, @@ -1328,7 +1328,7 @@ static struct gl_format_info format_info[MESA_FORMAT_COUNT] = "MESA_FORMAT_SIGNED_RED_RGTC1", GL_RED, GL_SIGNED_NORMALIZED, - 4, 0, 0, 0, + 8, 0, 0, 0, 0, 0, 0, 0, 0, 4, 4, 8 /* 8 bytes per 4x4 block */ }, @@ -1337,7 +1337,7 @@ static struct gl_format_info format_info[MESA_FORMAT_COUNT] = "MESA_FORMAT_RG_RGTC2", GL_RG, GL_UNSIGNED_NORMALIZED, - 4, 4, 0, 0, + 8, 8, 0, 0, 0, 0, 0, 0, 0, 4, 4, 16 /* 16 bytes per 4x4 block */ }, @@ -1346,7 +1346,7 @@ static struct gl_format_info format_info[MESA_FORMAT_COUNT] = "MESA_FORMAT_SIGNED_RG_RGTC2", GL_RG, GL_SIGNED_NORMALIZED, - 4, 4, 0, 0, + 8, 8, 0, 0, 0, 0, 0, 0, 0, 4, 4, 16 /* 16 bytes per 4x4 block */ }, diff --git a/mesalib/src/mesa/main/hint.h b/mesalib/src/mesa/main/hint.h index 3cdd2d650..6c6206874 100644 --- a/mesalib/src/mesa/main/hint.h +++ b/mesalib/src/mesa/main/hint.h @@ -1,59 +1,59 @@ -/** - * \file hint.h - * Hints operations. - * - * \if subset - * (No-op) - * - * \endif - */ - -/* - * Mesa 3-D graphics library - * Version: 4.1 - * - * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef HINT_H -#define HINT_H - -#include "glheader.h" -#include "mfeatures.h" - -struct gl_context; - -#if _HAVE_FULL_GL - -extern void GLAPIENTRY -_mesa_Hint( GLenum target, GLenum mode ); - -extern void -_mesa_init_hint( struct gl_context * ctx ); - -#else - -/** No-op */ -#define _mesa_init_hint( c ) ((void) 0) - -#endif - -#endif +/** + * \file hint.h + * Hints operations. + * + * \if subset + * (No-op) + * + * \endif + */ + +/* + * Mesa 3-D graphics library + * Version: 4.1 + * + * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef HINT_H +#define HINT_H + +#include "glheader.h" +#include "mfeatures.h" + +struct gl_context; + +#if _HAVE_FULL_GL + +extern void GLAPIENTRY +_mesa_Hint( GLenum target, GLenum mode ); + +extern void +_mesa_init_hint( struct gl_context * ctx ); + +#else + +/** No-op */ +#define _mesa_init_hint( c ) ((void) 0) + +#endif + +#endif diff --git a/mesalib/src/mesa/main/lines.h b/mesalib/src/mesa/main/lines.h index c882fc679..8e8b3f8d6 100644 --- a/mesalib/src/mesa/main/lines.h +++ b/mesalib/src/mesa/main/lines.h @@ -1,49 +1,49 @@ -/** - * \file lines.h - * Line operations. - */ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - - -#ifndef LINES_H -#define LINES_H - - -#include "glheader.h" - -struct gl_context; - -extern void GLAPIENTRY -_mesa_LineWidth( GLfloat width ); - -extern void GLAPIENTRY -_mesa_LineStipple( GLint factor, GLushort pattern ); - -extern void GLAPIENTRY -_mesa_init_line( struct gl_context * ctx ); - -#endif +/** + * \file lines.h + * Line operations. + */ + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + + +#ifndef LINES_H +#define LINES_H + + +#include "glheader.h" + +struct gl_context; + +extern void GLAPIENTRY +_mesa_LineWidth( GLfloat width ); + +extern void GLAPIENTRY +_mesa_LineStipple( GLint factor, GLushort pattern ); + +extern void GLAPIENTRY +_mesa_init_line( struct gl_context * ctx ); + +#endif diff --git a/mesalib/src/mesa/main/matrix.h b/mesalib/src/mesa/main/matrix.h index a4e7f7cfd..2878cc134 100644 --- a/mesalib/src/mesa/main/matrix.h +++ b/mesalib/src/mesa/main/matrix.h @@ -1,113 +1,113 @@ -/* - * Mesa 3-D graphics library - * Version: 7.5 - * - * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. - * Copyright (C) 2009 VMware, Inc. All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef MATRIX_H -#define MATRIX_H - - -#include "glheader.h" - -struct gl_context; - -extern void GLAPIENTRY -_mesa_Frustum( GLdouble left, GLdouble right, - GLdouble bottom, GLdouble top, - GLdouble nearval, GLdouble farval ); - -extern void GLAPIENTRY -_mesa_Ortho( GLdouble left, GLdouble right, - GLdouble bottom, GLdouble top, - GLdouble nearval, GLdouble farval ); - -extern void GLAPIENTRY -_mesa_PushMatrix( void ); - -extern void GLAPIENTRY -_mesa_PopMatrix( void ); - -extern void GLAPIENTRY -_mesa_LoadIdentity( void ); - -extern void GLAPIENTRY -_mesa_LoadMatrixf( const GLfloat *m ); - -extern void GLAPIENTRY -_mesa_LoadMatrixd( const GLdouble *m ); - -extern void GLAPIENTRY -_mesa_MatrixMode( GLenum mode ); - -extern void GLAPIENTRY -_mesa_MultMatrixf( const GLfloat *m ); - -extern void GLAPIENTRY -_mesa_MultMatrixd( const GLdouble *m ); - -extern void GLAPIENTRY -_mesa_Rotatef( GLfloat angle, GLfloat x, GLfloat y, GLfloat z ); - -extern void GLAPIENTRY -_mesa_Rotated( GLdouble angle, GLdouble x, GLdouble y, GLdouble z ); - -extern void GLAPIENTRY -_mesa_Scalef( GLfloat x, GLfloat y, GLfloat z ); - -extern void GLAPIENTRY -_mesa_Scaled( GLdouble x, GLdouble y, GLdouble z ); - -extern void GLAPIENTRY -_mesa_Translatef( GLfloat x, GLfloat y, GLfloat z ); - -extern void GLAPIENTRY -_mesa_Translated( GLdouble x, GLdouble y, GLdouble z ); - -extern void GLAPIENTRY -_mesa_LoadTransposeMatrixfARB( const GLfloat *m ); - -extern void GLAPIENTRY -_mesa_LoadTransposeMatrixdARB( const GLdouble *m ); - -extern void GLAPIENTRY -_mesa_MultTransposeMatrixfARB( const GLfloat *m ); - -extern void GLAPIENTRY -_mesa_MultTransposeMatrixdARB( const GLdouble *m ); - - -extern void -_mesa_init_matrix( struct gl_context * ctx ); - -extern void -_mesa_init_transform( struct gl_context *ctx ); - -extern void -_mesa_free_matrix_data( struct gl_context *ctx ); - -extern void -_mesa_update_modelview_project( struct gl_context *ctx, GLuint newstate ); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.5 + * + * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. + * Copyright (C) 2009 VMware, Inc. All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef MATRIX_H +#define MATRIX_H + + +#include "glheader.h" + +struct gl_context; + +extern void GLAPIENTRY +_mesa_Frustum( GLdouble left, GLdouble right, + GLdouble bottom, GLdouble top, + GLdouble nearval, GLdouble farval ); + +extern void GLAPIENTRY +_mesa_Ortho( GLdouble left, GLdouble right, + GLdouble bottom, GLdouble top, + GLdouble nearval, GLdouble farval ); + +extern void GLAPIENTRY +_mesa_PushMatrix( void ); + +extern void GLAPIENTRY +_mesa_PopMatrix( void ); + +extern void GLAPIENTRY +_mesa_LoadIdentity( void ); + +extern void GLAPIENTRY +_mesa_LoadMatrixf( const GLfloat *m ); + +extern void GLAPIENTRY +_mesa_LoadMatrixd( const GLdouble *m ); + +extern void GLAPIENTRY +_mesa_MatrixMode( GLenum mode ); + +extern void GLAPIENTRY +_mesa_MultMatrixf( const GLfloat *m ); + +extern void GLAPIENTRY +_mesa_MultMatrixd( const GLdouble *m ); + +extern void GLAPIENTRY +_mesa_Rotatef( GLfloat angle, GLfloat x, GLfloat y, GLfloat z ); + +extern void GLAPIENTRY +_mesa_Rotated( GLdouble angle, GLdouble x, GLdouble y, GLdouble z ); + +extern void GLAPIENTRY +_mesa_Scalef( GLfloat x, GLfloat y, GLfloat z ); + +extern void GLAPIENTRY +_mesa_Scaled( GLdouble x, GLdouble y, GLdouble z ); + +extern void GLAPIENTRY +_mesa_Translatef( GLfloat x, GLfloat y, GLfloat z ); + +extern void GLAPIENTRY +_mesa_Translated( GLdouble x, GLdouble y, GLdouble z ); + +extern void GLAPIENTRY +_mesa_LoadTransposeMatrixfARB( const GLfloat *m ); + +extern void GLAPIENTRY +_mesa_LoadTransposeMatrixdARB( const GLdouble *m ); + +extern void GLAPIENTRY +_mesa_MultTransposeMatrixfARB( const GLfloat *m ); + +extern void GLAPIENTRY +_mesa_MultTransposeMatrixdARB( const GLdouble *m ); + + +extern void +_mesa_init_matrix( struct gl_context * ctx ); + +extern void +_mesa_init_transform( struct gl_context *ctx ); + +extern void +_mesa_free_matrix_data( struct gl_context *ctx ); + +extern void +_mesa_update_modelview_project( struct gl_context *ctx, GLuint newstate ); + + +#endif diff --git a/mesalib/src/mesa/main/mipmap.c b/mesalib/src/mesa/main/mipmap.c index abd26b8d3..250d3c6dc 100644 --- a/mesalib/src/mesa/main/mipmap.c +++ b/mesalib/src/mesa/main/mipmap.c @@ -2133,10 +2133,10 @@ generate_mipmap_compressed(struct gl_context *ctx, GLenum target, } /* The image space was allocated above so use glTexSubImage now */ - ctx->Driver.TexSubImage2D(ctx, dstImage, - 0, 0, dstWidth, dstHeight, - temp_base_format, temp_datatype, - temp_dst, &ctx->DefaultPacking); + ctx->Driver.TexSubImage(ctx, 2, dstImage, + 0, 0, 0, dstWidth, dstHeight, 1, + temp_base_format, temp_datatype, + temp_dst, &ctx->DefaultPacking); /* swap src and dest pointers */ { diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h index 3a8cac9c5..ea40ffc9c 100644 --- a/mesalib/src/mesa/main/mtypes.h +++ b/mesalib/src/mesa/main/mtypes.h @@ -2895,6 +2895,7 @@ struct gl_extensions GLboolean ARB_point_sprite; GLboolean ARB_sampler_objects; GLboolean ARB_seamless_cube_map; + GLboolean ARB_shader_bit_encoding; GLboolean ARB_shader_objects; GLboolean ARB_shader_stencil_export; GLboolean ARB_shader_texture_lod; diff --git a/mesalib/src/mesa/main/multisample.c b/mesalib/src/mesa/main/multisample.c index 4eb1ab3a3..67fdb8c9d 100644 --- a/mesalib/src/mesa/main/multisample.c +++ b/mesalib/src/mesa/main/multisample.c @@ -1,62 +1,62 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#include "main/glheader.h" -#include "main/context.h" -#include "main/macros.h" -#include "main/multisample.h" -#include "main/mtypes.h" - - -/** - * Called via glSampleCoverageARB - */ -void GLAPIENTRY -_mesa_SampleCoverageARB(GLclampf value, GLboolean invert) -{ - GET_CURRENT_CONTEXT(ctx); - - ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH( ctx ); - - ctx->Multisample.SampleCoverageValue = (GLfloat) CLAMP(value, 0.0, 1.0); - ctx->Multisample.SampleCoverageInvert = invert; - ctx->NewState |= _NEW_MULTISAMPLE; -} - - -/** - * Initialize the context's multisample state. - * \param ctx the GL context. - */ -void -_mesa_init_multisample(struct gl_context *ctx) -{ - ctx->Multisample.Enabled = GL_TRUE; - ctx->Multisample.SampleAlphaToCoverage = GL_FALSE; - ctx->Multisample.SampleAlphaToOne = GL_FALSE; - ctx->Multisample.SampleCoverage = GL_FALSE; - ctx->Multisample.SampleCoverageValue = 1.0; - ctx->Multisample.SampleCoverageInvert = GL_FALSE; -} +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#include "main/glheader.h" +#include "main/context.h" +#include "main/macros.h" +#include "main/multisample.h" +#include "main/mtypes.h" + + +/** + * Called via glSampleCoverageARB + */ +void GLAPIENTRY +_mesa_SampleCoverageARB(GLclampf value, GLboolean invert) +{ + GET_CURRENT_CONTEXT(ctx); + + ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH( ctx ); + + ctx->Multisample.SampleCoverageValue = (GLfloat) CLAMP(value, 0.0, 1.0); + ctx->Multisample.SampleCoverageInvert = invert; + ctx->NewState |= _NEW_MULTISAMPLE; +} + + +/** + * Initialize the context's multisample state. + * \param ctx the GL context. + */ +void +_mesa_init_multisample(struct gl_context *ctx) +{ + ctx->Multisample.Enabled = GL_TRUE; + ctx->Multisample.SampleAlphaToCoverage = GL_FALSE; + ctx->Multisample.SampleAlphaToOne = GL_FALSE; + ctx->Multisample.SampleCoverage = GL_FALSE; + ctx->Multisample.SampleCoverageValue = 1.0; + ctx->Multisample.SampleCoverageInvert = GL_FALSE; +} diff --git a/mesalib/src/mesa/main/multisample.h b/mesalib/src/mesa/main/multisample.h index c1c967bbe..e86d4092b 100644 --- a/mesalib/src/mesa/main/multisample.h +++ b/mesalib/src/mesa/main/multisample.h @@ -1,41 +1,41 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef MULTISAMPLE_H -#define MULTISAMPLE_H - -#include "glheader.h" - -struct gl_context; - -extern void GLAPIENTRY -_mesa_SampleCoverageARB(GLclampf value, GLboolean invert); - - -extern void -_mesa_init_multisample(struct gl_context *ctx); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef MULTISAMPLE_H +#define MULTISAMPLE_H + +#include "glheader.h" + +struct gl_context; + +extern void GLAPIENTRY +_mesa_SampleCoverageARB(GLclampf value, GLboolean invert); + + +extern void +_mesa_init_multisample(struct gl_context *ctx); + + +#endif diff --git a/mesalib/src/mesa/main/pixelstore.h b/mesalib/src/mesa/main/pixelstore.h index 2995beb10..eb5081974 100644 --- a/mesalib/src/mesa/main/pixelstore.h +++ b/mesalib/src/mesa/main/pixelstore.h @@ -1,52 +1,52 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file pixelstore.h - * glPixelStore functions. - */ - - -#ifndef PIXELSTORE_H -#define PIXELSTORE_H - - -#include "glheader.h" - -struct gl_context; - - -extern void GLAPIENTRY -_mesa_PixelStorei( GLenum pname, GLint param ); - - -extern void GLAPIENTRY -_mesa_PixelStoref( GLenum pname, GLfloat param ); - - -extern void -_mesa_init_pixelstore( struct gl_context *ctx ); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file pixelstore.h + * glPixelStore functions. + */ + + +#ifndef PIXELSTORE_H +#define PIXELSTORE_H + + +#include "glheader.h" + +struct gl_context; + + +extern void GLAPIENTRY +_mesa_PixelStorei( GLenum pname, GLint param ); + + +extern void GLAPIENTRY +_mesa_PixelStoref( GLenum pname, GLfloat param ); + + +extern void +_mesa_init_pixelstore( struct gl_context *ctx ); + + +#endif diff --git a/mesalib/src/mesa/main/points.c b/mesalib/src/mesa/main/points.c index 78d40f61c..87bfae27e 100644 --- a/mesalib/src/mesa/main/points.c +++ b/mesalib/src/mesa/main/points.c @@ -1,272 +1,272 @@ -/** - * \file points.c - * Point operations. - */ - -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#include "glheader.h" -#include "context.h" -#include "macros.h" -#include "points.h" -#include "mtypes.h" - - -/** - * Set current point size. - * \param size point diameter in pixels - * \sa glPointSize(). - */ -void GLAPIENTRY -_mesa_PointSize( GLfloat size ) -{ - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (size <= 0.0) { - _mesa_error( ctx, GL_INVALID_VALUE, "glPointSize" ); - return; - } - - if (ctx->Point.Size == size) - return; - - FLUSH_VERTICES(ctx, _NEW_POINT); - ctx->Point.Size = size; - - if (ctx->Driver.PointSize) - ctx->Driver.PointSize(ctx, size); -} - - -#if _HAVE_FULL_GL - - -void GLAPIENTRY -_mesa_PointParameteri( GLenum pname, GLint param ) -{ - GLfloat p[3]; - p[0] = (GLfloat) param; - p[1] = p[2] = 0.0F; - _mesa_PointParameterfv(pname, p); -} - - -void GLAPIENTRY -_mesa_PointParameteriv( GLenum pname, const GLint *params ) -{ - GLfloat p[3]; - p[0] = (GLfloat) params[0]; - if (pname == GL_DISTANCE_ATTENUATION_EXT) { - p[1] = (GLfloat) params[1]; - p[2] = (GLfloat) params[2]; - } - _mesa_PointParameterfv(pname, p); -} - - -void GLAPIENTRY -_mesa_PointParameterf( GLenum pname, GLfloat param) -{ - GLfloat p[3]; - p[0] = param; - p[1] = p[2] = 0.0F; - _mesa_PointParameterfv(pname, p); -} - - -void GLAPIENTRY -_mesa_PointParameterfv( GLenum pname, const GLfloat *params) -{ - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - switch (pname) { - case GL_DISTANCE_ATTENUATION_EXT: - if (ctx->Extensions.EXT_point_parameters) { - if (TEST_EQ_3V(ctx->Point.Params, params)) - return; - FLUSH_VERTICES(ctx, _NEW_POINT); - COPY_3V(ctx->Point.Params, params); - ctx->Point._Attenuated = (ctx->Point.Params[0] != 1.0 || - ctx->Point.Params[1] != 0.0 || - ctx->Point.Params[2] != 0.0); - - if (ctx->Point._Attenuated) - ctx->_TriangleCaps |= DD_POINT_ATTEN; - else - ctx->_TriangleCaps &= ~DD_POINT_ATTEN; - } - else { - _mesa_error(ctx, GL_INVALID_ENUM, - "glPointParameterf[v]{EXT,ARB}(pname)"); - return; - } - break; - case GL_POINT_SIZE_MIN_EXT: - if (ctx->Extensions.EXT_point_parameters) { - if (params[0] < 0.0F) { - _mesa_error( ctx, GL_INVALID_VALUE, - "glPointParameterf[v]{EXT,ARB}(param)" ); - return; - } - if (ctx->Point.MinSize == params[0]) - return; - FLUSH_VERTICES(ctx, _NEW_POINT); - ctx->Point.MinSize = params[0]; - } - else { - _mesa_error(ctx, GL_INVALID_ENUM, - "glPointParameterf[v]{EXT,ARB}(pname)"); - return; - } - break; - case GL_POINT_SIZE_MAX_EXT: - if (ctx->Extensions.EXT_point_parameters) { - if (params[0] < 0.0F) { - _mesa_error( ctx, GL_INVALID_VALUE, - "glPointParameterf[v]{EXT,ARB}(param)" ); - return; - } - if (ctx->Point.MaxSize == params[0]) - return; - FLUSH_VERTICES(ctx, _NEW_POINT); - ctx->Point.MaxSize = params[0]; - } - else { - _mesa_error(ctx, GL_INVALID_ENUM, - "glPointParameterf[v]{EXT,ARB}(pname)"); - return; - } - break; - case GL_POINT_FADE_THRESHOLD_SIZE_EXT: - if (ctx->Extensions.EXT_point_parameters) { - if (params[0] < 0.0F) { - _mesa_error( ctx, GL_INVALID_VALUE, - "glPointParameterf[v]{EXT,ARB}(param)" ); - return; - } - if (ctx->Point.Threshold == params[0]) - return; - FLUSH_VERTICES(ctx, _NEW_POINT); - ctx->Point.Threshold = params[0]; - } - else { - _mesa_error(ctx, GL_INVALID_ENUM, - "glPointParameterf[v]{EXT,ARB}(pname)"); - return; - } - break; - case GL_POINT_SPRITE_R_MODE_NV: - /* This is one area where ARB_point_sprite and NV_point_sprite - * differ. In ARB_point_sprite the POINT_SPRITE_R_MODE is - * always ZERO. NV_point_sprite adds the S and R modes. - */ - if (ctx->Extensions.NV_point_sprite) { - GLenum value = (GLenum) params[0]; - if (value != GL_ZERO && value != GL_S && value != GL_R) { - _mesa_error(ctx, GL_INVALID_VALUE, - "glPointParameterf[v]{EXT,ARB}(param)"); - return; - } - if (ctx->Point.SpriteRMode == value) - return; - FLUSH_VERTICES(ctx, _NEW_POINT); - ctx->Point.SpriteRMode = value; - } - else { - _mesa_error(ctx, GL_INVALID_ENUM, - "glPointParameterf[v]{EXT,ARB}(pname)"); - return; - } - break; - case GL_POINT_SPRITE_COORD_ORIGIN: - /* This is not completely correct. GL_POINT_SPRITE_COORD_ORIGIN was - * added to point sprites when the extension was merged into OpenGL - * 2.0. It is expected that an implementation supporting OpenGL 1.4 - * and GL_ARB_point_sprite will generate an error here. - */ - if (ctx->Extensions.ARB_point_sprite) { - GLenum value = (GLenum) params[0]; - if (value != GL_LOWER_LEFT && value != GL_UPPER_LEFT) { - _mesa_error(ctx, GL_INVALID_VALUE, - "glPointParameterf[v]{EXT,ARB}(param)"); - return; - } - if (ctx->Point.SpriteOrigin == value) - return; - FLUSH_VERTICES(ctx, _NEW_POINT); - ctx->Point.SpriteOrigin = value; - } - else { - _mesa_error(ctx, GL_INVALID_ENUM, - "glPointParameterf[v]{EXT,ARB}(pname)"); - return; - } - break; - default: - _mesa_error( ctx, GL_INVALID_ENUM, - "glPointParameterf[v]{EXT,ARB}(pname)" ); - return; - } - - if (ctx->Driver.PointParameterfv) - (*ctx->Driver.PointParameterfv)(ctx, pname, params); -} -#endif - - - -/** - * Initialize the context point state. - * - * \param ctx GL context. - * - * Initializes __struct gl_contextRec::Point and point related constants in - * __struct gl_contextRec::Const. - */ -void -_mesa_init_point(struct gl_context *ctx) -{ - GLuint i; - - ctx->Point.SmoothFlag = GL_FALSE; - ctx->Point.Size = 1.0; - ctx->Point.Params[0] = 1.0; - ctx->Point.Params[1] = 0.0; - ctx->Point.Params[2] = 0.0; - ctx->Point._Attenuated = GL_FALSE; - ctx->Point.MinSize = 0.0; - ctx->Point.MaxSize - = MAX2(ctx->Const.MaxPointSize, ctx->Const.MaxPointSizeAA); - ctx->Point.Threshold = 1.0; - ctx->Point.PointSprite = GL_FALSE; /* GL_ARB/NV_point_sprite */ - ctx->Point.SpriteRMode = GL_ZERO; /* GL_NV_point_sprite (only!) */ - ctx->Point.SpriteOrigin = GL_UPPER_LEFT; /* GL_ARB_point_sprite */ - for (i = 0; i < Elements(ctx->Point.CoordReplace); i++) { - ctx->Point.CoordReplace[i] = GL_FALSE; /* GL_ARB/NV_point_sprite */ - } -} +/** + * \file points.c + * Point operations. + */ + +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#include "glheader.h" +#include "context.h" +#include "macros.h" +#include "points.h" +#include "mtypes.h" + + +/** + * Set current point size. + * \param size point diameter in pixels + * \sa glPointSize(). + */ +void GLAPIENTRY +_mesa_PointSize( GLfloat size ) +{ + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (size <= 0.0) { + _mesa_error( ctx, GL_INVALID_VALUE, "glPointSize" ); + return; + } + + if (ctx->Point.Size == size) + return; + + FLUSH_VERTICES(ctx, _NEW_POINT); + ctx->Point.Size = size; + + if (ctx->Driver.PointSize) + ctx->Driver.PointSize(ctx, size); +} + + +#if _HAVE_FULL_GL + + +void GLAPIENTRY +_mesa_PointParameteri( GLenum pname, GLint param ) +{ + GLfloat p[3]; + p[0] = (GLfloat) param; + p[1] = p[2] = 0.0F; + _mesa_PointParameterfv(pname, p); +} + + +void GLAPIENTRY +_mesa_PointParameteriv( GLenum pname, const GLint *params ) +{ + GLfloat p[3]; + p[0] = (GLfloat) params[0]; + if (pname == GL_DISTANCE_ATTENUATION_EXT) { + p[1] = (GLfloat) params[1]; + p[2] = (GLfloat) params[2]; + } + _mesa_PointParameterfv(pname, p); +} + + +void GLAPIENTRY +_mesa_PointParameterf( GLenum pname, GLfloat param) +{ + GLfloat p[3]; + p[0] = param; + p[1] = p[2] = 0.0F; + _mesa_PointParameterfv(pname, p); +} + + +void GLAPIENTRY +_mesa_PointParameterfv( GLenum pname, const GLfloat *params) +{ + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + switch (pname) { + case GL_DISTANCE_ATTENUATION_EXT: + if (ctx->Extensions.EXT_point_parameters) { + if (TEST_EQ_3V(ctx->Point.Params, params)) + return; + FLUSH_VERTICES(ctx, _NEW_POINT); + COPY_3V(ctx->Point.Params, params); + ctx->Point._Attenuated = (ctx->Point.Params[0] != 1.0 || + ctx->Point.Params[1] != 0.0 || + ctx->Point.Params[2] != 0.0); + + if (ctx->Point._Attenuated) + ctx->_TriangleCaps |= DD_POINT_ATTEN; + else + ctx->_TriangleCaps &= ~DD_POINT_ATTEN; + } + else { + _mesa_error(ctx, GL_INVALID_ENUM, + "glPointParameterf[v]{EXT,ARB}(pname)"); + return; + } + break; + case GL_POINT_SIZE_MIN_EXT: + if (ctx->Extensions.EXT_point_parameters) { + if (params[0] < 0.0F) { + _mesa_error( ctx, GL_INVALID_VALUE, + "glPointParameterf[v]{EXT,ARB}(param)" ); + return; + } + if (ctx->Point.MinSize == params[0]) + return; + FLUSH_VERTICES(ctx, _NEW_POINT); + ctx->Point.MinSize = params[0]; + } + else { + _mesa_error(ctx, GL_INVALID_ENUM, + "glPointParameterf[v]{EXT,ARB}(pname)"); + return; + } + break; + case GL_POINT_SIZE_MAX_EXT: + if (ctx->Extensions.EXT_point_parameters) { + if (params[0] < 0.0F) { + _mesa_error( ctx, GL_INVALID_VALUE, + "glPointParameterf[v]{EXT,ARB}(param)" ); + return; + } + if (ctx->Point.MaxSize == params[0]) + return; + FLUSH_VERTICES(ctx, _NEW_POINT); + ctx->Point.MaxSize = params[0]; + } + else { + _mesa_error(ctx, GL_INVALID_ENUM, + "glPointParameterf[v]{EXT,ARB}(pname)"); + return; + } + break; + case GL_POINT_FADE_THRESHOLD_SIZE_EXT: + if (ctx->Extensions.EXT_point_parameters) { + if (params[0] < 0.0F) { + _mesa_error( ctx, GL_INVALID_VALUE, + "glPointParameterf[v]{EXT,ARB}(param)" ); + return; + } + if (ctx->Point.Threshold == params[0]) + return; + FLUSH_VERTICES(ctx, _NEW_POINT); + ctx->Point.Threshold = params[0]; + } + else { + _mesa_error(ctx, GL_INVALID_ENUM, + "glPointParameterf[v]{EXT,ARB}(pname)"); + return; + } + break; + case GL_POINT_SPRITE_R_MODE_NV: + /* This is one area where ARB_point_sprite and NV_point_sprite + * differ. In ARB_point_sprite the POINT_SPRITE_R_MODE is + * always ZERO. NV_point_sprite adds the S and R modes. + */ + if (ctx->Extensions.NV_point_sprite) { + GLenum value = (GLenum) params[0]; + if (value != GL_ZERO && value != GL_S && value != GL_R) { + _mesa_error(ctx, GL_INVALID_VALUE, + "glPointParameterf[v]{EXT,ARB}(param)"); + return; + } + if (ctx->Point.SpriteRMode == value) + return; + FLUSH_VERTICES(ctx, _NEW_POINT); + ctx->Point.SpriteRMode = value; + } + else { + _mesa_error(ctx, GL_INVALID_ENUM, + "glPointParameterf[v]{EXT,ARB}(pname)"); + return; + } + break; + case GL_POINT_SPRITE_COORD_ORIGIN: + /* This is not completely correct. GL_POINT_SPRITE_COORD_ORIGIN was + * added to point sprites when the extension was merged into OpenGL + * 2.0. It is expected that an implementation supporting OpenGL 1.4 + * and GL_ARB_point_sprite will generate an error here. + */ + if (ctx->Extensions.ARB_point_sprite) { + GLenum value = (GLenum) params[0]; + if (value != GL_LOWER_LEFT && value != GL_UPPER_LEFT) { + _mesa_error(ctx, GL_INVALID_VALUE, + "glPointParameterf[v]{EXT,ARB}(param)"); + return; + } + if (ctx->Point.SpriteOrigin == value) + return; + FLUSH_VERTICES(ctx, _NEW_POINT); + ctx->Point.SpriteOrigin = value; + } + else { + _mesa_error(ctx, GL_INVALID_ENUM, + "glPointParameterf[v]{EXT,ARB}(pname)"); + return; + } + break; + default: + _mesa_error( ctx, GL_INVALID_ENUM, + "glPointParameterf[v]{EXT,ARB}(pname)" ); + return; + } + + if (ctx->Driver.PointParameterfv) + (*ctx->Driver.PointParameterfv)(ctx, pname, params); +} +#endif + + + +/** + * Initialize the context point state. + * + * \param ctx GL context. + * + * Initializes __struct gl_contextRec::Point and point related constants in + * __struct gl_contextRec::Const. + */ +void +_mesa_init_point(struct gl_context *ctx) +{ + GLuint i; + + ctx->Point.SmoothFlag = GL_FALSE; + ctx->Point.Size = 1.0; + ctx->Point.Params[0] = 1.0; + ctx->Point.Params[1] = 0.0; + ctx->Point.Params[2] = 0.0; + ctx->Point._Attenuated = GL_FALSE; + ctx->Point.MinSize = 0.0; + ctx->Point.MaxSize + = MAX2(ctx->Const.MaxPointSize, ctx->Const.MaxPointSizeAA); + ctx->Point.Threshold = 1.0; + ctx->Point.PointSprite = GL_FALSE; /* GL_ARB/NV_point_sprite */ + ctx->Point.SpriteRMode = GL_ZERO; /* GL_NV_point_sprite (only!) */ + ctx->Point.SpriteOrigin = GL_UPPER_LEFT; /* GL_ARB_point_sprite */ + for (i = 0; i < Elements(ctx->Point.CoordReplace); i++) { + ctx->Point.CoordReplace[i] = GL_FALSE; /* GL_ARB/NV_point_sprite */ + } +} diff --git a/mesalib/src/mesa/main/points.h b/mesalib/src/mesa/main/points.h index 442053c8f..306a8a572 100644 --- a/mesalib/src/mesa/main/points.h +++ b/mesalib/src/mesa/main/points.h @@ -1,59 +1,59 @@ -/** - * \file points.h - * Point operations. - */ - -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef POINTS_H -#define POINTS_H - - -#include "glheader.h" - -struct gl_context; - - -extern void GLAPIENTRY -_mesa_PointSize( GLfloat size ); - -extern void GLAPIENTRY -_mesa_PointParameteri( GLenum pname, GLint param ); - -extern void GLAPIENTRY -_mesa_PointParameteriv( GLenum pname, const GLint *params ); - -extern void GLAPIENTRY -_mesa_PointParameterf( GLenum pname, GLfloat param ); - -extern void GLAPIENTRY -_mesa_PointParameterfv( GLenum pname, const GLfloat *params ); - -extern void -_mesa_init_point( struct gl_context * ctx ); - - -#endif +/** + * \file points.h + * Point operations. + */ + +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef POINTS_H +#define POINTS_H + + +#include "glheader.h" + +struct gl_context; + + +extern void GLAPIENTRY +_mesa_PointSize( GLfloat size ); + +extern void GLAPIENTRY +_mesa_PointParameteri( GLenum pname, GLint param ); + +extern void GLAPIENTRY +_mesa_PointParameteriv( GLenum pname, const GLint *params ); + +extern void GLAPIENTRY +_mesa_PointParameterf( GLenum pname, GLfloat param ); + +extern void GLAPIENTRY +_mesa_PointParameterfv( GLenum pname, const GLfloat *params ); + +extern void +_mesa_init_point( struct gl_context * ctx ); + + +#endif diff --git a/mesalib/src/mesa/main/rastpos.c b/mesalib/src/mesa/main/rastpos.c index 70f67d398..a1f6f4234 100644 --- a/mesalib/src/mesa/main/rastpos.c +++ b/mesalib/src/mesa/main/rastpos.c @@ -1,566 +1,566 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.3 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -/** - * \file rastpos.c - * Raster position operations. - */ - -#include "glheader.h" -#include "context.h" -#include "feedback.h" -#include "macros.h" -#include "mfeatures.h" -#include "mtypes.h" -#include "rastpos.h" -#include "state.h" -#include "main/dispatch.h" - - -#if FEATURE_rastpos - - -/** - * Helper function for all the RasterPos functions. - */ -static void -rasterpos(GLfloat x, GLfloat y, GLfloat z, GLfloat w) -{ - GET_CURRENT_CONTEXT(ctx); - GLfloat p[4]; - - p[0] = x; - p[1] = y; - p[2] = z; - p[3] = w; - - ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); - FLUSH_CURRENT(ctx, 0); - - if (ctx->NewState) - _mesa_update_state( ctx ); - - ctx->Driver.RasterPos(ctx, p); -} - - -static void GLAPIENTRY -_mesa_RasterPos2d(GLdouble x, GLdouble y) -{ - rasterpos((GLfloat)x, (GLfloat)y, (GLfloat)0.0, (GLfloat)1.0); -} - -static void GLAPIENTRY -_mesa_RasterPos2f(GLfloat x, GLfloat y) -{ - rasterpos(x, y, 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos2i(GLint x, GLint y) -{ - rasterpos((GLfloat) x, (GLfloat) y, 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos2s(GLshort x, GLshort y) -{ - rasterpos(x, y, 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3d(GLdouble x, GLdouble y, GLdouble z) -{ - rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3f(GLfloat x, GLfloat y, GLfloat z) -{ - rasterpos(x, y, z, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3i(GLint x, GLint y, GLint z) -{ - rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3s(GLshort x, GLshort y, GLshort z) -{ - rasterpos(x, y, z, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos4d(GLdouble x, GLdouble y, GLdouble z, GLdouble w) -{ - rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -_mesa_RasterPos4f(GLfloat x, GLfloat y, GLfloat z, GLfloat w) -{ - rasterpos(x, y, z, w); -} - -static void GLAPIENTRY -_mesa_RasterPos4i(GLint x, GLint y, GLint z, GLint w) -{ - rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -_mesa_RasterPos4s(GLshort x, GLshort y, GLshort z, GLshort w) -{ - rasterpos(x, y, z, w); -} - -static void GLAPIENTRY -_mesa_RasterPos2dv(const GLdouble *v) -{ - rasterpos((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos2fv(const GLfloat *v) -{ - rasterpos(v[0], v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos2iv(const GLint *v) -{ - rasterpos((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos2sv(const GLshort *v) -{ - rasterpos(v[0], v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3dv(const GLdouble *v) -{ - rasterpos((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3fv(const GLfloat *v) -{ - rasterpos(v[0], v[1], v[2], 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3iv(const GLint *v) -{ - rasterpos((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos3sv(const GLshort *v) -{ - rasterpos(v[0], v[1], v[2], 1.0F); -} - -static void GLAPIENTRY -_mesa_RasterPos4dv(const GLdouble *v) -{ - rasterpos((GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -_mesa_RasterPos4fv(const GLfloat *v) -{ - rasterpos(v[0], v[1], v[2], v[3]); -} - -static void GLAPIENTRY -_mesa_RasterPos4iv(const GLint *v) -{ - rasterpos((GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -_mesa_RasterPos4sv(const GLshort *v) -{ - rasterpos(v[0], v[1], v[2], v[3]); -} - - -/**********************************************************************/ -/*** GL_ARB_window_pos / GL_MESA_window_pos ***/ -/**********************************************************************/ - - -/** - * All glWindowPosMESA and glWindowPosARB commands call this function to - * update the current raster position. - */ -static void -window_pos3f(GLfloat x, GLfloat y, GLfloat z) -{ - GET_CURRENT_CONTEXT(ctx); - GLfloat z2; - - ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); - FLUSH_CURRENT(ctx, 0); - - z2 = CLAMP(z, 0.0F, 1.0F) * (ctx->Viewport.Far - ctx->Viewport.Near) - + ctx->Viewport.Near; - - /* set raster position */ - ctx->Current.RasterPos[0] = x; - ctx->Current.RasterPos[1] = y; - ctx->Current.RasterPos[2] = z2; - ctx->Current.RasterPos[3] = 1.0F; - - ctx->Current.RasterPosValid = GL_TRUE; - - if (ctx->Fog.FogCoordinateSource == GL_FOG_COORDINATE_EXT) - ctx->Current.RasterDistance = ctx->Current.Attrib[VERT_ATTRIB_FOG][0]; - else - ctx->Current.RasterDistance = 0.0; - - /* raster color = current color or index */ - ctx->Current.RasterColor[0] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][0], 0.0F, 1.0F); - ctx->Current.RasterColor[1] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][1], 0.0F, 1.0F); - ctx->Current.RasterColor[2] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][2], 0.0F, 1.0F); - ctx->Current.RasterColor[3] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][3], 0.0F, 1.0F); - ctx->Current.RasterSecondaryColor[0] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][0], 0.0F, 1.0F); - ctx->Current.RasterSecondaryColor[1] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][1], 0.0F, 1.0F); - ctx->Current.RasterSecondaryColor[2] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][2], 0.0F, 1.0F); - ctx->Current.RasterSecondaryColor[3] - = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][3], 0.0F, 1.0F); - - /* raster texcoord = current texcoord */ - { - GLuint texSet; - for (texSet = 0; texSet < ctx->Const.MaxTextureCoordUnits; texSet++) { - assert(texSet < Elements(ctx->Current.RasterTexCoords)); - COPY_4FV( ctx->Current.RasterTexCoords[texSet], - ctx->Current.Attrib[VERT_ATTRIB_TEX0 + texSet] ); - } - } - - if (ctx->RenderMode==GL_SELECT) { - _mesa_update_hitflag( ctx, ctx->Current.RasterPos[2] ); - } -} - - -/* This is just to support the GL_MESA_window_pos version */ -static void -window_pos4f(GLfloat x, GLfloat y, GLfloat z, GLfloat w) -{ - GET_CURRENT_CONTEXT(ctx); - window_pos3f(x, y, z); - ctx->Current.RasterPos[3] = w; -} - - -static void GLAPIENTRY -_mesa_WindowPos2dMESA(GLdouble x, GLdouble y) -{ - window_pos4f((GLfloat) x, (GLfloat) y, 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos2fMESA(GLfloat x, GLfloat y) -{ - window_pos4f(x, y, 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos2iMESA(GLint x, GLint y) -{ - window_pos4f((GLfloat) x, (GLfloat) y, 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos2sMESA(GLshort x, GLshort y) -{ - window_pos4f(x, y, 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos3dMESA(GLdouble x, GLdouble y, GLdouble z) -{ - window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos3fMESA(GLfloat x, GLfloat y, GLfloat z) -{ - window_pos4f(x, y, z, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos3iMESA(GLint x, GLint y, GLint z) -{ - window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos3sMESA(GLshort x, GLshort y, GLshort z) -{ - window_pos4f(x, y, z, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos4dMESA(GLdouble x, GLdouble y, GLdouble z, GLdouble w) -{ - window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -_mesa_WindowPos4fMESA(GLfloat x, GLfloat y, GLfloat z, GLfloat w) -{ - window_pos4f(x, y, z, w); -} - -static void GLAPIENTRY -_mesa_WindowPos4iMESA(GLint x, GLint y, GLint z, GLint w) -{ - window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); -} - -static void GLAPIENTRY -_mesa_WindowPos4sMESA(GLshort x, GLshort y, GLshort z, GLshort w) -{ - window_pos4f(x, y, z, w); -} - -static void GLAPIENTRY -_mesa_WindowPos2dvMESA(const GLdouble *v) -{ - window_pos4f((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos2fvMESA(const GLfloat *v) -{ - window_pos4f(v[0], v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos2ivMESA(const GLint *v) -{ - window_pos4f((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos2svMESA(const GLshort *v) -{ - window_pos4f(v[0], v[1], 0.0F, 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos3dvMESA(const GLdouble *v) -{ - window_pos4f((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos3fvMESA(const GLfloat *v) -{ - window_pos4f(v[0], v[1], v[2], 1.0); -} - -static void GLAPIENTRY -_mesa_WindowPos3ivMESA(const GLint *v) -{ - window_pos4f((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos3svMESA(const GLshort *v) -{ - window_pos4f(v[0], v[1], v[2], 1.0F); -} - -static void GLAPIENTRY -_mesa_WindowPos4dvMESA(const GLdouble *v) -{ - window_pos4f((GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -_mesa_WindowPos4fvMESA(const GLfloat *v) -{ - window_pos4f(v[0], v[1], v[2], v[3]); -} - -static void GLAPIENTRY -_mesa_WindowPos4ivMESA(const GLint *v) -{ - window_pos4f((GLfloat) v[0], (GLfloat) v[1], - (GLfloat) v[2], (GLfloat) v[3]); -} - -static void GLAPIENTRY -_mesa_WindowPos4svMESA(const GLshort *v) -{ - window_pos4f(v[0], v[1], v[2], v[3]); -} - - -#if 0 - -/* - * OpenGL implementation of glWindowPos*MESA() - */ -void glWindowPos4fMESA( GLfloat x, GLfloat y, GLfloat z, GLfloat w ) -{ - GLfloat fx, fy; - - /* Push current matrix mode and viewport attributes */ - glPushAttrib( GL_TRANSFORM_BIT | GL_VIEWPORT_BIT ); - - /* Setup projection parameters */ - glMatrixMode( GL_PROJECTION ); - glPushMatrix(); - glLoadIdentity(); - glMatrixMode( GL_MODELVIEW ); - glPushMatrix(); - glLoadIdentity(); - - glDepthRange( z, z ); - glViewport( (int) x - 1, (int) y - 1, 2, 2 ); - - /* set the raster (window) position */ - fx = x - (int) x; - fy = y - (int) y; - glRasterPos4f( fx, fy, 0.0, w ); - - /* restore matrices, viewport and matrix mode */ - glPopMatrix(); - glMatrixMode( GL_PROJECTION ); - glPopMatrix(); - - glPopAttrib(); -} - -#endif - - -void -_mesa_init_rastpos_dispatch(struct _glapi_table *disp) -{ - SET_RasterPos2f(disp, _mesa_RasterPos2f); - SET_RasterPos2fv(disp, _mesa_RasterPos2fv); - SET_RasterPos2i(disp, _mesa_RasterPos2i); - SET_RasterPos2iv(disp, _mesa_RasterPos2iv); - SET_RasterPos2d(disp, _mesa_RasterPos2d); - SET_RasterPos2dv(disp, _mesa_RasterPos2dv); - SET_RasterPos2s(disp, _mesa_RasterPos2s); - SET_RasterPos2sv(disp, _mesa_RasterPos2sv); - SET_RasterPos3d(disp, _mesa_RasterPos3d); - SET_RasterPos3dv(disp, _mesa_RasterPos3dv); - SET_RasterPos3f(disp, _mesa_RasterPos3f); - SET_RasterPos3fv(disp, _mesa_RasterPos3fv); - SET_RasterPos3i(disp, _mesa_RasterPos3i); - SET_RasterPos3iv(disp, _mesa_RasterPos3iv); - SET_RasterPos3s(disp, _mesa_RasterPos3s); - SET_RasterPos3sv(disp, _mesa_RasterPos3sv); - SET_RasterPos4d(disp, _mesa_RasterPos4d); - SET_RasterPos4dv(disp, _mesa_RasterPos4dv); - SET_RasterPos4f(disp, _mesa_RasterPos4f); - SET_RasterPos4fv(disp, _mesa_RasterPos4fv); - SET_RasterPos4i(disp, _mesa_RasterPos4i); - SET_RasterPos4iv(disp, _mesa_RasterPos4iv); - SET_RasterPos4s(disp, _mesa_RasterPos4s); - SET_RasterPos4sv(disp, _mesa_RasterPos4sv); - - /* 197. GL_MESA_window_pos */ - SET_WindowPos2dMESA(disp, _mesa_WindowPos2dMESA); - SET_WindowPos2dvMESA(disp, _mesa_WindowPos2dvMESA); - SET_WindowPos2fMESA(disp, _mesa_WindowPos2fMESA); - SET_WindowPos2fvMESA(disp, _mesa_WindowPos2fvMESA); - SET_WindowPos2iMESA(disp, _mesa_WindowPos2iMESA); - SET_WindowPos2ivMESA(disp, _mesa_WindowPos2ivMESA); - SET_WindowPos2sMESA(disp, _mesa_WindowPos2sMESA); - SET_WindowPos2svMESA(disp, _mesa_WindowPos2svMESA); - SET_WindowPos3dMESA(disp, _mesa_WindowPos3dMESA); - SET_WindowPos3dvMESA(disp, _mesa_WindowPos3dvMESA); - SET_WindowPos3fMESA(disp, _mesa_WindowPos3fMESA); - SET_WindowPos3fvMESA(disp, _mesa_WindowPos3fvMESA); - SET_WindowPos3iMESA(disp, _mesa_WindowPos3iMESA); - SET_WindowPos3ivMESA(disp, _mesa_WindowPos3ivMESA); - SET_WindowPos3sMESA(disp, _mesa_WindowPos3sMESA); - SET_WindowPos3svMESA(disp, _mesa_WindowPos3svMESA); - SET_WindowPos4dMESA(disp, _mesa_WindowPos4dMESA); - SET_WindowPos4dvMESA(disp, _mesa_WindowPos4dvMESA); - SET_WindowPos4fMESA(disp, _mesa_WindowPos4fMESA); - SET_WindowPos4fvMESA(disp, _mesa_WindowPos4fvMESA); - SET_WindowPos4iMESA(disp, _mesa_WindowPos4iMESA); - SET_WindowPos4ivMESA(disp, _mesa_WindowPos4ivMESA); - SET_WindowPos4sMESA(disp, _mesa_WindowPos4sMESA); - SET_WindowPos4svMESA(disp, _mesa_WindowPos4svMESA); -} - - -#endif /* FEATURE_rastpos */ - - -/**********************************************************************/ -/** \name Initialization */ -/**********************************************************************/ -/*@{*/ - -/** - * Initialize the context current raster position information. - * - * \param ctx GL context. - * - * Initialize the current raster position information in - * __struct gl_contextRec::Current, and adds the extension entry points to the - * dispatcher. - */ -void _mesa_init_rastpos( struct gl_context * ctx ) -{ - int i; - - ASSIGN_4V( ctx->Current.RasterPos, 0.0, 0.0, 0.0, 1.0 ); - ctx->Current.RasterDistance = 0.0; - ASSIGN_4V( ctx->Current.RasterColor, 1.0, 1.0, 1.0, 1.0 ); - ASSIGN_4V( ctx->Current.RasterSecondaryColor, 0.0, 0.0, 0.0, 1.0 ); - for (i = 0; i < Elements(ctx->Current.RasterTexCoords); i++) - ASSIGN_4V( ctx->Current.RasterTexCoords[i], 0.0, 0.0, 0.0, 1.0 ); - ctx->Current.RasterPosValid = GL_TRUE; -} - -/*@}*/ +/* + * Mesa 3-D graphics library + * Version: 6.5.3 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +/** + * \file rastpos.c + * Raster position operations. + */ + +#include "glheader.h" +#include "context.h" +#include "feedback.h" +#include "macros.h" +#include "mfeatures.h" +#include "mtypes.h" +#include "rastpos.h" +#include "state.h" +#include "main/dispatch.h" + + +#if FEATURE_rastpos + + +/** + * Helper function for all the RasterPos functions. + */ +static void +rasterpos(GLfloat x, GLfloat y, GLfloat z, GLfloat w) +{ + GET_CURRENT_CONTEXT(ctx); + GLfloat p[4]; + + p[0] = x; + p[1] = y; + p[2] = z; + p[3] = w; + + ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); + FLUSH_CURRENT(ctx, 0); + + if (ctx->NewState) + _mesa_update_state( ctx ); + + ctx->Driver.RasterPos(ctx, p); +} + + +static void GLAPIENTRY +_mesa_RasterPos2d(GLdouble x, GLdouble y) +{ + rasterpos((GLfloat)x, (GLfloat)y, (GLfloat)0.0, (GLfloat)1.0); +} + +static void GLAPIENTRY +_mesa_RasterPos2f(GLfloat x, GLfloat y) +{ + rasterpos(x, y, 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos2i(GLint x, GLint y) +{ + rasterpos((GLfloat) x, (GLfloat) y, 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos2s(GLshort x, GLshort y) +{ + rasterpos(x, y, 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3d(GLdouble x, GLdouble y, GLdouble z) +{ + rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3f(GLfloat x, GLfloat y, GLfloat z) +{ + rasterpos(x, y, z, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3i(GLint x, GLint y, GLint z) +{ + rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3s(GLshort x, GLshort y, GLshort z) +{ + rasterpos(x, y, z, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos4d(GLdouble x, GLdouble y, GLdouble z, GLdouble w) +{ + rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +_mesa_RasterPos4f(GLfloat x, GLfloat y, GLfloat z, GLfloat w) +{ + rasterpos(x, y, z, w); +} + +static void GLAPIENTRY +_mesa_RasterPos4i(GLint x, GLint y, GLint z, GLint w) +{ + rasterpos((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +_mesa_RasterPos4s(GLshort x, GLshort y, GLshort z, GLshort w) +{ + rasterpos(x, y, z, w); +} + +static void GLAPIENTRY +_mesa_RasterPos2dv(const GLdouble *v) +{ + rasterpos((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos2fv(const GLfloat *v) +{ + rasterpos(v[0], v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos2iv(const GLint *v) +{ + rasterpos((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos2sv(const GLshort *v) +{ + rasterpos(v[0], v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3dv(const GLdouble *v) +{ + rasterpos((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3fv(const GLfloat *v) +{ + rasterpos(v[0], v[1], v[2], 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3iv(const GLint *v) +{ + rasterpos((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos3sv(const GLshort *v) +{ + rasterpos(v[0], v[1], v[2], 1.0F); +} + +static void GLAPIENTRY +_mesa_RasterPos4dv(const GLdouble *v) +{ + rasterpos((GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +_mesa_RasterPos4fv(const GLfloat *v) +{ + rasterpos(v[0], v[1], v[2], v[3]); +} + +static void GLAPIENTRY +_mesa_RasterPos4iv(const GLint *v) +{ + rasterpos((GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +_mesa_RasterPos4sv(const GLshort *v) +{ + rasterpos(v[0], v[1], v[2], v[3]); +} + + +/**********************************************************************/ +/*** GL_ARB_window_pos / GL_MESA_window_pos ***/ +/**********************************************************************/ + + +/** + * All glWindowPosMESA and glWindowPosARB commands call this function to + * update the current raster position. + */ +static void +window_pos3f(GLfloat x, GLfloat y, GLfloat z) +{ + GET_CURRENT_CONTEXT(ctx); + GLfloat z2; + + ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); + FLUSH_CURRENT(ctx, 0); + + z2 = CLAMP(z, 0.0F, 1.0F) * (ctx->Viewport.Far - ctx->Viewport.Near) + + ctx->Viewport.Near; + + /* set raster position */ + ctx->Current.RasterPos[0] = x; + ctx->Current.RasterPos[1] = y; + ctx->Current.RasterPos[2] = z2; + ctx->Current.RasterPos[3] = 1.0F; + + ctx->Current.RasterPosValid = GL_TRUE; + + if (ctx->Fog.FogCoordinateSource == GL_FOG_COORDINATE_EXT) + ctx->Current.RasterDistance = ctx->Current.Attrib[VERT_ATTRIB_FOG][0]; + else + ctx->Current.RasterDistance = 0.0; + + /* raster color = current color or index */ + ctx->Current.RasterColor[0] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][0], 0.0F, 1.0F); + ctx->Current.RasterColor[1] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][1], 0.0F, 1.0F); + ctx->Current.RasterColor[2] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][2], 0.0F, 1.0F); + ctx->Current.RasterColor[3] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR0][3], 0.0F, 1.0F); + ctx->Current.RasterSecondaryColor[0] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][0], 0.0F, 1.0F); + ctx->Current.RasterSecondaryColor[1] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][1], 0.0F, 1.0F); + ctx->Current.RasterSecondaryColor[2] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][2], 0.0F, 1.0F); + ctx->Current.RasterSecondaryColor[3] + = CLAMP(ctx->Current.Attrib[VERT_ATTRIB_COLOR1][3], 0.0F, 1.0F); + + /* raster texcoord = current texcoord */ + { + GLuint texSet; + for (texSet = 0; texSet < ctx->Const.MaxTextureCoordUnits; texSet++) { + assert(texSet < Elements(ctx->Current.RasterTexCoords)); + COPY_4FV( ctx->Current.RasterTexCoords[texSet], + ctx->Current.Attrib[VERT_ATTRIB_TEX0 + texSet] ); + } + } + + if (ctx->RenderMode==GL_SELECT) { + _mesa_update_hitflag( ctx, ctx->Current.RasterPos[2] ); + } +} + + +/* This is just to support the GL_MESA_window_pos version */ +static void +window_pos4f(GLfloat x, GLfloat y, GLfloat z, GLfloat w) +{ + GET_CURRENT_CONTEXT(ctx); + window_pos3f(x, y, z); + ctx->Current.RasterPos[3] = w; +} + + +static void GLAPIENTRY +_mesa_WindowPos2dMESA(GLdouble x, GLdouble y) +{ + window_pos4f((GLfloat) x, (GLfloat) y, 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos2fMESA(GLfloat x, GLfloat y) +{ + window_pos4f(x, y, 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos2iMESA(GLint x, GLint y) +{ + window_pos4f((GLfloat) x, (GLfloat) y, 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos2sMESA(GLshort x, GLshort y) +{ + window_pos4f(x, y, 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos3dMESA(GLdouble x, GLdouble y, GLdouble z) +{ + window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos3fMESA(GLfloat x, GLfloat y, GLfloat z) +{ + window_pos4f(x, y, z, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos3iMESA(GLint x, GLint y, GLint z) +{ + window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos3sMESA(GLshort x, GLshort y, GLshort z) +{ + window_pos4f(x, y, z, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos4dMESA(GLdouble x, GLdouble y, GLdouble z, GLdouble w) +{ + window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +_mesa_WindowPos4fMESA(GLfloat x, GLfloat y, GLfloat z, GLfloat w) +{ + window_pos4f(x, y, z, w); +} + +static void GLAPIENTRY +_mesa_WindowPos4iMESA(GLint x, GLint y, GLint z, GLint w) +{ + window_pos4f((GLfloat) x, (GLfloat) y, (GLfloat) z, (GLfloat) w); +} + +static void GLAPIENTRY +_mesa_WindowPos4sMESA(GLshort x, GLshort y, GLshort z, GLshort w) +{ + window_pos4f(x, y, z, w); +} + +static void GLAPIENTRY +_mesa_WindowPos2dvMESA(const GLdouble *v) +{ + window_pos4f((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos2fvMESA(const GLfloat *v) +{ + window_pos4f(v[0], v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos2ivMESA(const GLint *v) +{ + window_pos4f((GLfloat) v[0], (GLfloat) v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos2svMESA(const GLshort *v) +{ + window_pos4f(v[0], v[1], 0.0F, 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos3dvMESA(const GLdouble *v) +{ + window_pos4f((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos3fvMESA(const GLfloat *v) +{ + window_pos4f(v[0], v[1], v[2], 1.0); +} + +static void GLAPIENTRY +_mesa_WindowPos3ivMESA(const GLint *v) +{ + window_pos4f((GLfloat) v[0], (GLfloat) v[1], (GLfloat) v[2], 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos3svMESA(const GLshort *v) +{ + window_pos4f(v[0], v[1], v[2], 1.0F); +} + +static void GLAPIENTRY +_mesa_WindowPos4dvMESA(const GLdouble *v) +{ + window_pos4f((GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +_mesa_WindowPos4fvMESA(const GLfloat *v) +{ + window_pos4f(v[0], v[1], v[2], v[3]); +} + +static void GLAPIENTRY +_mesa_WindowPos4ivMESA(const GLint *v) +{ + window_pos4f((GLfloat) v[0], (GLfloat) v[1], + (GLfloat) v[2], (GLfloat) v[3]); +} + +static void GLAPIENTRY +_mesa_WindowPos4svMESA(const GLshort *v) +{ + window_pos4f(v[0], v[1], v[2], v[3]); +} + + +#if 0 + +/* + * OpenGL implementation of glWindowPos*MESA() + */ +void glWindowPos4fMESA( GLfloat x, GLfloat y, GLfloat z, GLfloat w ) +{ + GLfloat fx, fy; + + /* Push current matrix mode and viewport attributes */ + glPushAttrib( GL_TRANSFORM_BIT | GL_VIEWPORT_BIT ); + + /* Setup projection parameters */ + glMatrixMode( GL_PROJECTION ); + glPushMatrix(); + glLoadIdentity(); + glMatrixMode( GL_MODELVIEW ); + glPushMatrix(); + glLoadIdentity(); + + glDepthRange( z, z ); + glViewport( (int) x - 1, (int) y - 1, 2, 2 ); + + /* set the raster (window) position */ + fx = x - (int) x; + fy = y - (int) y; + glRasterPos4f( fx, fy, 0.0, w ); + + /* restore matrices, viewport and matrix mode */ + glPopMatrix(); + glMatrixMode( GL_PROJECTION ); + glPopMatrix(); + + glPopAttrib(); +} + +#endif + + +void +_mesa_init_rastpos_dispatch(struct _glapi_table *disp) +{ + SET_RasterPos2f(disp, _mesa_RasterPos2f); + SET_RasterPos2fv(disp, _mesa_RasterPos2fv); + SET_RasterPos2i(disp, _mesa_RasterPos2i); + SET_RasterPos2iv(disp, _mesa_RasterPos2iv); + SET_RasterPos2d(disp, _mesa_RasterPos2d); + SET_RasterPos2dv(disp, _mesa_RasterPos2dv); + SET_RasterPos2s(disp, _mesa_RasterPos2s); + SET_RasterPos2sv(disp, _mesa_RasterPos2sv); + SET_RasterPos3d(disp, _mesa_RasterPos3d); + SET_RasterPos3dv(disp, _mesa_RasterPos3dv); + SET_RasterPos3f(disp, _mesa_RasterPos3f); + SET_RasterPos3fv(disp, _mesa_RasterPos3fv); + SET_RasterPos3i(disp, _mesa_RasterPos3i); + SET_RasterPos3iv(disp, _mesa_RasterPos3iv); + SET_RasterPos3s(disp, _mesa_RasterPos3s); + SET_RasterPos3sv(disp, _mesa_RasterPos3sv); + SET_RasterPos4d(disp, _mesa_RasterPos4d); + SET_RasterPos4dv(disp, _mesa_RasterPos4dv); + SET_RasterPos4f(disp, _mesa_RasterPos4f); + SET_RasterPos4fv(disp, _mesa_RasterPos4fv); + SET_RasterPos4i(disp, _mesa_RasterPos4i); + SET_RasterPos4iv(disp, _mesa_RasterPos4iv); + SET_RasterPos4s(disp, _mesa_RasterPos4s); + SET_RasterPos4sv(disp, _mesa_RasterPos4sv); + + /* 197. GL_MESA_window_pos */ + SET_WindowPos2dMESA(disp, _mesa_WindowPos2dMESA); + SET_WindowPos2dvMESA(disp, _mesa_WindowPos2dvMESA); + SET_WindowPos2fMESA(disp, _mesa_WindowPos2fMESA); + SET_WindowPos2fvMESA(disp, _mesa_WindowPos2fvMESA); + SET_WindowPos2iMESA(disp, _mesa_WindowPos2iMESA); + SET_WindowPos2ivMESA(disp, _mesa_WindowPos2ivMESA); + SET_WindowPos2sMESA(disp, _mesa_WindowPos2sMESA); + SET_WindowPos2svMESA(disp, _mesa_WindowPos2svMESA); + SET_WindowPos3dMESA(disp, _mesa_WindowPos3dMESA); + SET_WindowPos3dvMESA(disp, _mesa_WindowPos3dvMESA); + SET_WindowPos3fMESA(disp, _mesa_WindowPos3fMESA); + SET_WindowPos3fvMESA(disp, _mesa_WindowPos3fvMESA); + SET_WindowPos3iMESA(disp, _mesa_WindowPos3iMESA); + SET_WindowPos3ivMESA(disp, _mesa_WindowPos3ivMESA); + SET_WindowPos3sMESA(disp, _mesa_WindowPos3sMESA); + SET_WindowPos3svMESA(disp, _mesa_WindowPos3svMESA); + SET_WindowPos4dMESA(disp, _mesa_WindowPos4dMESA); + SET_WindowPos4dvMESA(disp, _mesa_WindowPos4dvMESA); + SET_WindowPos4fMESA(disp, _mesa_WindowPos4fMESA); + SET_WindowPos4fvMESA(disp, _mesa_WindowPos4fvMESA); + SET_WindowPos4iMESA(disp, _mesa_WindowPos4iMESA); + SET_WindowPos4ivMESA(disp, _mesa_WindowPos4ivMESA); + SET_WindowPos4sMESA(disp, _mesa_WindowPos4sMESA); + SET_WindowPos4svMESA(disp, _mesa_WindowPos4svMESA); +} + + +#endif /* FEATURE_rastpos */ + + +/**********************************************************************/ +/** \name Initialization */ +/**********************************************************************/ +/*@{*/ + +/** + * Initialize the context current raster position information. + * + * \param ctx GL context. + * + * Initialize the current raster position information in + * __struct gl_contextRec::Current, and adds the extension entry points to the + * dispatcher. + */ +void _mesa_init_rastpos( struct gl_context * ctx ) +{ + int i; + + ASSIGN_4V( ctx->Current.RasterPos, 0.0, 0.0, 0.0, 1.0 ); + ctx->Current.RasterDistance = 0.0; + ASSIGN_4V( ctx->Current.RasterColor, 1.0, 1.0, 1.0, 1.0 ); + ASSIGN_4V( ctx->Current.RasterSecondaryColor, 0.0, 0.0, 0.0, 1.0 ); + for (i = 0; i < Elements(ctx->Current.RasterTexCoords); i++) + ASSIGN_4V( ctx->Current.RasterTexCoords[i], 0.0, 0.0, 0.0, 1.0 ); + ctx->Current.RasterPosValid = GL_TRUE; +} + +/*@}*/ diff --git a/mesalib/src/mesa/main/remap.c b/mesalib/src/mesa/main/remap.c index 6da31deb3..c89fba453 100644 --- a/mesalib/src/mesa/main/remap.c +++ b/mesalib/src/mesa/main/remap.c @@ -1,226 +1,226 @@ -/* - * Mesa 3-D graphics library - * Version: 7.7 - * - * Copyright (C) 2009 Chia-I Wu - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - - -/** - * \file remap.c - * Remap table management. - * - * Entries in the dispatch table are either static or dynamic. The - * dispatch table is shared by mesa core and glapi. When they are - * built separately, it is possible that a static entry in mesa core - * is dynamic, or assigned a different static offset, in glapi. The - * remap table is in charge of mapping a static entry in mesa core to - * a dynamic entry, or the corresponding static entry, in glapi. - */ - -#include "mfeatures.h" - -#if FEATURE_remap_table - -#include "remap.h" -#include "imports.h" -#include "glapi/glapi.h" - -#define MAX_ENTRY_POINTS 16 - -#define need_MESA_remap_table -#include "main/remap_helper.h" - - -/* this is global for quick access */ -int driDispatchRemapTable[driDispatchRemapTable_size]; - - -/** - * Return the spec string associated with the given function index. - * The index is available from including remap_helper.h. - * - * \param func_index an opaque function index. - * - * \return the spec string associated with the function index, or NULL. - */ -const char * -_mesa_get_function_spec(GLint func_index) -{ - if (func_index < Elements(_mesa_function_pool)) - return _mesa_function_pool + func_index; - else - return NULL; -} - - -/** - * Map a function by its spec. The function will be added to glapi, - * and the dispatch offset will be returned. - * - * \param spec a '\0'-separated string array specifying a function. - * It begins with the parameter signature of the function, - * followed by the names of the entry points. An empty entry - * point name terminates the array. - * - * \return the offset of the (re-)mapped function in the dispatch - * table, or -1. - */ -GLint -_mesa_map_function_spec(const char *spec) -{ - const char *signature; - const char *names[MAX_ENTRY_POINTS + 1]; - GLint num_names = 0; - - if (!spec) - return -1; - - signature = spec; - spec += strlen(spec) + 1; - - /* spec is terminated by an empty string */ - while (*spec) { - names[num_names] = spec; - num_names++; - if (num_names >= MAX_ENTRY_POINTS) - break; - spec += strlen(spec) + 1; - } - if (!num_names) - return -1; - - names[num_names] = NULL; - - /* add the entry points to the dispatch table */ - return _glapi_add_dispatch(names, signature); -} - - -/** - * Map an array of functions. This is a convenient function for - * use with arrays available from including remap_helper.h. - * - * Note that the dispatch offsets of the functions are not returned. - * If they are needed, _mesa_map_function_spec() should be used. - * - * \param func_array an array of function remaps. - */ -void -_mesa_map_function_array(const struct gl_function_remap *func_array) -{ - GLint i; - - if (!func_array) - return; - - for (i = 0; func_array[i].func_index != -1; i++) { - const char *spec; - GLint offset; - - spec = _mesa_get_function_spec(func_array[i].func_index); - if (!spec) { - _mesa_problem(NULL, "invalid function index %d", - func_array[i].func_index); - continue; - } - - offset = _mesa_map_function_spec(spec); - /* error checks */ - if (offset < 0) { - const char *name = spec + strlen(spec) + 1; - _mesa_warning(NULL, "failed to remap %s", name); - } - else if (func_array[i].dispatch_offset >= 0 && - offset != func_array[i].dispatch_offset) { - const char *name = spec + strlen(spec) + 1; - _mesa_problem(NULL, "%s should be mapped to %d, not %d", - name, func_array[i].dispatch_offset, offset); - } - } -} - - -/** - * Map the functions which are already static. - * - * When a extension function are incorporated into the ABI, the - * extension suffix is usually stripped. Mapping such functions - * makes sure the alternative names are available. - * - * Note that functions mapped by _mesa_init_remap_table() are - * excluded. - */ -void -_mesa_map_static_functions(void) -{ - /* Remap static functions which have alternative names and are in the ABI. - * This is to be on the safe side. glapi should have defined those names. - */ - _mesa_map_function_array(MESA_alt_functions); -} - - -/** - * Initialize the remap table. This is called in one_time_init(). - * The remap table needs to be initialized before calling the - * CALL/GET/SET macros defined in main/dispatch.h. - */ -static void -_mesa_do_init_remap_table(const char *pool, - int size, - const struct gl_function_pool_remap *remap) -{ - static GLboolean initialized = GL_FALSE; - GLint i; - - if (initialized) - return; - initialized = GL_TRUE; - - /* initialize the remap table */ - for (i = 0; i < size; i++) { - GLint offset; - const char *spec; - - /* sanity check */ - ASSERT(i == remap[i].remap_index); - spec = _mesa_function_pool + remap[i].pool_index; - - offset = _mesa_map_function_spec(spec); - /* store the dispatch offset in the remap table */ - driDispatchRemapTable[i] = offset; - if (offset < 0) - _mesa_warning(NULL, "failed to remap index %d", i); - } -} - - -void -_mesa_init_remap_table(void) -{ - _mesa_do_init_remap_table(_mesa_function_pool, - driDispatchRemapTable_size, - MESA_remap_table_functions); -} - - -#endif /* FEATURE_remap_table */ +/* + * Mesa 3-D graphics library + * Version: 7.7 + * + * Copyright (C) 2009 Chia-I Wu + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ + + +/** + * \file remap.c + * Remap table management. + * + * Entries in the dispatch table are either static or dynamic. The + * dispatch table is shared by mesa core and glapi. When they are + * built separately, it is possible that a static entry in mesa core + * is dynamic, or assigned a different static offset, in glapi. The + * remap table is in charge of mapping a static entry in mesa core to + * a dynamic entry, or the corresponding static entry, in glapi. + */ + +#include "mfeatures.h" + +#if FEATURE_remap_table + +#include "remap.h" +#include "imports.h" +#include "glapi/glapi.h" + +#define MAX_ENTRY_POINTS 16 + +#define need_MESA_remap_table +#include "main/remap_helper.h" + + +/* this is global for quick access */ +int driDispatchRemapTable[driDispatchRemapTable_size]; + + +/** + * Return the spec string associated with the given function index. + * The index is available from including remap_helper.h. + * + * \param func_index an opaque function index. + * + * \return the spec string associated with the function index, or NULL. + */ +const char * +_mesa_get_function_spec(GLint func_index) +{ + if (func_index < Elements(_mesa_function_pool)) + return _mesa_function_pool + func_index; + else + return NULL; +} + + +/** + * Map a function by its spec. The function will be added to glapi, + * and the dispatch offset will be returned. + * + * \param spec a '\0'-separated string array specifying a function. + * It begins with the parameter signature of the function, + * followed by the names of the entry points. An empty entry + * point name terminates the array. + * + * \return the offset of the (re-)mapped function in the dispatch + * table, or -1. + */ +GLint +_mesa_map_function_spec(const char *spec) +{ + const char *signature; + const char *names[MAX_ENTRY_POINTS + 1]; + GLint num_names = 0; + + if (!spec) + return -1; + + signature = spec; + spec += strlen(spec) + 1; + + /* spec is terminated by an empty string */ + while (*spec) { + names[num_names] = spec; + num_names++; + if (num_names >= MAX_ENTRY_POINTS) + break; + spec += strlen(spec) + 1; + } + if (!num_names) + return -1; + + names[num_names] = NULL; + + /* add the entry points to the dispatch table */ + return _glapi_add_dispatch(names, signature); +} + + +/** + * Map an array of functions. This is a convenient function for + * use with arrays available from including remap_helper.h. + * + * Note that the dispatch offsets of the functions are not returned. + * If they are needed, _mesa_map_function_spec() should be used. + * + * \param func_array an array of function remaps. + */ +void +_mesa_map_function_array(const struct gl_function_remap *func_array) +{ + GLint i; + + if (!func_array) + return; + + for (i = 0; func_array[i].func_index != -1; i++) { + const char *spec; + GLint offset; + + spec = _mesa_get_function_spec(func_array[i].func_index); + if (!spec) { + _mesa_problem(NULL, "invalid function index %d", + func_array[i].func_index); + continue; + } + + offset = _mesa_map_function_spec(spec); + /* error checks */ + if (offset < 0) { + const char *name = spec + strlen(spec) + 1; + _mesa_warning(NULL, "failed to remap %s", name); + } + else if (func_array[i].dispatch_offset >= 0 && + offset != func_array[i].dispatch_offset) { + const char *name = spec + strlen(spec) + 1; + _mesa_problem(NULL, "%s should be mapped to %d, not %d", + name, func_array[i].dispatch_offset, offset); + } + } +} + + +/** + * Map the functions which are already static. + * + * When a extension function are incorporated into the ABI, the + * extension suffix is usually stripped. Mapping such functions + * makes sure the alternative names are available. + * + * Note that functions mapped by _mesa_init_remap_table() are + * excluded. + */ +void +_mesa_map_static_functions(void) +{ + /* Remap static functions which have alternative names and are in the ABI. + * This is to be on the safe side. glapi should have defined those names. + */ + _mesa_map_function_array(MESA_alt_functions); +} + + +/** + * Initialize the remap table. This is called in one_time_init(). + * The remap table needs to be initialized before calling the + * CALL/GET/SET macros defined in main/dispatch.h. + */ +static void +_mesa_do_init_remap_table(const char *pool, + int size, + const struct gl_function_pool_remap *remap) +{ + static GLboolean initialized = GL_FALSE; + GLint i; + + if (initialized) + return; + initialized = GL_TRUE; + + /* initialize the remap table */ + for (i = 0; i < size; i++) { + GLint offset; + const char *spec; + + /* sanity check */ + ASSERT(i == remap[i].remap_index); + spec = _mesa_function_pool + remap[i].pool_index; + + offset = _mesa_map_function_spec(spec); + /* store the dispatch offset in the remap table */ + driDispatchRemapTable[i] = offset; + if (offset < 0) + _mesa_warning(NULL, "failed to remap index %d", i); + } +} + + +void +_mesa_init_remap_table(void) +{ + _mesa_do_init_remap_table(_mesa_function_pool, + driDispatchRemapTable_size, + MESA_remap_table_functions); +} + + +#endif /* FEATURE_remap_table */ diff --git a/mesalib/src/mesa/main/scissor.c b/mesalib/src/mesa/main/scissor.c index 16e8438c5..df665994a 100644 --- a/mesalib/src/mesa/main/scissor.c +++ b/mesalib/src/mesa/main/scissor.c @@ -1,100 +1,100 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#include "main/glheader.h" -#include "main/context.h" -#include "main/mtypes.h" -#include "main/scissor.h" - - -/** - * Called via glScissor - */ -void GLAPIENTRY -_mesa_Scissor( GLint x, GLint y, GLsizei width, GLsizei height ) -{ - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (MESA_VERBOSE & VERBOSE_API) - _mesa_debug(ctx, "glScissor %d %d %d %d\n", x, y, width, height); - - if (width < 0 || height < 0) { - _mesa_error( ctx, GL_INVALID_VALUE, "glScissor" ); - return; - } - - _mesa_set_scissor(ctx, x, y, width, height); -} - - -/** - * Define the scissor box. - * - * \param x, y coordinates of the scissor box lower-left corner. - * \param width width of the scissor box. - * \param height height of the scissor box. - * - * \sa glScissor(). - * - * Verifies the parameters and updates __struct gl_contextRec::Scissor. On a - * change flushes the vertices and notifies the driver via - * the dd_function_table::Scissor callback. - */ -void -_mesa_set_scissor(struct gl_context *ctx, - GLint x, GLint y, GLsizei width, GLsizei height) -{ - if (x == ctx->Scissor.X && - y == ctx->Scissor.Y && - width == ctx->Scissor.Width && - height == ctx->Scissor.Height) - return; - - FLUSH_VERTICES(ctx, _NEW_SCISSOR); - ctx->Scissor.X = x; - ctx->Scissor.Y = y; - ctx->Scissor.Width = width; - ctx->Scissor.Height = height; - - if (ctx->Driver.Scissor) - ctx->Driver.Scissor( ctx, x, y, width, height ); -} - - -/** - * Initialize the context's scissor state. - * \param ctx the GL context. - */ -void -_mesa_init_scissor(struct gl_context *ctx) -{ - /* Scissor group */ - ctx->Scissor.Enabled = GL_FALSE; - ctx->Scissor.X = 0; - ctx->Scissor.Y = 0; - ctx->Scissor.Width = 0; - ctx->Scissor.Height = 0; -} +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#include "main/glheader.h" +#include "main/context.h" +#include "main/mtypes.h" +#include "main/scissor.h" + + +/** + * Called via glScissor + */ +void GLAPIENTRY +_mesa_Scissor( GLint x, GLint y, GLsizei width, GLsizei height ) +{ + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (MESA_VERBOSE & VERBOSE_API) + _mesa_debug(ctx, "glScissor %d %d %d %d\n", x, y, width, height); + + if (width < 0 || height < 0) { + _mesa_error( ctx, GL_INVALID_VALUE, "glScissor" ); + return; + } + + _mesa_set_scissor(ctx, x, y, width, height); +} + + +/** + * Define the scissor box. + * + * \param x, y coordinates of the scissor box lower-left corner. + * \param width width of the scissor box. + * \param height height of the scissor box. + * + * \sa glScissor(). + * + * Verifies the parameters and updates __struct gl_contextRec::Scissor. On a + * change flushes the vertices and notifies the driver via + * the dd_function_table::Scissor callback. + */ +void +_mesa_set_scissor(struct gl_context *ctx, + GLint x, GLint y, GLsizei width, GLsizei height) +{ + if (x == ctx->Scissor.X && + y == ctx->Scissor.Y && + width == ctx->Scissor.Width && + height == ctx->Scissor.Height) + return; + + FLUSH_VERTICES(ctx, _NEW_SCISSOR); + ctx->Scissor.X = x; + ctx->Scissor.Y = y; + ctx->Scissor.Width = width; + ctx->Scissor.Height = height; + + if (ctx->Driver.Scissor) + ctx->Driver.Scissor( ctx, x, y, width, height ); +} + + +/** + * Initialize the context's scissor state. + * \param ctx the GL context. + */ +void +_mesa_init_scissor(struct gl_context *ctx) +{ + /* Scissor group */ + ctx->Scissor.Enabled = GL_FALSE; + ctx->Scissor.X = 0; + ctx->Scissor.Y = 0; + ctx->Scissor.Width = 0; + ctx->Scissor.Height = 0; +} diff --git a/mesalib/src/mesa/main/scissor.h b/mesalib/src/mesa/main/scissor.h index 8735ec0cd..da9385e2b 100644 --- a/mesalib/src/mesa/main/scissor.h +++ b/mesalib/src/mesa/main/scissor.h @@ -1,47 +1,47 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef SCISSOR_H -#define SCISSOR_H - - -#include "glheader.h" - -struct gl_context; - -extern void GLAPIENTRY -_mesa_Scissor( GLint x, GLint y, GLsizei width, GLsizei height ); - - -extern void -_mesa_set_scissor(struct gl_context *ctx, - GLint x, GLint y, GLsizei width, GLsizei height); - - -extern void -_mesa_init_scissor(struct gl_context *ctx); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef SCISSOR_H +#define SCISSOR_H + + +#include "glheader.h" + +struct gl_context; + +extern void GLAPIENTRY +_mesa_Scissor( GLint x, GLint y, GLsizei width, GLsizei height ); + + +extern void +_mesa_set_scissor(struct gl_context *ctx, + GLint x, GLint y, GLsizei width, GLsizei height); + + +extern void +_mesa_init_scissor(struct gl_context *ctx); + + +#endif diff --git a/mesalib/src/mesa/main/stencil.h b/mesalib/src/mesa/main/stencil.h index f3a33db70..b772bfd6e 100644 --- a/mesalib/src/mesa/main/stencil.h +++ b/mesalib/src/mesa/main/stencil.h @@ -1,81 +1,81 @@ -/** - * \file stencil.h - * Stencil operations. - */ - -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef STENCIL_H -#define STENCIL_H - - -#include "glheader.h" - -struct gl_context; - -extern void GLAPIENTRY -_mesa_ClearStencil( GLint s ); - - -extern void GLAPIENTRY -_mesa_StencilFunc( GLenum func, GLint ref, GLuint mask ); - - -extern void GLAPIENTRY -_mesa_StencilMask( GLuint mask ); - - -extern void GLAPIENTRY -_mesa_StencilOp( GLenum fail, GLenum zfail, GLenum zpass ); - - -extern void GLAPIENTRY -_mesa_ActiveStencilFaceEXT(GLenum face); - - -extern void GLAPIENTRY -_mesa_StencilOpSeparate(GLenum face, GLenum fail, GLenum zfail, GLenum zpass); - - -extern void GLAPIENTRY -_mesa_StencilFuncSeparate(GLenum face, GLenum func, GLint ref, GLuint mask); - - -extern void GLAPIENTRY -_mesa_StencilFuncSeparateATI(GLenum frontfunc, GLenum backfunc, GLint ref, GLuint mask); - -extern void GLAPIENTRY -_mesa_StencilMaskSeparate(GLenum face, GLuint mask); - - -extern void -_mesa_update_stencil(struct gl_context *ctx); - - -extern void -_mesa_init_stencil( struct gl_context * ctx ); - -#endif +/** + * \file stencil.h + * Stencil operations. + */ + +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef STENCIL_H +#define STENCIL_H + + +#include "glheader.h" + +struct gl_context; + +extern void GLAPIENTRY +_mesa_ClearStencil( GLint s ); + + +extern void GLAPIENTRY +_mesa_StencilFunc( GLenum func, GLint ref, GLuint mask ); + + +extern void GLAPIENTRY +_mesa_StencilMask( GLuint mask ); + + +extern void GLAPIENTRY +_mesa_StencilOp( GLenum fail, GLenum zfail, GLenum zpass ); + + +extern void GLAPIENTRY +_mesa_ActiveStencilFaceEXT(GLenum face); + + +extern void GLAPIENTRY +_mesa_StencilOpSeparate(GLenum face, GLenum fail, GLenum zfail, GLenum zpass); + + +extern void GLAPIENTRY +_mesa_StencilFuncSeparate(GLenum face, GLenum func, GLint ref, GLuint mask); + + +extern void GLAPIENTRY +_mesa_StencilFuncSeparateATI(GLenum frontfunc, GLenum backfunc, GLint ref, GLuint mask); + +extern void GLAPIENTRY +_mesa_StencilMaskSeparate(GLenum face, GLuint mask); + + +extern void +_mesa_update_stencil(struct gl_context *ctx); + + +extern void +_mesa_init_stencil( struct gl_context * ctx ); + +#endif diff --git a/mesalib/src/mesa/main/syncobj.c b/mesalib/src/mesa/main/syncobj.c index 676038430..e1a5c6c0f 100644 --- a/mesalib/src/mesa/main/syncobj.c +++ b/mesalib/src/mesa/main/syncobj.c @@ -1,426 +1,426 @@ -/* - * Copyright © 2009 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -/** - * \file syncobj.c - * Sync object management. - * - * Unlike textures and other objects that are shared between contexts, sync - * objects are not bound to the context. As a result, the reference counting - * and delete behavior of sync objects is slightly different. References to - * sync objects are added: - * - * - By \c glFencSynce. This sets the initial reference count to 1. - * - At the start of \c glClientWaitSync. The reference is held for the - * duration of the wait call. - * - * References are removed: - * - * - By \c glDeleteSync. - * - At the end of \c glClientWaitSync. - * - * Additionally, drivers may call \c _mesa_ref_sync_object and - * \c _mesa_unref_sync_object as needed to implement \c ServerWaitSync. - * - * As with shader objects, sync object names become invalid as soon as - * \c glDeleteSync is called. For this reason \c glDeleteSync sets the - * \c DeletePending flag. All functions validate object handles by testing - * this flag. - * - * \note - * Only \c GL_ARB_sync objects are shared between contexts. If support is ever - * added for either \c GL_NV_fence or \c GL_APPLE_fence different semantics - * will need to be implemented. - * - * \author Ian Romanick - */ - -#include "glheader.h" -#include "imports.h" -#include "context.h" -#include "macros.h" -#include "mfeatures.h" -#include "get.h" -#include "dispatch.h" -#include "mtypes.h" - -#if FEATURE_ARB_sync -#include "syncobj.h" - -static struct gl_sync_object * -_mesa_new_sync_object(struct gl_context *ctx, GLenum type) -{ - struct gl_sync_object *s = MALLOC_STRUCT(gl_sync_object); - (void) ctx; - (void) type; - - return s; -} - - -static void -_mesa_delete_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj) -{ - (void) ctx; - free(syncObj); -} - - -static void -_mesa_fence_sync(struct gl_context *ctx, struct gl_sync_object *syncObj, - GLenum condition, GLbitfield flags) -{ - (void) ctx; - (void) condition; - (void) flags; - - syncObj->StatusFlag = 1; -} - - -static void -_mesa_check_sync(struct gl_context *ctx, struct gl_sync_object *syncObj) -{ - (void) ctx; - (void) syncObj; - - /* No-op for software rendering. Hardware drivers will need to determine - * whether the state of the sync object has changed. - */ -} - - -static void -_mesa_wait_sync(struct gl_context *ctx, struct gl_sync_object *syncObj, - GLbitfield flags, GLuint64 timeout) -{ - (void) ctx; - (void) syncObj; - (void) flags; - (void) timeout; - - /* No-op for software rendering. Hardware drivers will need to wait until - * the state of the sync object changes or the timeout expires. - */ -} - - -void -_mesa_init_sync_object_functions(struct dd_function_table *driver) -{ - driver->NewSyncObject = _mesa_new_sync_object; - driver->FenceSync = _mesa_fence_sync; - driver->DeleteSyncObject = _mesa_delete_sync_object; - driver->CheckSync = _mesa_check_sync; - - /* Use the same no-op wait function for both. - */ - driver->ClientWaitSync = _mesa_wait_sync; - driver->ServerWaitSync = _mesa_wait_sync; -} - - -void -_mesa_init_sync_dispatch(struct _glapi_table *disp) -{ - SET_IsSync(disp, _mesa_IsSync); - SET_DeleteSync(disp, _mesa_DeleteSync); - SET_FenceSync(disp, _mesa_FenceSync); - SET_ClientWaitSync(disp, _mesa_ClientWaitSync); - SET_WaitSync(disp, _mesa_WaitSync); - SET_GetInteger64v(disp, _mesa_GetInteger64v); - SET_GetSynciv(disp, _mesa_GetSynciv); -} - - -/** - * Allocate/init the context state related to sync objects. - */ -void -_mesa_init_sync(struct gl_context *ctx) -{ - (void) ctx; -} - - -/** - * Free the context state related to sync objects. - */ -void -_mesa_free_sync_data(struct gl_context *ctx) -{ - (void) ctx; -} - - -static int -_mesa_validate_sync(struct gl_sync_object *syncObj) -{ - return (syncObj != NULL) - && (syncObj->Type == GL_SYNC_FENCE) - && !syncObj->DeletePending; -} - - -void -_mesa_ref_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj) -{ - _glthread_LOCK_MUTEX(ctx->Shared->Mutex); - syncObj->RefCount++; - _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); -} - - -void -_mesa_unref_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj) -{ - _glthread_LOCK_MUTEX(ctx->Shared->Mutex); - syncObj->RefCount--; - if (syncObj->RefCount == 0) { - remove_from_list(& syncObj->link); - _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); - - ctx->Driver.DeleteSyncObject(ctx, syncObj); - } else { - _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); - } -} - - -GLboolean GLAPIENTRY -_mesa_IsSync(GLsync sync) -{ - GET_CURRENT_CONTEXT(ctx); - struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; - ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE); - - return _mesa_validate_sync(syncObj) ? GL_TRUE : GL_FALSE; -} - - -void GLAPIENTRY -_mesa_DeleteSync(GLsync sync) -{ - GET_CURRENT_CONTEXT(ctx); - struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; - ASSERT_OUTSIDE_BEGIN_END(ctx); - - /* From the GL_ARB_sync spec: - * - * DeleteSync will silently ignore a value of zero. An - * INVALID_VALUE error is generated if is neither zero nor the - * name of a sync object. - */ - if (sync == 0) { - return; - } - - if (!_mesa_validate_sync(syncObj)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glDeleteSync"); - return; - } - - /* If there are no client-waits or server-waits pending on this sync, delete - * the underlying object. - */ - syncObj->DeletePending = GL_TRUE; - _mesa_unref_sync_object(ctx, syncObj); -} - - -GLsync GLAPIENTRY -_mesa_FenceSync(GLenum condition, GLbitfield flags) -{ - GET_CURRENT_CONTEXT(ctx); - struct gl_sync_object *syncObj; - ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, 0); - - if (condition != GL_SYNC_GPU_COMMANDS_COMPLETE) { - _mesa_error(ctx, GL_INVALID_ENUM, "glFenceSync(condition=0x%x)", - condition); - return 0; - } - - if (flags != 0) { - _mesa_error(ctx, GL_INVALID_VALUE, "glFenceSync(flags=0x%x)", - condition); - return 0; - } - - syncObj = ctx->Driver.NewSyncObject(ctx, GL_SYNC_FENCE); - if (syncObj != NULL) { - syncObj->Type = GL_SYNC_FENCE; - /* The name is not currently used, and it is never visible to - * applications. If sync support is extended to provide support for - * NV_fence, this field will be used. We'll also need to add an object - * ID hashtable. - */ - syncObj->Name = 1; - syncObj->RefCount = 1; - syncObj->DeletePending = GL_FALSE; - syncObj->SyncCondition = condition; - syncObj->Flags = flags; - syncObj->StatusFlag = 0; - - ctx->Driver.FenceSync(ctx, syncObj, condition, flags); - - _glthread_LOCK_MUTEX(ctx->Shared->Mutex); - insert_at_tail(& ctx->Shared->SyncObjects, & syncObj->link); - _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); - - return (GLsync) syncObj; - } - - return NULL; -} - - -GLenum GLAPIENTRY -_mesa_ClientWaitSync(GLsync sync, GLbitfield flags, GLuint64 timeout) -{ - GET_CURRENT_CONTEXT(ctx); - struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; - GLenum ret; - ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_WAIT_FAILED); - - if (!_mesa_validate_sync(syncObj)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glClientWaitSync"); - return GL_WAIT_FAILED; - } - - if ((flags & ~GL_SYNC_FLUSH_COMMANDS_BIT) != 0) { - _mesa_error(ctx, GL_INVALID_ENUM, "glClientWaitSync(flags=0x%x)", flags); - return GL_WAIT_FAILED; - } - - _mesa_ref_sync_object(ctx, syncObj); - - /* From the GL_ARB_sync spec: - * - * ClientWaitSync returns one of four status values. A return value of - * ALREADY_SIGNALED indicates that was signaled at the time - * ClientWaitSync was called. ALREADY_SIGNALED will always be returned - * if was signaled, even if the value of is zero. - */ - ctx->Driver.CheckSync(ctx, syncObj); - if (syncObj->StatusFlag) { - ret = GL_ALREADY_SIGNALED; - } else { - ctx->Driver.ClientWaitSync(ctx, syncObj, flags, timeout); - - ret = syncObj->StatusFlag ? GL_CONDITION_SATISFIED : GL_TIMEOUT_EXPIRED; - } - - _mesa_unref_sync_object(ctx, syncObj); - return ret; -} - - -void GLAPIENTRY -_mesa_WaitSync(GLsync sync, GLbitfield flags, GLuint64 timeout) -{ - GET_CURRENT_CONTEXT(ctx); - struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (!_mesa_validate_sync(syncObj)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glWaitSync"); - return; - } - - if (flags != 0) { - _mesa_error(ctx, GL_INVALID_ENUM, "glWaitSync(flags=0x%x)", flags); - return; - } - - /* From the GL_ARB_sync spec: - * - * If the value of is zero, then WaitSync does nothing. - */ - if (timeout == 0) { - return; - } - - ctx->Driver.ServerWaitSync(ctx, syncObj, flags, timeout); -} - - -void GLAPIENTRY -_mesa_GetSynciv(GLsync sync, GLenum pname, GLsizei bufSize, GLsizei *length, - GLint *values) -{ - GET_CURRENT_CONTEXT(ctx); - struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; - GLsizei size = 0; - GLint v[1]; - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (!_mesa_validate_sync(syncObj)) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glGetSynciv"); - return; - } - - switch (pname) { - case GL_OBJECT_TYPE: - v[0] = syncObj->Type; - size = 1; - break; - - case GL_SYNC_CONDITION: - v[0] = syncObj->SyncCondition; - size = 1; - break; - - case GL_SYNC_STATUS: - /* Update the state of the sync by dipping into the driver. Note that - * this call won't block. It just updates state in the common object - * data from the current driver state. - */ - ctx->Driver.CheckSync(ctx, syncObj); - - v[0] = (syncObj->StatusFlag) ? GL_SIGNALED : GL_UNSIGNALED; - size = 1; - break; - - case GL_SYNC_FLAGS: - v[0] = syncObj->Flags; - size = 1; - break; - - default: - _mesa_error(ctx, GL_INVALID_ENUM, "glGetSynciv(pname=0x%x)\n", pname); - return; - } - - if (size > 0) { - const GLsizei copy_count = MIN2(size, bufSize); - - memcpy(values, v, sizeof(GLint) * copy_count); - } - - if (length != NULL) { - *length = size; - } -} - -#endif /* FEATURE_ARB_sync */ +/* + * Copyright © 2009 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ + +/** + * \file syncobj.c + * Sync object management. + * + * Unlike textures and other objects that are shared between contexts, sync + * objects are not bound to the context. As a result, the reference counting + * and delete behavior of sync objects is slightly different. References to + * sync objects are added: + * + * - By \c glFencSynce. This sets the initial reference count to 1. + * - At the start of \c glClientWaitSync. The reference is held for the + * duration of the wait call. + * + * References are removed: + * + * - By \c glDeleteSync. + * - At the end of \c glClientWaitSync. + * + * Additionally, drivers may call \c _mesa_ref_sync_object and + * \c _mesa_unref_sync_object as needed to implement \c ServerWaitSync. + * + * As with shader objects, sync object names become invalid as soon as + * \c glDeleteSync is called. For this reason \c glDeleteSync sets the + * \c DeletePending flag. All functions validate object handles by testing + * this flag. + * + * \note + * Only \c GL_ARB_sync objects are shared between contexts. If support is ever + * added for either \c GL_NV_fence or \c GL_APPLE_fence different semantics + * will need to be implemented. + * + * \author Ian Romanick + */ + +#include "glheader.h" +#include "imports.h" +#include "context.h" +#include "macros.h" +#include "mfeatures.h" +#include "get.h" +#include "dispatch.h" +#include "mtypes.h" + +#if FEATURE_ARB_sync +#include "syncobj.h" + +static struct gl_sync_object * +_mesa_new_sync_object(struct gl_context *ctx, GLenum type) +{ + struct gl_sync_object *s = MALLOC_STRUCT(gl_sync_object); + (void) ctx; + (void) type; + + return s; +} + + +static void +_mesa_delete_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj) +{ + (void) ctx; + free(syncObj); +} + + +static void +_mesa_fence_sync(struct gl_context *ctx, struct gl_sync_object *syncObj, + GLenum condition, GLbitfield flags) +{ + (void) ctx; + (void) condition; + (void) flags; + + syncObj->StatusFlag = 1; +} + + +static void +_mesa_check_sync(struct gl_context *ctx, struct gl_sync_object *syncObj) +{ + (void) ctx; + (void) syncObj; + + /* No-op for software rendering. Hardware drivers will need to determine + * whether the state of the sync object has changed. + */ +} + + +static void +_mesa_wait_sync(struct gl_context *ctx, struct gl_sync_object *syncObj, + GLbitfield flags, GLuint64 timeout) +{ + (void) ctx; + (void) syncObj; + (void) flags; + (void) timeout; + + /* No-op for software rendering. Hardware drivers will need to wait until + * the state of the sync object changes or the timeout expires. + */ +} + + +void +_mesa_init_sync_object_functions(struct dd_function_table *driver) +{ + driver->NewSyncObject = _mesa_new_sync_object; + driver->FenceSync = _mesa_fence_sync; + driver->DeleteSyncObject = _mesa_delete_sync_object; + driver->CheckSync = _mesa_check_sync; + + /* Use the same no-op wait function for both. + */ + driver->ClientWaitSync = _mesa_wait_sync; + driver->ServerWaitSync = _mesa_wait_sync; +} + + +void +_mesa_init_sync_dispatch(struct _glapi_table *disp) +{ + SET_IsSync(disp, _mesa_IsSync); + SET_DeleteSync(disp, _mesa_DeleteSync); + SET_FenceSync(disp, _mesa_FenceSync); + SET_ClientWaitSync(disp, _mesa_ClientWaitSync); + SET_WaitSync(disp, _mesa_WaitSync); + SET_GetInteger64v(disp, _mesa_GetInteger64v); + SET_GetSynciv(disp, _mesa_GetSynciv); +} + + +/** + * Allocate/init the context state related to sync objects. + */ +void +_mesa_init_sync(struct gl_context *ctx) +{ + (void) ctx; +} + + +/** + * Free the context state related to sync objects. + */ +void +_mesa_free_sync_data(struct gl_context *ctx) +{ + (void) ctx; +} + + +static int +_mesa_validate_sync(struct gl_sync_object *syncObj) +{ + return (syncObj != NULL) + && (syncObj->Type == GL_SYNC_FENCE) + && !syncObj->DeletePending; +} + + +void +_mesa_ref_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj) +{ + _glthread_LOCK_MUTEX(ctx->Shared->Mutex); + syncObj->RefCount++; + _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); +} + + +void +_mesa_unref_sync_object(struct gl_context *ctx, struct gl_sync_object *syncObj) +{ + _glthread_LOCK_MUTEX(ctx->Shared->Mutex); + syncObj->RefCount--; + if (syncObj->RefCount == 0) { + remove_from_list(& syncObj->link); + _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); + + ctx->Driver.DeleteSyncObject(ctx, syncObj); + } else { + _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); + } +} + + +GLboolean GLAPIENTRY +_mesa_IsSync(GLsync sync) +{ + GET_CURRENT_CONTEXT(ctx); + struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; + ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_FALSE); + + return _mesa_validate_sync(syncObj) ? GL_TRUE : GL_FALSE; +} + + +void GLAPIENTRY +_mesa_DeleteSync(GLsync sync) +{ + GET_CURRENT_CONTEXT(ctx); + struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; + ASSERT_OUTSIDE_BEGIN_END(ctx); + + /* From the GL_ARB_sync spec: + * + * DeleteSync will silently ignore a value of zero. An + * INVALID_VALUE error is generated if is neither zero nor the + * name of a sync object. + */ + if (sync == 0) { + return; + } + + if (!_mesa_validate_sync(syncObj)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glDeleteSync"); + return; + } + + /* If there are no client-waits or server-waits pending on this sync, delete + * the underlying object. + */ + syncObj->DeletePending = GL_TRUE; + _mesa_unref_sync_object(ctx, syncObj); +} + + +GLsync GLAPIENTRY +_mesa_FenceSync(GLenum condition, GLbitfield flags) +{ + GET_CURRENT_CONTEXT(ctx); + struct gl_sync_object *syncObj; + ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, 0); + + if (condition != GL_SYNC_GPU_COMMANDS_COMPLETE) { + _mesa_error(ctx, GL_INVALID_ENUM, "glFenceSync(condition=0x%x)", + condition); + return 0; + } + + if (flags != 0) { + _mesa_error(ctx, GL_INVALID_VALUE, "glFenceSync(flags=0x%x)", + condition); + return 0; + } + + syncObj = ctx->Driver.NewSyncObject(ctx, GL_SYNC_FENCE); + if (syncObj != NULL) { + syncObj->Type = GL_SYNC_FENCE; + /* The name is not currently used, and it is never visible to + * applications. If sync support is extended to provide support for + * NV_fence, this field will be used. We'll also need to add an object + * ID hashtable. + */ + syncObj->Name = 1; + syncObj->RefCount = 1; + syncObj->DeletePending = GL_FALSE; + syncObj->SyncCondition = condition; + syncObj->Flags = flags; + syncObj->StatusFlag = 0; + + ctx->Driver.FenceSync(ctx, syncObj, condition, flags); + + _glthread_LOCK_MUTEX(ctx->Shared->Mutex); + insert_at_tail(& ctx->Shared->SyncObjects, & syncObj->link); + _glthread_UNLOCK_MUTEX(ctx->Shared->Mutex); + + return (GLsync) syncObj; + } + + return NULL; +} + + +GLenum GLAPIENTRY +_mesa_ClientWaitSync(GLsync sync, GLbitfield flags, GLuint64 timeout) +{ + GET_CURRENT_CONTEXT(ctx); + struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; + GLenum ret; + ASSERT_OUTSIDE_BEGIN_END_WITH_RETVAL(ctx, GL_WAIT_FAILED); + + if (!_mesa_validate_sync(syncObj)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glClientWaitSync"); + return GL_WAIT_FAILED; + } + + if ((flags & ~GL_SYNC_FLUSH_COMMANDS_BIT) != 0) { + _mesa_error(ctx, GL_INVALID_ENUM, "glClientWaitSync(flags=0x%x)", flags); + return GL_WAIT_FAILED; + } + + _mesa_ref_sync_object(ctx, syncObj); + + /* From the GL_ARB_sync spec: + * + * ClientWaitSync returns one of four status values. A return value of + * ALREADY_SIGNALED indicates that was signaled at the time + * ClientWaitSync was called. ALREADY_SIGNALED will always be returned + * if was signaled, even if the value of is zero. + */ + ctx->Driver.CheckSync(ctx, syncObj); + if (syncObj->StatusFlag) { + ret = GL_ALREADY_SIGNALED; + } else { + ctx->Driver.ClientWaitSync(ctx, syncObj, flags, timeout); + + ret = syncObj->StatusFlag ? GL_CONDITION_SATISFIED : GL_TIMEOUT_EXPIRED; + } + + _mesa_unref_sync_object(ctx, syncObj); + return ret; +} + + +void GLAPIENTRY +_mesa_WaitSync(GLsync sync, GLbitfield flags, GLuint64 timeout) +{ + GET_CURRENT_CONTEXT(ctx); + struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (!_mesa_validate_sync(syncObj)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glWaitSync"); + return; + } + + if (flags != 0) { + _mesa_error(ctx, GL_INVALID_ENUM, "glWaitSync(flags=0x%x)", flags); + return; + } + + /* From the GL_ARB_sync spec: + * + * If the value of is zero, then WaitSync does nothing. + */ + if (timeout == 0) { + return; + } + + ctx->Driver.ServerWaitSync(ctx, syncObj, flags, timeout); +} + + +void GLAPIENTRY +_mesa_GetSynciv(GLsync sync, GLenum pname, GLsizei bufSize, GLsizei *length, + GLint *values) +{ + GET_CURRENT_CONTEXT(ctx); + struct gl_sync_object *const syncObj = (struct gl_sync_object *) sync; + GLsizei size = 0; + GLint v[1]; + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (!_mesa_validate_sync(syncObj)) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glGetSynciv"); + return; + } + + switch (pname) { + case GL_OBJECT_TYPE: + v[0] = syncObj->Type; + size = 1; + break; + + case GL_SYNC_CONDITION: + v[0] = syncObj->SyncCondition; + size = 1; + break; + + case GL_SYNC_STATUS: + /* Update the state of the sync by dipping into the driver. Note that + * this call won't block. It just updates state in the common object + * data from the current driver state. + */ + ctx->Driver.CheckSync(ctx, syncObj); + + v[0] = (syncObj->StatusFlag) ? GL_SIGNALED : GL_UNSIGNALED; + size = 1; + break; + + case GL_SYNC_FLAGS: + v[0] = syncObj->Flags; + size = 1; + break; + + default: + _mesa_error(ctx, GL_INVALID_ENUM, "glGetSynciv(pname=0x%x)\n", pname); + return; + } + + if (size > 0) { + const GLsizei copy_count = MIN2(size, bufSize); + + memcpy(values, v, sizeof(GLint) * copy_count); + } + + if (length != NULL) { + *length = size; + } +} + +#endif /* FEATURE_ARB_sync */ diff --git a/mesalib/src/mesa/main/texformat.h b/mesalib/src/mesa/main/texformat.h index b84117995..3cf09213a 100644 --- a/mesalib/src/mesa/main/texformat.h +++ b/mesalib/src/mesa/main/texformat.h @@ -1,39 +1,39 @@ -/* - * Mesa 3-D graphics library - * Version: 7.75 - * - * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. - * Copyright (c) 2008-2009 VMware, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#ifndef TEXFORMAT_H -#define TEXFORMAT_H - - -#include "formats.h" - -struct gl_context; - -extern gl_format -_mesa_choose_tex_format( struct gl_context *ctx, GLint internalFormat, - GLenum format, GLenum type ); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.75 + * + * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. + * Copyright (c) 2008-2009 VMware, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef TEXFORMAT_H +#define TEXFORMAT_H + + +#include "formats.h" + +struct gl_context; + +extern gl_format +_mesa_choose_tex_format( struct gl_context *ctx, GLint internalFormat, + GLenum format, GLenum type ); + + +#endif diff --git a/mesalib/src/mesa/main/texgen.c b/mesalib/src/mesa/main/texgen.c index 10eaa89ad..0ace0b936 100644 --- a/mesalib/src/mesa/main/texgen.c +++ b/mesalib/src/mesa/main/texgen.c @@ -1,401 +1,401 @@ -/* - * Mesa 3-D graphics library - * Version: 7.5 - * - * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. - * Copyright (C) 2009 VMware, Inc. All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file texgen.c - * - * glTexGen-related functions - */ - - -#include "main/glheader.h" -#include "main/context.h" -#include "main/enums.h" -#include "main/macros.h" -#include "main/mfeatures.h" -#include "main/texgen.h" -#include "main/texstate.h" -#include "math/m_matrix.h" -#include "main/dispatch.h" - - -#if FEATURE_texgen - - -/** - * Return texgen state for given coordinate - */ -static struct gl_texgen * -get_texgen(struct gl_texture_unit *texUnit, GLenum coord) -{ - switch (coord) { - case GL_S: - return &texUnit->GenS; - case GL_T: - return &texUnit->GenT; - case GL_R: - return &texUnit->GenR; - case GL_Q: - return &texUnit->GenQ; - default: - return NULL; - } -} - - -void GLAPIENTRY -_mesa_TexGenfv( GLenum coord, GLenum pname, const GLfloat *params ) -{ - struct gl_texture_unit *texUnit; - struct gl_texgen *texgen; - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (MESA_VERBOSE&(VERBOSE_API|VERBOSE_TEXTURE)) - _mesa_debug(ctx, "glTexGen %s %s %.1f(%s)...\n", - _mesa_lookup_enum_by_nr(coord), - _mesa_lookup_enum_by_nr(pname), - *params, - _mesa_lookup_enum_by_nr((GLenum) (GLint) *params)); - - if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glTexGen(current unit)"); - return; - } - - texUnit = _mesa_get_current_tex_unit(ctx); - - texgen = get_texgen(texUnit, coord); - if (!texgen) { - _mesa_error(ctx, GL_INVALID_ENUM, "glTexGen(coord)"); - return; - } - - switch (pname) { - case GL_TEXTURE_GEN_MODE: - { - GLenum mode = (GLenum) (GLint) params[0]; - GLbitfield bit = 0x0; - if (texgen->Mode == mode) - return; - switch (mode) { - case GL_OBJECT_LINEAR: - bit = TEXGEN_OBJ_LINEAR; - break; - case GL_EYE_LINEAR: - bit = TEXGEN_EYE_LINEAR; - break; - case GL_SPHERE_MAP: - if (coord == GL_S || coord == GL_T) - bit = TEXGEN_SPHERE_MAP; - break; - case GL_REFLECTION_MAP_NV: - if (coord != GL_Q) - bit = TEXGEN_REFLECTION_MAP_NV; - break; - case GL_NORMAL_MAP_NV: - if (coord != GL_Q) - bit = TEXGEN_NORMAL_MAP_NV; - break; - default: - ; /* nop */ - } - if (!bit) { - _mesa_error( ctx, GL_INVALID_ENUM, "glTexGenfv(param)" ); - return; - } - FLUSH_VERTICES(ctx, _NEW_TEXTURE); - texgen->Mode = mode; - texgen->_ModeBit = bit; - } - break; - - case GL_OBJECT_PLANE: - { - if (TEST_EQ_4V(texgen->ObjectPlane, params)) - return; - FLUSH_VERTICES(ctx, _NEW_TEXTURE); - COPY_4FV(texgen->ObjectPlane, params); - } - break; - - case GL_EYE_PLANE: - { - GLfloat tmp[4]; - /* Transform plane equation by the inverse modelview matrix */ - if (_math_matrix_is_dirty(ctx->ModelviewMatrixStack.Top)) { - _math_matrix_analyse(ctx->ModelviewMatrixStack.Top); - } - _mesa_transform_vector(tmp, params, - ctx->ModelviewMatrixStack.Top->inv); - if (TEST_EQ_4V(texgen->EyePlane, tmp)) - return; - FLUSH_VERTICES(ctx, _NEW_TEXTURE); - COPY_4FV(texgen->EyePlane, tmp); - } - break; - - default: - _mesa_error( ctx, GL_INVALID_ENUM, "glTexGenfv(pname)" ); - return; - } - - if (ctx->Driver.TexGen) - ctx->Driver.TexGen( ctx, coord, pname, params ); -} - - -static void GLAPIENTRY -_mesa_TexGeniv(GLenum coord, GLenum pname, const GLint *params ) -{ - GLfloat p[4]; - p[0] = (GLfloat) params[0]; - if (pname == GL_TEXTURE_GEN_MODE) { - p[1] = p[2] = p[3] = 0.0F; - } - else { - p[1] = (GLfloat) params[1]; - p[2] = (GLfloat) params[2]; - p[3] = (GLfloat) params[3]; - } - _mesa_TexGenfv(coord, pname, p); -} - - -static void GLAPIENTRY -_mesa_TexGend(GLenum coord, GLenum pname, GLdouble param ) -{ - GLfloat p[4]; - p[0] = (GLfloat) param; - p[1] = p[2] = p[3] = 0.0F; - _mesa_TexGenfv( coord, pname, p ); -} - -#if FEATURE_ES1 - -void GLAPIENTRY -_es_GetTexGenfv(GLenum coord, GLenum pname, GLfloat *params) -{ - ASSERT(coord == GL_TEXTURE_GEN_STR_OES); - _mesa_GetTexGenfv(GL_S, pname, params); -} - - -void GLAPIENTRY -_es_TexGenf(GLenum coord, GLenum pname, GLfloat param) -{ - ASSERT(coord == GL_TEXTURE_GEN_STR_OES); - /* set S, T, and R at the same time */ - _mesa_TexGenf(GL_S, pname, param); - _mesa_TexGenf(GL_T, pname, param); - _mesa_TexGenf(GL_R, pname, param); -} - - -void GLAPIENTRY -_es_TexGenfv(GLenum coord, GLenum pname, const GLfloat *params) -{ - ASSERT(coord == GL_TEXTURE_GEN_STR_OES); - /* set S, T, and R at the same time */ - _mesa_TexGenfv(GL_S, pname, params); - _mesa_TexGenfv(GL_T, pname, params); - _mesa_TexGenfv(GL_R, pname, params); -} - -#endif - -static void GLAPIENTRY -_mesa_TexGendv(GLenum coord, GLenum pname, const GLdouble *params ) -{ - GLfloat p[4]; - p[0] = (GLfloat) params[0]; - if (pname == GL_TEXTURE_GEN_MODE) { - p[1] = p[2] = p[3] = 0.0F; - } - else { - p[1] = (GLfloat) params[1]; - p[2] = (GLfloat) params[2]; - p[3] = (GLfloat) params[3]; - } - _mesa_TexGenfv( coord, pname, p ); -} - - -void GLAPIENTRY -_mesa_TexGenf( GLenum coord, GLenum pname, GLfloat param ) -{ - GLfloat p[4]; - p[0] = param; - p[1] = p[2] = p[3] = 0.0F; - _mesa_TexGenfv(coord, pname, p); -} - - -void GLAPIENTRY -_mesa_TexGeni( GLenum coord, GLenum pname, GLint param ) -{ - GLint p[4]; - p[0] = param; - p[1] = p[2] = p[3] = 0; - _mesa_TexGeniv( coord, pname, p ); -} - - - -static void GLAPIENTRY -_mesa_GetTexGendv( GLenum coord, GLenum pname, GLdouble *params ) -{ - struct gl_texture_unit *texUnit; - struct gl_texgen *texgen; - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexGendv(current unit)"); - return; - } - - texUnit = _mesa_get_current_tex_unit(ctx); - - texgen = get_texgen(texUnit, coord); - if (!texgen) { - _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexGendv(coord)"); - return; - } - - switch (pname) { - case GL_TEXTURE_GEN_MODE: - params[0] = ENUM_TO_DOUBLE(texgen->Mode); - break; - case GL_OBJECT_PLANE: - COPY_4V(params, texgen->ObjectPlane); - break; - case GL_EYE_PLANE: - COPY_4V(params, texgen->EyePlane); - break; - default: - _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexGendv(pname)" ); - } -} - - - -void GLAPIENTRY -_mesa_GetTexGenfv( GLenum coord, GLenum pname, GLfloat *params ) -{ - struct gl_texture_unit *texUnit; - struct gl_texgen *texgen; - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexGenfv(current unit)"); - return; - } - - texUnit = _mesa_get_current_tex_unit(ctx); - - texgen = get_texgen(texUnit, coord); - if (!texgen) { - _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexGenfv(coord)"); - return; - } - - switch (pname) { - case GL_TEXTURE_GEN_MODE: - params[0] = ENUM_TO_FLOAT(texgen->Mode); - break; - case GL_OBJECT_PLANE: - COPY_4V(params, texgen->ObjectPlane); - break; - case GL_EYE_PLANE: - COPY_4V(params, texgen->EyePlane); - break; - default: - _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexGenfv(pname)" ); - } -} - - - -static void GLAPIENTRY -_mesa_GetTexGeniv( GLenum coord, GLenum pname, GLint *params ) -{ - struct gl_texture_unit *texUnit; - struct gl_texgen *texgen; - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { - _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexGeniv(current unit)"); - return; - } - - texUnit = _mesa_get_current_tex_unit(ctx); - - texgen = get_texgen(texUnit, coord); - if (!texgen) { - _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexGeniv(coord)"); - return; - } - - switch (pname) { - case GL_TEXTURE_GEN_MODE: - params[0] = texgen->Mode; - break; - case GL_OBJECT_PLANE: - params[0] = (GLint) texgen->ObjectPlane[0]; - params[1] = (GLint) texgen->ObjectPlane[1]; - params[2] = (GLint) texgen->ObjectPlane[2]; - params[3] = (GLint) texgen->ObjectPlane[3]; - break; - case GL_EYE_PLANE: - params[0] = (GLint) texgen->EyePlane[0]; - params[1] = (GLint) texgen->EyePlane[1]; - params[2] = (GLint) texgen->EyePlane[2]; - params[3] = (GLint) texgen->EyePlane[3]; - break; - default: - _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexGeniv(pname)" ); - } -} - - -void -_mesa_init_texgen_dispatch(struct _glapi_table *disp) -{ - SET_GetTexGendv(disp, _mesa_GetTexGendv); - SET_GetTexGenfv(disp, _mesa_GetTexGenfv); - SET_GetTexGeniv(disp, _mesa_GetTexGeniv); - SET_TexGend(disp, _mesa_TexGend); - SET_TexGendv(disp, _mesa_TexGendv); - SET_TexGenf(disp, _mesa_TexGenf); - SET_TexGenfv(disp, _mesa_TexGenfv); - SET_TexGeni(disp, _mesa_TexGeni); - SET_TexGeniv(disp, _mesa_TexGeniv); -} - - -#endif /* FEATURE_texgen */ +/* + * Mesa 3-D graphics library + * Version: 7.5 + * + * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. + * Copyright (C) 2009 VMware, Inc. All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file texgen.c + * + * glTexGen-related functions + */ + + +#include "main/glheader.h" +#include "main/context.h" +#include "main/enums.h" +#include "main/macros.h" +#include "main/mfeatures.h" +#include "main/texgen.h" +#include "main/texstate.h" +#include "math/m_matrix.h" +#include "main/dispatch.h" + + +#if FEATURE_texgen + + +/** + * Return texgen state for given coordinate + */ +static struct gl_texgen * +get_texgen(struct gl_texture_unit *texUnit, GLenum coord) +{ + switch (coord) { + case GL_S: + return &texUnit->GenS; + case GL_T: + return &texUnit->GenT; + case GL_R: + return &texUnit->GenR; + case GL_Q: + return &texUnit->GenQ; + default: + return NULL; + } +} + + +void GLAPIENTRY +_mesa_TexGenfv( GLenum coord, GLenum pname, const GLfloat *params ) +{ + struct gl_texture_unit *texUnit; + struct gl_texgen *texgen; + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (MESA_VERBOSE&(VERBOSE_API|VERBOSE_TEXTURE)) + _mesa_debug(ctx, "glTexGen %s %s %.1f(%s)...\n", + _mesa_lookup_enum_by_nr(coord), + _mesa_lookup_enum_by_nr(pname), + *params, + _mesa_lookup_enum_by_nr((GLenum) (GLint) *params)); + + if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glTexGen(current unit)"); + return; + } + + texUnit = _mesa_get_current_tex_unit(ctx); + + texgen = get_texgen(texUnit, coord); + if (!texgen) { + _mesa_error(ctx, GL_INVALID_ENUM, "glTexGen(coord)"); + return; + } + + switch (pname) { + case GL_TEXTURE_GEN_MODE: + { + GLenum mode = (GLenum) (GLint) params[0]; + GLbitfield bit = 0x0; + if (texgen->Mode == mode) + return; + switch (mode) { + case GL_OBJECT_LINEAR: + bit = TEXGEN_OBJ_LINEAR; + break; + case GL_EYE_LINEAR: + bit = TEXGEN_EYE_LINEAR; + break; + case GL_SPHERE_MAP: + if (coord == GL_S || coord == GL_T) + bit = TEXGEN_SPHERE_MAP; + break; + case GL_REFLECTION_MAP_NV: + if (coord != GL_Q) + bit = TEXGEN_REFLECTION_MAP_NV; + break; + case GL_NORMAL_MAP_NV: + if (coord != GL_Q) + bit = TEXGEN_NORMAL_MAP_NV; + break; + default: + ; /* nop */ + } + if (!bit) { + _mesa_error( ctx, GL_INVALID_ENUM, "glTexGenfv(param)" ); + return; + } + FLUSH_VERTICES(ctx, _NEW_TEXTURE); + texgen->Mode = mode; + texgen->_ModeBit = bit; + } + break; + + case GL_OBJECT_PLANE: + { + if (TEST_EQ_4V(texgen->ObjectPlane, params)) + return; + FLUSH_VERTICES(ctx, _NEW_TEXTURE); + COPY_4FV(texgen->ObjectPlane, params); + } + break; + + case GL_EYE_PLANE: + { + GLfloat tmp[4]; + /* Transform plane equation by the inverse modelview matrix */ + if (_math_matrix_is_dirty(ctx->ModelviewMatrixStack.Top)) { + _math_matrix_analyse(ctx->ModelviewMatrixStack.Top); + } + _mesa_transform_vector(tmp, params, + ctx->ModelviewMatrixStack.Top->inv); + if (TEST_EQ_4V(texgen->EyePlane, tmp)) + return; + FLUSH_VERTICES(ctx, _NEW_TEXTURE); + COPY_4FV(texgen->EyePlane, tmp); + } + break; + + default: + _mesa_error( ctx, GL_INVALID_ENUM, "glTexGenfv(pname)" ); + return; + } + + if (ctx->Driver.TexGen) + ctx->Driver.TexGen( ctx, coord, pname, params ); +} + + +static void GLAPIENTRY +_mesa_TexGeniv(GLenum coord, GLenum pname, const GLint *params ) +{ + GLfloat p[4]; + p[0] = (GLfloat) params[0]; + if (pname == GL_TEXTURE_GEN_MODE) { + p[1] = p[2] = p[3] = 0.0F; + } + else { + p[1] = (GLfloat) params[1]; + p[2] = (GLfloat) params[2]; + p[3] = (GLfloat) params[3]; + } + _mesa_TexGenfv(coord, pname, p); +} + + +static void GLAPIENTRY +_mesa_TexGend(GLenum coord, GLenum pname, GLdouble param ) +{ + GLfloat p[4]; + p[0] = (GLfloat) param; + p[1] = p[2] = p[3] = 0.0F; + _mesa_TexGenfv( coord, pname, p ); +} + +#if FEATURE_ES1 + +void GLAPIENTRY +_es_GetTexGenfv(GLenum coord, GLenum pname, GLfloat *params) +{ + ASSERT(coord == GL_TEXTURE_GEN_STR_OES); + _mesa_GetTexGenfv(GL_S, pname, params); +} + + +void GLAPIENTRY +_es_TexGenf(GLenum coord, GLenum pname, GLfloat param) +{ + ASSERT(coord == GL_TEXTURE_GEN_STR_OES); + /* set S, T, and R at the same time */ + _mesa_TexGenf(GL_S, pname, param); + _mesa_TexGenf(GL_T, pname, param); + _mesa_TexGenf(GL_R, pname, param); +} + + +void GLAPIENTRY +_es_TexGenfv(GLenum coord, GLenum pname, const GLfloat *params) +{ + ASSERT(coord == GL_TEXTURE_GEN_STR_OES); + /* set S, T, and R at the same time */ + _mesa_TexGenfv(GL_S, pname, params); + _mesa_TexGenfv(GL_T, pname, params); + _mesa_TexGenfv(GL_R, pname, params); +} + +#endif + +static void GLAPIENTRY +_mesa_TexGendv(GLenum coord, GLenum pname, const GLdouble *params ) +{ + GLfloat p[4]; + p[0] = (GLfloat) params[0]; + if (pname == GL_TEXTURE_GEN_MODE) { + p[1] = p[2] = p[3] = 0.0F; + } + else { + p[1] = (GLfloat) params[1]; + p[2] = (GLfloat) params[2]; + p[3] = (GLfloat) params[3]; + } + _mesa_TexGenfv( coord, pname, p ); +} + + +void GLAPIENTRY +_mesa_TexGenf( GLenum coord, GLenum pname, GLfloat param ) +{ + GLfloat p[4]; + p[0] = param; + p[1] = p[2] = p[3] = 0.0F; + _mesa_TexGenfv(coord, pname, p); +} + + +void GLAPIENTRY +_mesa_TexGeni( GLenum coord, GLenum pname, GLint param ) +{ + GLint p[4]; + p[0] = param; + p[1] = p[2] = p[3] = 0; + _mesa_TexGeniv( coord, pname, p ); +} + + + +static void GLAPIENTRY +_mesa_GetTexGendv( GLenum coord, GLenum pname, GLdouble *params ) +{ + struct gl_texture_unit *texUnit; + struct gl_texgen *texgen; + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexGendv(current unit)"); + return; + } + + texUnit = _mesa_get_current_tex_unit(ctx); + + texgen = get_texgen(texUnit, coord); + if (!texgen) { + _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexGendv(coord)"); + return; + } + + switch (pname) { + case GL_TEXTURE_GEN_MODE: + params[0] = ENUM_TO_DOUBLE(texgen->Mode); + break; + case GL_OBJECT_PLANE: + COPY_4V(params, texgen->ObjectPlane); + break; + case GL_EYE_PLANE: + COPY_4V(params, texgen->EyePlane); + break; + default: + _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexGendv(pname)" ); + } +} + + + +void GLAPIENTRY +_mesa_GetTexGenfv( GLenum coord, GLenum pname, GLfloat *params ) +{ + struct gl_texture_unit *texUnit; + struct gl_texgen *texgen; + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexGenfv(current unit)"); + return; + } + + texUnit = _mesa_get_current_tex_unit(ctx); + + texgen = get_texgen(texUnit, coord); + if (!texgen) { + _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexGenfv(coord)"); + return; + } + + switch (pname) { + case GL_TEXTURE_GEN_MODE: + params[0] = ENUM_TO_FLOAT(texgen->Mode); + break; + case GL_OBJECT_PLANE: + COPY_4V(params, texgen->ObjectPlane); + break; + case GL_EYE_PLANE: + COPY_4V(params, texgen->EyePlane); + break; + default: + _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexGenfv(pname)" ); + } +} + + + +static void GLAPIENTRY +_mesa_GetTexGeniv( GLenum coord, GLenum pname, GLint *params ) +{ + struct gl_texture_unit *texUnit; + struct gl_texgen *texgen; + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + if (ctx->Texture.CurrentUnit >= ctx->Const.MaxTextureCoordUnits) { + _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexGeniv(current unit)"); + return; + } + + texUnit = _mesa_get_current_tex_unit(ctx); + + texgen = get_texgen(texUnit, coord); + if (!texgen) { + _mesa_error(ctx, GL_INVALID_ENUM, "glGetTexGeniv(coord)"); + return; + } + + switch (pname) { + case GL_TEXTURE_GEN_MODE: + params[0] = texgen->Mode; + break; + case GL_OBJECT_PLANE: + params[0] = (GLint) texgen->ObjectPlane[0]; + params[1] = (GLint) texgen->ObjectPlane[1]; + params[2] = (GLint) texgen->ObjectPlane[2]; + params[3] = (GLint) texgen->ObjectPlane[3]; + break; + case GL_EYE_PLANE: + params[0] = (GLint) texgen->EyePlane[0]; + params[1] = (GLint) texgen->EyePlane[1]; + params[2] = (GLint) texgen->EyePlane[2]; + params[3] = (GLint) texgen->EyePlane[3]; + break; + default: + _mesa_error( ctx, GL_INVALID_ENUM, "glGetTexGeniv(pname)" ); + } +} + + +void +_mesa_init_texgen_dispatch(struct _glapi_table *disp) +{ + SET_GetTexGendv(disp, _mesa_GetTexGendv); + SET_GetTexGenfv(disp, _mesa_GetTexGenfv); + SET_GetTexGeniv(disp, _mesa_GetTexGeniv); + SET_TexGend(disp, _mesa_TexGend); + SET_TexGendv(disp, _mesa_TexGendv); + SET_TexGenf(disp, _mesa_TexGenf); + SET_TexGenfv(disp, _mesa_TexGenfv); + SET_TexGeni(disp, _mesa_TexGeni); + SET_TexGeniv(disp, _mesa_TexGeniv); +} + + +#endif /* FEATURE_texgen */ diff --git a/mesalib/src/mesa/main/teximage.c b/mesalib/src/mesa/main/teximage.c index 694f6fa00..b16baaf91 100644 --- a/mesalib/src/mesa/main/teximage.c +++ b/mesalib/src/mesa/main/teximage.c @@ -2606,26 +2606,9 @@ teximage(struct gl_context *ctx, GLuint dims, border, internalFormat, texFormat); /* Give the texture to the driver. may be null. */ - ASSERT(ctx->Driver.TexImage3D); - switch (dims) { - case 1: - ctx->Driver.TexImage1D(ctx, texImage, internalFormat, - width, border, format, - type, pixels, unpack); - break; - case 2: - ctx->Driver.TexImage2D(ctx, texImage, internalFormat, - width, height, border, format, - type, pixels, unpack); - break; - case 3: - ctx->Driver.TexImage3D(ctx, texImage, internalFormat, - width, height, depth, border, format, - type, pixels, unpack); - break; - default: - _mesa_problem(ctx, "invalid dims=%u in teximage()", dims); - } + ctx->Driver.TexImage(ctx, dims, texImage, internalFormat, + width, height, depth, border, format, + type, pixels, unpack); check_gen_mipmap(ctx, target, texObj, level); @@ -2809,26 +2792,10 @@ texsubimage(struct gl_context *ctx, GLuint dims, GLenum target, GLint level, xoffset += texImage->Border; } - switch (dims) { - case 1: - ctx->Driver.TexSubImage1D(ctx, texImage, - xoffset, width, - format, type, pixels, &ctx->Unpack); - break; - case 2: - ctx->Driver.TexSubImage2D(ctx, texImage, - xoffset, yoffset, width, height, - format, type, pixels, &ctx->Unpack); - break; - case 3: - ctx->Driver.TexSubImage3D(ctx, texImage, - xoffset, yoffset, zoffset, - width, height, depth, - format, type, pixels, &ctx->Unpack); - break; - default: - _mesa_problem(ctx, "unexpected dims in subteximage()"); - } + ctx->Driver.TexSubImage(ctx, dims, texImage, + xoffset, yoffset, zoffset, + width, height, depth, + format, type, pixels, &ctx->Unpack); check_gen_mipmap(ctx, target, texObj, level); @@ -2959,7 +2926,7 @@ copyteximage(struct gl_context *ctx, GLuint dims, GL_NONE, GL_NONE); if (legal_texture_size(ctx, texFormat, width, height, 1)) { - GLint srcX = x, srcY = y, dstX = 0, dstY = 0; + GLint srcX = x, srcY = y, dstX = 0, dstY = 0, dstZ = 0; /* Free old texture image */ ctx->Driver.FreeTextureImageBuffer(ctx, texImage); @@ -2968,29 +2935,17 @@ copyteximage(struct gl_context *ctx, GLuint dims, border, internalFormat, texFormat); /* Allocate texture memory (no pixel data yet) */ - if (dims == 1) { - ctx->Driver.TexImage1D(ctx, texImage, internalFormat, - width, border, GL_NONE, GL_NONE, NULL, - &ctx->Unpack); - } - else { - ctx->Driver.TexImage2D(ctx, texImage, internalFormat, - width, height, border, GL_NONE, GL_NONE, - NULL, &ctx->Unpack); - } + ctx->Driver.TexImage(ctx, dims, texImage, internalFormat, + width, height, 1, border, GL_NONE, GL_NONE, + NULL, &ctx->Unpack); if (_mesa_clip_copytexsubimage(ctx, &dstX, &dstY, &srcX, &srcY, &width, &height)) { struct gl_renderbuffer *srcRb = get_copy_tex_image_source(ctx, texImage->TexFormat); - if (dims == 1) - ctx->Driver.CopyTexSubImage1D(ctx, texImage, dstX, - srcRb, srcX, srcY, width); - - else - ctx->Driver.CopyTexSubImage2D(ctx, texImage, dstX, dstY, - srcRb, srcX, srcY, width, height); + ctx->Driver.CopyTexSubImage(ctx, dims, texImage, dstX, dstY, dstZ, + srcRb, srcX, srcY, width, height); } check_gen_mipmap(ctx, target, texObj, level); @@ -3089,23 +3044,9 @@ copytexsubimage(struct gl_context *ctx, GLuint dims, GLenum target, GLint level, struct gl_renderbuffer *srcRb = get_copy_tex_image_source(ctx, texImage->TexFormat); - switch (dims) { - case 1: - ctx->Driver.CopyTexSubImage1D(ctx, texImage, xoffset, - srcRb, x, y, width); - break; - case 2: - ctx->Driver.CopyTexSubImage2D(ctx, texImage, xoffset, yoffset, - srcRb, x, y, width, height); - break; - case 3: - ctx->Driver.CopyTexSubImage3D(ctx, texImage, - xoffset, yoffset, zoffset, - srcRb, x, y, width, height); - break; - default: - _mesa_problem(ctx, "bad dims in copytexsubimage()"); - } + ctx->Driver.CopyTexSubImage(ctx, dims, texImage, + xoffset, yoffset, zoffset, + srcRb, x, y, width, height); check_gen_mipmap(ctx, target, texObj, level); @@ -3600,31 +3541,10 @@ compressedteximage(struct gl_context *ctx, GLuint dims, width, height, depth, border, internalFormat, texFormat); - switch (dims) { - case 1: - ASSERT(ctx->Driver.CompressedTexImage1D); - ctx->Driver.CompressedTexImage1D(ctx, texImage, - internalFormat, - width, - border, imageSize, data); - break; - case 2: - ASSERT(ctx->Driver.CompressedTexImage2D); - ctx->Driver.CompressedTexImage2D(ctx, texImage, - internalFormat, - width, height, - border, imageSize, data); - break; - case 3: - ASSERT(ctx->Driver.CompressedTexImage3D); - ctx->Driver.CompressedTexImage3D(ctx, texImage, - internalFormat, - width, height, depth, - border, imageSize, data); - break; - default: - _mesa_problem(ctx, "bad dims in compressedteximage"); - } + ctx->Driver.CompressedTexImage(ctx, dims, texImage, + internalFormat, + width, height, depth, + border, imageSize, data); check_gen_mipmap(ctx, target, texObj, level); @@ -3713,33 +3633,10 @@ compressed_tex_sub_image(GLuint dims, GLenum target, GLint level, /* error was recorded */ } else if (width > 0 && height > 0 && depth > 0) { - switch (dims) { - case 1: - if (ctx->Driver.CompressedTexSubImage1D) { - ctx->Driver.CompressedTexSubImage1D(ctx, texImage, - xoffset, width, - format, imageSize, data); - } - break; - case 2: - if (ctx->Driver.CompressedTexSubImage2D) { - ctx->Driver.CompressedTexSubImage2D(ctx, texImage, - xoffset, yoffset, - width, height, - format, imageSize, data); - } - break; - case 3: - if (ctx->Driver.CompressedTexSubImage3D) { - ctx->Driver.CompressedTexSubImage3D(ctx, texImage, - xoffset, yoffset, zoffset, - width, height, depth, - format, imageSize, data); - } - break; - default: - ; - } + ctx->Driver.CompressedTexSubImage(ctx, dims, texImage, + xoffset, yoffset, zoffset, + width, height, depth, + format, imageSize, data); check_gen_mipmap(ctx, target, texObj, level); diff --git a/mesalib/src/mesa/main/texobj.c b/mesalib/src/mesa/main/texobj.c index a471bad22..e0494c940 100644 --- a/mesalib/src/mesa/main/texobj.c +++ b/mesalib/src/mesa/main/texobj.c @@ -807,30 +807,10 @@ _mesa_get_fallback_texture(struct gl_context *ctx, gl_texture_index tex) 0, /* border */ GL_RGBA, texFormat); - switch (dims) { - case 0: - break; - case 1: - ctx->Driver.TexImage1D(ctx, texImage, GL_RGBA, - width, 0, - GL_RGBA, GL_UNSIGNED_BYTE, texel, - &ctx->DefaultPacking); - break; - case 2: - ctx->Driver.TexImage2D(ctx, texImage, GL_RGBA, - width, height, 0, - GL_RGBA, GL_UNSIGNED_BYTE, texel, - &ctx->DefaultPacking); - break; - case 3: - ctx->Driver.TexImage3D(ctx, texImage, GL_RGBA, - width, height, depth, 0, - GL_RGBA, GL_UNSIGNED_BYTE, texel, - &ctx->DefaultPacking); - break; - default: - _mesa_problem(ctx, "bad dims in _mesa_get_fallback_texture()"); - } + ctx->Driver.TexImage(ctx, dims, texImage, GL_RGBA, + width, height, depth, 0, + GL_RGBA, GL_UNSIGNED_BYTE, texel, + &ctx->DefaultPacking); } _mesa_test_texobj_completeness(ctx, texObj); diff --git a/mesalib/src/mesa/main/texstore.c b/mesalib/src/mesa/main/texstore.c index d368b1868..1ced8aac9 100644 --- a/mesalib/src/mesa/main/texstore.c +++ b/mesalib/src/mesa/main/texstore.c @@ -4261,170 +4261,71 @@ store_texsubimage(struct gl_context *ctx, /** - * This is the fallback for Driver.TexImage1D(). + * Fallback code for ctx->Driver.TexImage(). + * Basically, allocate storage for the texture image, then copy the + * user's image into it. */ void -_mesa_store_teximage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint border, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing) +_mesa_store_teximage(struct gl_context *ctx, + GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLint width, GLint height, GLint depth, GLint border, + GLenum format, GLenum type, const GLvoid *pixels, + const struct gl_pixelstore_attrib *packing) { - if (width == 0) - return; - - /* allocate storage for texture data */ - if (!ctx->Driver.AllocTextureImageBuffer(ctx, texImage, texImage->TexFormat, - width, 1, 1)) { - _mesa_error(ctx, GL_OUT_OF_MEMORY, "glTexImage1D"); - return; - } - - store_texsubimage(ctx, texImage, - 0, 0, 0, width, 1, 1, - format, type, pixels, packing, "glTexImage1D"); -} - - -/** - * This is the fallback for Driver.TexImage2D(). - */ -void -_mesa_store_teximage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint border, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *packing) -{ - if (width == 0 || height == 0) - return; - - /* allocate storage for texture data */ - if (!ctx->Driver.AllocTextureImageBuffer(ctx, texImage, texImage->TexFormat, - width, height, 1)) { - _mesa_error(ctx, GL_OUT_OF_MEMORY, "glTexImage2D"); - return; - } - - store_texsubimage(ctx, texImage, - 0, 0, 0, width, height, 1, - format, type, pixels, packing, "glTexImage2D"); -} - + assert(dims == 1 || dims == 2 || dims == 3); - -/** - * This is the fallback for Driver.TexImage3D(). - */ -void -_mesa_store_teximage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint depth, GLint border, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *packing) -{ if (width == 0 || height == 0 || depth == 0) return; /* allocate storage for texture data */ if (!ctx->Driver.AllocTextureImageBuffer(ctx, texImage, texImage->TexFormat, width, height, depth)) { - _mesa_error(ctx, GL_OUT_OF_MEMORY, "glTexImage3D"); + _mesa_error(ctx, GL_OUT_OF_MEMORY, "glTexImage%uD", dims); return; } store_texsubimage(ctx, texImage, 0, 0, 0, width, height, depth, - format, type, pixels, packing, "glTexImage3D"); + format, type, pixels, packing, "glTexImage"); } - - /* - * This is the fallback for Driver.TexSubImage1D(). + * Fallback for Driver.TexSubImage(). */ void -_mesa_store_texsubimage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint width, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *packing) -{ - store_texsubimage(ctx, texImage, - xoffset, 0, 0, width, 1, 1, - format, type, pixels, packing, "glTexSubImage1D"); -} - - - -/** - * This is the fallback for Driver.TexSubImage2D(). - */ -void -_mesa_store_texsubimage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - GLint width, GLint height, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *packing) -{ - store_texsubimage(ctx, texImage, - xoffset, yoffset, 0, width, height, 1, - format, type, pixels, packing, "glTexSubImage2D"); -} - - -/* - * This is the fallback for Driver.TexSubImage3D(). - */ -void -_mesa_store_texsubimage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - GLint width, GLint height, GLint depth, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *packing) +_mesa_store_texsubimage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + GLint width, GLint height, GLint depth, + GLenum format, GLenum type, const void *pixels, + const struct gl_pixelstore_attrib *packing) { store_texsubimage(ctx, texImage, xoffset, yoffset, zoffset, width, height, depth, - format, type, pixels, packing, "glTexSubImage3D"); -} - - -/* - * Fallback for Driver.CompressedTexImage1D() - */ -void -_mesa_store_compressed_teximage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint border, - GLsizei imageSize, const GLvoid *data) -{ - /* no compressed 1D image formats at this time */ - (void) ctx; - (void) internalFormat; - (void) width; (void) border; - (void) imageSize; (void) data; - (void) texImage; + format, type, pixels, packing, "glTexSubImage"); } - /** - * Fallback for Driver.CompressedTexImage2D() + * Fallback for Driver.CompressedTexImage() */ void -_mesa_store_compressed_teximage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint border, - GLsizei imageSize, const GLvoid *data) +_mesa_store_compressed_teximage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLint width, GLint height, GLint depth, + GLint border, + GLsizei imageSize, const GLvoid *data) { + /* only 2D compressed images are supported at this time */ + if (dims != 2) { + _mesa_problem(ctx, "Unexpected glCompressedTexImage1D/3D call"); + return; + } + /* This is pretty simple, because unlike the general texstore path we don't * have to worry about the usual image unpacking or image transfer * operations. @@ -4441,66 +4342,24 @@ _mesa_store_compressed_teximage2d(struct gl_context *ctx, return; } - _mesa_store_compressed_texsubimage2d(ctx, texImage, - 0, 0, - width, height, - texImage->TexFormat, - imageSize, data); -} - - - -/* - * Fallback for Driver.CompressedTexImage3D() - */ -void -_mesa_store_compressed_teximage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint depth, - GLint border, - GLsizei imageSize, const GLvoid *data) -{ - /* this space intentionally left blank */ - (void) ctx; - (void) internalFormat; - (void) width; (void) height; (void) depth; - (void) border; - (void) imageSize; (void) data; - (void) texImage; -} - - - -/** - * Fallback for Driver.CompressedTexSubImage1D() - */ -void -_mesa_store_compressed_texsubimage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLsizei width, - GLenum format, - GLsizei imageSize, const GLvoid *data) -{ - /* there are no compressed 1D texture formats yet */ - (void) ctx; - (void) xoffset; (void) width; - (void) format; - (void) imageSize; (void) data; - (void) texImage; + _mesa_store_compressed_texsubimage(ctx, dims, texImage, + 0, 0, 0, + width, height, depth, + texImage->TexFormat, + imageSize, data); } /** - * Fallback for Driver.CompressedTexSubImage2D() + * Fallback for Driver.CompressedTexSubImage() */ void -_mesa_store_compressed_texsubimage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - GLsizei width, GLsizei height, - GLenum format, - GLsizei imageSize, const GLvoid *data) +_mesa_store_compressed_texsubimage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + GLsizei width, GLsizei height, GLsizei depth, + GLenum format, + GLsizei imageSize, const GLvoid *data) { GLint bytesPerRow, dstRowStride, srcRowStride; GLint i, rows; @@ -4509,6 +4368,11 @@ _mesa_store_compressed_texsubimage2d(struct gl_context *ctx, const gl_format texFormat = texImage->TexFormat; GLuint bw, bh; + if (dims != 2) { + _mesa_problem(ctx, "Unexpected 1D/3D compressed texsubimage call"); + return; + } + _mesa_get_format_block_size(texFormat, &bw, &bh); /* these should have been caught sooner */ @@ -4552,24 +4416,3 @@ _mesa_store_compressed_texsubimage2d(struct gl_context *ctx, _mesa_unmap_teximage_pbo(ctx, &ctx->Unpack); } - - -/** - * Fallback for Driver.CompressedTexSubImage3D() - */ -void -_mesa_store_compressed_texsubimage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - GLsizei width, GLsizei height, GLsizei depth, - GLenum format, - GLsizei imageSize, const GLvoid *data) -{ - /* there are no compressed 3D texture formats yet */ - (void) ctx; - (void) xoffset; (void) yoffset; (void) zoffset; - (void) width; (void) height; (void) depth; - (void) format; - (void) imageSize; (void) data; - (void) texImage; -} diff --git a/mesalib/src/mesa/main/texstore.h b/mesalib/src/mesa/main/texstore.h index 85e33b0fe..5a1c01423 100644 --- a/mesalib/src/mesa/main/texstore.h +++ b/mesalib/src/mesa/main/texstore.h @@ -90,103 +90,40 @@ _mesa_make_temp_float_image(struct gl_context *ctx, GLuint dims, GLbitfield transferOps); extern void -_mesa_store_teximage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint border, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); +_mesa_store_teximage(struct gl_context *ctx, + GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLint width, GLint height, GLint depth, GLint border, + GLenum format, GLenum type, const GLvoid *pixels, + const struct gl_pixelstore_attrib *packing); extern void -_mesa_store_teximage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint border, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); +_mesa_store_texsubimage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + GLint width, GLint height, GLint depth, + GLenum format, GLenum type, const GLvoid *pixels, + const struct gl_pixelstore_attrib *packing); extern void -_mesa_store_teximage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint depth, GLint border, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); +_mesa_store_compressed_teximage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLint width, GLint height, GLint depth, + GLint border, + GLsizei imageSize, const GLvoid *data); extern void -_mesa_store_texsubimage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint width, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); - - -extern void -_mesa_store_texsubimage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - GLint width, GLint height, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); - - -extern void -_mesa_store_texsubimage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - GLint width, GLint height, GLint depth, - GLenum format, GLenum type, const GLvoid *pixels, - const struct gl_pixelstore_attrib *packing); - - -extern void -_mesa_store_compressed_teximage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint border, - GLsizei imageSize, const GLvoid *data); - -extern void -_mesa_store_compressed_teximage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint border, - GLsizei imageSize, const GLvoid *data); - -extern void -_mesa_store_compressed_teximage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint depth, - GLint border, - GLsizei imageSize, const GLvoid *data); - - -extern void -_mesa_store_compressed_texsubimage1d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLsizei width, - GLenum format, - GLsizei imageSize, const GLvoid *data); - -extern void -_mesa_store_compressed_texsubimage2d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - GLsizei width, GLsizei height, - GLenum format, - GLsizei imageSize, const GLvoid *data); - -extern void -_mesa_store_compressed_texsubimage3d(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - GLsizei width, GLsizei height, GLsizei depth, - GLenum format, - GLsizei imageSize, const GLvoid *data); +_mesa_store_compressed_texsubimage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint xoffset, GLint yoffset, GLint zoffset, + GLsizei width, GLsizei height, GLsizei depth, + GLenum format, + GLsizei imageSize, const GLvoid *data); #endif diff --git a/mesalib/src/mesa/main/texturebarrier.c b/mesalib/src/mesa/main/texturebarrier.c index ef4bd45e6..56cc8718c 100644 --- a/mesalib/src/mesa/main/texturebarrier.c +++ b/mesalib/src/mesa/main/texturebarrier.c @@ -1,54 +1,54 @@ -/* - * Copyright © 2011 Marek Olšák - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -/** - * \file texturebarrier.c - * Implementation of glTextureBarrierNV. - * - * \author Marek Olšák - */ - -#include "context.h" -#include "texturebarrier.h" - - -static void -_mesa_texture_barrier(struct gl_context *ctx) -{ - /* no-op */ -} - -void -_mesa_init_texture_barrier_functions(struct dd_function_table *driver) -{ - driver->TextureBarrier = _mesa_texture_barrier; -} - -void GLAPIENTRY -_mesa_TextureBarrierNV(void) -{ - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END(ctx); - - ctx->Driver.TextureBarrier(ctx); -} +/* + * Copyright © 2011 Marek Olšák + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ + +/** + * \file texturebarrier.c + * Implementation of glTextureBarrierNV. + * + * \author Marek Olšák + */ + +#include "context.h" +#include "texturebarrier.h" + + +static void +_mesa_texture_barrier(struct gl_context *ctx) +{ + /* no-op */ +} + +void +_mesa_init_texture_barrier_functions(struct dd_function_table *driver) +{ + driver->TextureBarrier = _mesa_texture_barrier; +} + +void GLAPIENTRY +_mesa_TextureBarrierNV(void) +{ + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END(ctx); + + ctx->Driver.TextureBarrier(ctx); +} diff --git a/mesalib/src/mesa/main/texturebarrier.h b/mesalib/src/mesa/main/texturebarrier.h index ef35989b4..a84a85b21 100644 --- a/mesalib/src/mesa/main/texturebarrier.h +++ b/mesalib/src/mesa/main/texturebarrier.h @@ -1,44 +1,44 @@ -/* - * Copyright © 2011 Marek Olšák - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -/** - * \file texturebarrier.h - * GL_NV_texture_barrier - * - * \author Marek Olšák - */ - -#ifndef TEXTUREBARRIER_H -#define TEXTUREBARRIER_H - -#include "glheader.h" - -struct dd_function_table; - -extern void -_mesa_init_texture_barrier_functions(struct dd_function_table *driver); - -extern void GLAPIENTRY -_mesa_TextureBarrierNV(void); - -#endif /* TEXTUREBARRIER_H */ +/* + * Copyright © 2011 Marek Olšák + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ + +/** + * \file texturebarrier.h + * GL_NV_texture_barrier + * + * \author Marek Olšák + */ + +#ifndef TEXTUREBARRIER_H +#define TEXTUREBARRIER_H + +#include "glheader.h" + +struct dd_function_table; + +extern void +_mesa_init_texture_barrier_functions(struct dd_function_table *driver); + +extern void GLAPIENTRY +_mesa_TextureBarrierNV(void); + +#endif /* TEXTUREBARRIER_H */ diff --git a/mesalib/src/mesa/main/transformfeedback.c b/mesalib/src/mesa/main/transformfeedback.c index f2c1435d9..1bd76d130 100644 --- a/mesalib/src/mesa/main/transformfeedback.c +++ b/mesalib/src/mesa/main/transformfeedback.c @@ -470,7 +470,12 @@ _mesa_BindBufferRange(GLenum target, GLuint index, return; } - bufObj = _mesa_lookup_bufferobj(ctx, buffer); + if (buffer == 0) { + bufObj = ctx->Shared->NullBufferObj; + } else { + bufObj = _mesa_lookup_bufferobj(ctx, buffer); + } + if (!bufObj) { _mesa_error(ctx, GL_INVALID_OPERATION, "glBindBufferRange(invalid buffer=%u)", buffer); @@ -518,7 +523,12 @@ _mesa_BindBufferBase(GLenum target, GLuint index, GLuint buffer) return; } - bufObj = _mesa_lookup_bufferobj(ctx, buffer); + if (buffer == 0) { + bufObj = ctx->Shared->NullBufferObj; + } else { + bufObj = _mesa_lookup_bufferobj(ctx, buffer); + } + if (!bufObj) { _mesa_error(ctx, GL_INVALID_OPERATION, "glBindBufferBase(invalid buffer=%u)", buffer); @@ -574,7 +584,12 @@ _mesa_BindBufferOffsetEXT(GLenum target, GLuint index, GLuint buffer, return; } - bufObj = _mesa_lookup_bufferobj(ctx, buffer); + if (buffer == 0) { + bufObj = ctx->Shared->NullBufferObj; + } else { + bufObj = _mesa_lookup_bufferobj(ctx, buffer); + } + if (!bufObj) { _mesa_error(ctx, GL_INVALID_OPERATION, "glBindBufferOffsetEXT(invalid buffer=%u)", buffer); diff --git a/mesalib/src/mesa/main/version.c b/mesalib/src/mesa/main/version.c index 90afecb5d..efaaf58de 100644 --- a/mesalib/src/mesa/main/version.c +++ b/mesalib/src/mesa/main/version.c @@ -171,6 +171,7 @@ compute_version(struct gl_context *ctx) ctx->Extensions.ARB_instanced_arrays && ctx->Extensions.ARB_occlusion_query2 && ctx->Extensions.ARB_sampler_objects && + ctx->Extensions.ARB_shader_bit_encoding && ctx->Extensions.ARB_texture_rgb10_a2ui && ctx->Extensions.ARB_timer_query && ctx->Extensions.ARB_vertex_type_2_10_10_10_rev && diff --git a/mesalib/src/mesa/main/viewport.c b/mesalib/src/mesa/main/viewport.c index 9186d1bb8..a58697ab3 100644 --- a/mesalib/src/mesa/main/viewport.c +++ b/mesalib/src/mesa/main/viewport.c @@ -1,185 +1,185 @@ -/* - * Mesa 3-D graphics library - * Version: 7.5 - * - * Copyright (C) 2009 VMware, Inc. All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -/** - * \file viewport.c - * glViewport and glDepthRange functions. - */ - - -#include "context.h" -#include "macros.h" -#include "mtypes.h" -#include "viewport.h" - - -/** - * Set the viewport. - * \sa Called via glViewport() or display list execution. - * - * Flushes the vertices and calls _mesa_set_viewport() with the given - * parameters. - */ -void GLAPIENTRY -_mesa_Viewport(GLint x, GLint y, GLsizei width, GLsizei height) -{ - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); - _mesa_set_viewport(ctx, x, y, width, height); -} - - -/** - * Set new viewport parameters and update derived state (the _WindowMap - * matrix). Usually called from _mesa_Viewport(). - * - * \param ctx GL context. - * \param x, y coordinates of the lower left corner of the viewport rectangle. - * \param width width of the viewport rectangle. - * \param height height of the viewport rectangle. - */ -void -_mesa_set_viewport(struct gl_context *ctx, GLint x, GLint y, - GLsizei width, GLsizei height) -{ - if (MESA_VERBOSE & VERBOSE_API) - _mesa_debug(ctx, "glViewport %d %d %d %d\n", x, y, width, height); - - if (width < 0 || height < 0) { - _mesa_error(ctx, GL_INVALID_VALUE, - "glViewport(%d, %d, %d, %d)", x, y, width, height); - return; - } - - /* clamp width and height to the implementation dependent range */ - width = MIN2(width, (GLsizei) ctx->Const.MaxViewportWidth); - height = MIN2(height, (GLsizei) ctx->Const.MaxViewportHeight); - - ctx->Viewport.X = x; - ctx->Viewport.Width = width; - ctx->Viewport.Y = y; - ctx->Viewport.Height = height; - ctx->NewState |= _NEW_VIEWPORT; - -#if 1 - /* XXX remove this someday. Currently the DRI drivers rely on - * the WindowMap matrix being up to date in the driver's Viewport - * and DepthRange functions. - */ - _math_matrix_viewport(&ctx->Viewport._WindowMap, - ctx->Viewport.X, ctx->Viewport.Y, - ctx->Viewport.Width, ctx->Viewport.Height, - ctx->Viewport.Near, ctx->Viewport.Far, - ctx->DrawBuffer->_DepthMaxF); -#endif - - if (ctx->Driver.Viewport) { - /* Many drivers will use this call to check for window size changes - * and reallocate the z/stencil/accum/etc buffers if needed. - */ - ctx->Driver.Viewport(ctx, x, y, width, height); - } -} - - -/** - * Called by glDepthRange - * - * \param nearval specifies the Z buffer value which should correspond to - * the near clip plane - * \param farval specifies the Z buffer value which should correspond to - * the far clip plane - */ -void GLAPIENTRY -_mesa_DepthRange(GLclampd nearval, GLclampd farval) -{ - GET_CURRENT_CONTEXT(ctx); - ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); - - if (MESA_VERBOSE&VERBOSE_API) - _mesa_debug(ctx, "glDepthRange %f %f\n", nearval, farval); - - if (ctx->Viewport.Near == nearval && - ctx->Viewport.Far == farval) - return; - - ctx->Viewport.Near = (GLfloat) CLAMP(nearval, 0.0, 1.0); - ctx->Viewport.Far = (GLfloat) CLAMP(farval, 0.0, 1.0); - ctx->NewState |= _NEW_VIEWPORT; - -#if 1 - /* XXX remove this someday. Currently the DRI drivers rely on - * the WindowMap matrix being up to date in the driver's Viewport - * and DepthRange functions. - */ - _math_matrix_viewport(&ctx->Viewport._WindowMap, - ctx->Viewport.X, ctx->Viewport.Y, - ctx->Viewport.Width, ctx->Viewport.Height, - ctx->Viewport.Near, ctx->Viewport.Far, - ctx->DrawBuffer->_DepthMaxF); -#endif - - if (ctx->Driver.DepthRange) { - ctx->Driver.DepthRange(ctx, nearval, farval); - } -} - -void GLAPIENTRY -_mesa_DepthRangef(GLclampf nearval, GLclampf farval) -{ - _mesa_DepthRange(nearval, farval); -} - -/** - * Initialize the context viewport attribute group. - * \param ctx the GL context. - */ -void _mesa_init_viewport(struct gl_context *ctx) -{ - GLfloat depthMax = 65535.0F; /* sorf of arbitrary */ - - /* Viewport group */ - ctx->Viewport.X = 0; - ctx->Viewport.Y = 0; - ctx->Viewport.Width = 0; - ctx->Viewport.Height = 0; - ctx->Viewport.Near = 0.0; - ctx->Viewport.Far = 1.0; - _math_matrix_ctr(&ctx->Viewport._WindowMap); - - _math_matrix_viewport(&ctx->Viewport._WindowMap, 0, 0, 0, 0, - 0.0F, 1.0F, depthMax); -} - - -/** - * Free the context viewport attribute group data. - * \param ctx the GL context. - */ -void _mesa_free_viewport_data(struct gl_context *ctx) -{ - _math_matrix_dtr(&ctx->Viewport._WindowMap); -} - +/* + * Mesa 3-D graphics library + * Version: 7.5 + * + * Copyright (C) 2009 VMware, Inc. All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +/** + * \file viewport.c + * glViewport and glDepthRange functions. + */ + + +#include "context.h" +#include "macros.h" +#include "mtypes.h" +#include "viewport.h" + + +/** + * Set the viewport. + * \sa Called via glViewport() or display list execution. + * + * Flushes the vertices and calls _mesa_set_viewport() with the given + * parameters. + */ +void GLAPIENTRY +_mesa_Viewport(GLint x, GLint y, GLsizei width, GLsizei height) +{ + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); + _mesa_set_viewport(ctx, x, y, width, height); +} + + +/** + * Set new viewport parameters and update derived state (the _WindowMap + * matrix). Usually called from _mesa_Viewport(). + * + * \param ctx GL context. + * \param x, y coordinates of the lower left corner of the viewport rectangle. + * \param width width of the viewport rectangle. + * \param height height of the viewport rectangle. + */ +void +_mesa_set_viewport(struct gl_context *ctx, GLint x, GLint y, + GLsizei width, GLsizei height) +{ + if (MESA_VERBOSE & VERBOSE_API) + _mesa_debug(ctx, "glViewport %d %d %d %d\n", x, y, width, height); + + if (width < 0 || height < 0) { + _mesa_error(ctx, GL_INVALID_VALUE, + "glViewport(%d, %d, %d, %d)", x, y, width, height); + return; + } + + /* clamp width and height to the implementation dependent range */ + width = MIN2(width, (GLsizei) ctx->Const.MaxViewportWidth); + height = MIN2(height, (GLsizei) ctx->Const.MaxViewportHeight); + + ctx->Viewport.X = x; + ctx->Viewport.Width = width; + ctx->Viewport.Y = y; + ctx->Viewport.Height = height; + ctx->NewState |= _NEW_VIEWPORT; + +#if 1 + /* XXX remove this someday. Currently the DRI drivers rely on + * the WindowMap matrix being up to date in the driver's Viewport + * and DepthRange functions. + */ + _math_matrix_viewport(&ctx->Viewport._WindowMap, + ctx->Viewport.X, ctx->Viewport.Y, + ctx->Viewport.Width, ctx->Viewport.Height, + ctx->Viewport.Near, ctx->Viewport.Far, + ctx->DrawBuffer->_DepthMaxF); +#endif + + if (ctx->Driver.Viewport) { + /* Many drivers will use this call to check for window size changes + * and reallocate the z/stencil/accum/etc buffers if needed. + */ + ctx->Driver.Viewport(ctx, x, y, width, height); + } +} + + +/** + * Called by glDepthRange + * + * \param nearval specifies the Z buffer value which should correspond to + * the near clip plane + * \param farval specifies the Z buffer value which should correspond to + * the far clip plane + */ +void GLAPIENTRY +_mesa_DepthRange(GLclampd nearval, GLclampd farval) +{ + GET_CURRENT_CONTEXT(ctx); + ASSERT_OUTSIDE_BEGIN_END_AND_FLUSH(ctx); + + if (MESA_VERBOSE&VERBOSE_API) + _mesa_debug(ctx, "glDepthRange %f %f\n", nearval, farval); + + if (ctx->Viewport.Near == nearval && + ctx->Viewport.Far == farval) + return; + + ctx->Viewport.Near = (GLfloat) CLAMP(nearval, 0.0, 1.0); + ctx->Viewport.Far = (GLfloat) CLAMP(farval, 0.0, 1.0); + ctx->NewState |= _NEW_VIEWPORT; + +#if 1 + /* XXX remove this someday. Currently the DRI drivers rely on + * the WindowMap matrix being up to date in the driver's Viewport + * and DepthRange functions. + */ + _math_matrix_viewport(&ctx->Viewport._WindowMap, + ctx->Viewport.X, ctx->Viewport.Y, + ctx->Viewport.Width, ctx->Viewport.Height, + ctx->Viewport.Near, ctx->Viewport.Far, + ctx->DrawBuffer->_DepthMaxF); +#endif + + if (ctx->Driver.DepthRange) { + ctx->Driver.DepthRange(ctx, nearval, farval); + } +} + +void GLAPIENTRY +_mesa_DepthRangef(GLclampf nearval, GLclampf farval) +{ + _mesa_DepthRange(nearval, farval); +} + +/** + * Initialize the context viewport attribute group. + * \param ctx the GL context. + */ +void _mesa_init_viewport(struct gl_context *ctx) +{ + GLfloat depthMax = 65535.0F; /* sorf of arbitrary */ + + /* Viewport group */ + ctx->Viewport.X = 0; + ctx->Viewport.Y = 0; + ctx->Viewport.Width = 0; + ctx->Viewport.Height = 0; + ctx->Viewport.Near = 0.0; + ctx->Viewport.Far = 1.0; + _math_matrix_ctr(&ctx->Viewport._WindowMap); + + _math_matrix_viewport(&ctx->Viewport._WindowMap, 0, 0, 0, 0, + 0.0F, 1.0F, depthMax); +} + + +/** + * Free the context viewport attribute group data. + * \param ctx the GL context. + */ +void _mesa_free_viewport_data(struct gl_context *ctx) +{ + _math_matrix_dtr(&ctx->Viewport._WindowMap); +} + diff --git a/mesalib/src/mesa/main/viewport.h b/mesalib/src/mesa/main/viewport.h index 2f66c6a67..db4507bea 100644 --- a/mesalib/src/mesa/main/viewport.h +++ b/mesalib/src/mesa/main/viewport.h @@ -1,58 +1,58 @@ -/* - * Mesa 3-D graphics library - * Version: 7.5 - * - * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. - * Copyright (C) 2009 VMware, Inc. All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef VIEWPORT_H -#define VIEWPORT_H - -#include "glheader.h" - -struct gl_context; - -extern void GLAPIENTRY -_mesa_Viewport(GLint x, GLint y, GLsizei width, GLsizei height); - - -extern void -_mesa_set_viewport(struct gl_context *ctx, GLint x, GLint y, - GLsizei width, GLsizei height); - - -extern void GLAPIENTRY -_mesa_DepthRange(GLclampd nearval, GLclampd farval); - -extern void GLAPIENTRY -_mesa_DepthRangef(GLclampf nearval, GLclampf farval); - - -extern void -_mesa_init_viewport(struct gl_context *ctx); - - -extern void -_mesa_free_viewport_data(struct gl_context *ctx); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 7.5 + * + * Copyright (C) 1999-2008 Brian Paul All Rights Reserved. + * Copyright (C) 2009 VMware, Inc. All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef VIEWPORT_H +#define VIEWPORT_H + +#include "glheader.h" + +struct gl_context; + +extern void GLAPIENTRY +_mesa_Viewport(GLint x, GLint y, GLsizei width, GLsizei height); + + +extern void +_mesa_set_viewport(struct gl_context *ctx, GLint x, GLint y, + GLsizei width, GLsizei height); + + +extern void GLAPIENTRY +_mesa_DepthRange(GLclampd nearval, GLclampd farval); + +extern void GLAPIENTRY +_mesa_DepthRangef(GLclampf nearval, GLclampf farval); + + +extern void +_mesa_init_viewport(struct gl_context *ctx); + + +extern void +_mesa_free_viewport_data(struct gl_context *ctx); + + +#endif diff --git a/mesalib/src/mesa/math/m_debug_clip.c b/mesalib/src/mesa/math/m_debug_clip.c index 36d2a9e64..bbad6ef02 100644 --- a/mesalib/src/mesa/math/m_debug_clip.c +++ b/mesalib/src/mesa/math/m_debug_clip.c @@ -1,409 +1,409 @@ -/* - * Mesa 3-D graphics library - * Version: 6.1 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Gareth Hughes - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "main/macros.h" -#include "main/imports.h" - -#include "m_matrix.h" -#include "m_xform.h" - -#include "m_debug.h" -#include "m_debug_util.h" - -#ifdef __UNIXOS2__ -/* The linker doesn't like empty files */ -static char dummy; -#endif - -#ifdef DEBUG_MATH /* This code only used for debugging */ - -static clip_func *clip_tab[2] = { - _mesa_clip_tab, - _mesa_clip_np_tab -}; -static char *cnames[2] = { - "_mesa_clip_tab", - "_mesa_clip_np_tab" -}; -#ifdef RUN_DEBUG_BENCHMARK -static char *cstrings[2] = { - "clip, perspective divide", - "clip, no divide" -}; -#endif - - -/* ============================================================= - * Reference cliptests - */ - -static GLvector4f *ref_cliptest_points4( GLvector4f *clip_vec, - GLvector4f *proj_vec, - GLubyte clipMask[], - GLubyte *orMask, - GLubyte *andMask, - GLboolean viewport_z_clip ) -{ - const GLuint stride = clip_vec->stride; - const GLuint count = clip_vec->count; - const GLfloat *from = (GLfloat *)clip_vec->start; - GLuint c = 0; - GLfloat (*vProj)[4] = (GLfloat (*)[4])proj_vec->start; - GLubyte tmpAndMask = *andMask; - GLubyte tmpOrMask = *orMask; - GLuint i; - for ( i = 0 ; i < count ; i++, STRIDE_F(from, stride) ) { - const GLfloat cx = from[0]; - const GLfloat cy = from[1]; - const GLfloat cz = from[2]; - const GLfloat cw = from[3]; - GLubyte mask = 0; - if ( -cx + cw < 0 ) mask |= CLIP_RIGHT_BIT; - if ( cx + cw < 0 ) mask |= CLIP_LEFT_BIT; - if ( -cy + cw < 0 ) mask |= CLIP_TOP_BIT; - if ( cy + cw < 0 ) mask |= CLIP_BOTTOM_BIT; - if (viewport_z_clip) { - if ( -cz + cw < 0 ) mask |= CLIP_FAR_BIT; - if ( cz + cw < 0 ) mask |= CLIP_NEAR_BIT; - } - clipMask[i] = mask; - if ( mask ) { - c++; - tmpAndMask &= mask; - tmpOrMask |= mask; - vProj[i][0] = 0; - vProj[i][1] = 0; - vProj[i][2] = 0; - vProj[i][3] = 1; - } else { - GLfloat oow = 1.0F / cw; - vProj[i][0] = cx * oow; - vProj[i][1] = cy * oow; - vProj[i][2] = cz * oow; - vProj[i][3] = oow; - } - } - - *orMask = tmpOrMask; - *andMask = (GLubyte) (c < count ? 0 : tmpAndMask); - - proj_vec->flags |= VEC_SIZE_4; - proj_vec->size = 4; - proj_vec->count = clip_vec->count; - return proj_vec; -} - -/* Keep these here for now, even though we don't use them... - */ -static GLvector4f *ref_cliptest_points3( GLvector4f *clip_vec, - GLvector4f *proj_vec, - GLubyte clipMask[], - GLubyte *orMask, - GLubyte *andMask, - GLboolean viewport_z_clip ) -{ - const GLuint stride = clip_vec->stride; - const GLuint count = clip_vec->count; - const GLfloat *from = (GLfloat *)clip_vec->start; - - GLubyte tmpOrMask = *orMask; - GLubyte tmpAndMask = *andMask; - GLuint i; - for ( i = 0 ; i < count ; i++, STRIDE_F(from, stride) ) { - const GLfloat cx = from[0], cy = from[1], cz = from[2]; - GLubyte mask = 0; - if ( cx > 1.0 ) mask |= CLIP_RIGHT_BIT; - else if ( cx < -1.0 ) mask |= CLIP_LEFT_BIT; - if ( cy > 1.0 ) mask |= CLIP_TOP_BIT; - else if ( cy < -1.0 ) mask |= CLIP_BOTTOM_BIT; - if (viewport_z_clip) { - if ( cz > 1.0 ) mask |= CLIP_FAR_BIT; - else if ( cz < -1.0 ) mask |= CLIP_NEAR_BIT; - } - clipMask[i] = mask; - tmpOrMask |= mask; - tmpAndMask &= mask; - } - - *orMask = tmpOrMask; - *andMask = tmpAndMask; - return clip_vec; -} - -static GLvector4f * ref_cliptest_points2( GLvector4f *clip_vec, - GLvector4f *proj_vec, - GLubyte clipMask[], - GLubyte *orMask, - GLubyte *andMask, - GLboolean viewport_z_clip ) -{ - const GLuint stride = clip_vec->stride; - const GLuint count = clip_vec->count; - const GLfloat *from = (GLfloat *)clip_vec->start; - - GLubyte tmpOrMask = *orMask; - GLubyte tmpAndMask = *andMask; - GLuint i; - - (void) viewport_z_clip; - - for ( i = 0 ; i < count ; i++, STRIDE_F(from, stride) ) { - const GLfloat cx = from[0], cy = from[1]; - GLubyte mask = 0; - if ( cx > 1.0 ) mask |= CLIP_RIGHT_BIT; - else if ( cx < -1.0 ) mask |= CLIP_LEFT_BIT; - if ( cy > 1.0 ) mask |= CLIP_TOP_BIT; - else if ( cy < -1.0 ) mask |= CLIP_BOTTOM_BIT; - clipMask[i] = mask; - tmpOrMask |= mask; - tmpAndMask &= mask; - } - - *orMask = tmpOrMask; - *andMask = tmpAndMask; - return clip_vec; -} - -static clip_func ref_cliptest[5] = { - 0, - 0, - ref_cliptest_points2, - ref_cliptest_points3, - ref_cliptest_points4 -}; - - -/* ============================================================= - * Cliptest tests - */ - -ALIGN16(static GLfloat, s[TEST_COUNT][4]); -ALIGN16(static GLfloat, d[TEST_COUNT][4]); -ALIGN16(static GLfloat, r[TEST_COUNT][4]); - - -/** - * Check if X, Y or Z component of the coordinate is close to W, in terms - * of the clip test. - */ -static GLboolean -xyz_close_to_w(const GLfloat c[4]) -{ - float k = 0.0001; - return (fabs(c[0] - c[3]) < k || - fabs(c[1] - c[3]) < k || - fabs(c[2] - c[3]) < k || - fabs(-c[0] - c[3]) < k || - fabs(-c[1] - c[3]) < k || - fabs(-c[2] - c[3]) < k); -} - - - -static int test_cliptest_function( clip_func func, int np, - int psize, long *cycles ) -{ - GLvector4f source[1], dest[1], ref[1]; - GLubyte dm[TEST_COUNT], dco, dca; - GLubyte rm[TEST_COUNT], rco, rca; - int i, j; -#ifdef RUN_DEBUG_BENCHMARK - int cycle_i; /* the counter for the benchmarks we run */ -#endif - GLboolean viewport_z_clip = GL_TRUE; - - (void) cycles; - - if ( psize > 4 ) { - _mesa_problem( NULL, "test_cliptest_function called with psize > 4\n" ); - return 0; - } - - for ( i = 0 ; i < TEST_COUNT ; i++) { - ASSIGN_4V( d[i], 0.0, 0.0, 0.0, 1.0 ); - ASSIGN_4V( s[i], 0.0, 0.0, 0.0, 1.0 ); - for ( j = 0 ; j < psize ; j++ ) - s[i][j] = rnd(); - } - - source->data = (GLfloat(*)[4])s; - source->start = (GLfloat *)s; - source->count = TEST_COUNT; - source->stride = sizeof(s[0]); - source->size = 4; - source->flags = 0; - - dest->data = (GLfloat(*)[4])d; - dest->start = (GLfloat *)d; - dest->count = TEST_COUNT; - dest->stride = sizeof(float[4]); - dest->size = 0; - dest->flags = 0; - - ref->data = (GLfloat(*)[4])r; - ref->start = (GLfloat *)r; - ref->count = TEST_COUNT; - ref->stride = sizeof(float[4]); - ref->size = 0; - ref->flags = 0; - - dco = rco = 0; - dca = rca = CLIP_FRUSTUM_BITS; - - ref_cliptest[psize]( source, ref, rm, &rco, &rca, viewport_z_clip ); - - if ( mesa_profile ) { - BEGIN_RACE( *cycles ); - func( source, dest, dm, &dco, &dca, viewport_z_clip ); - END_RACE( *cycles ); - } - else { - func( source, dest, dm, &dco, &dca, viewport_z_clip ); - } - - if ( dco != rco ) { - printf( "\n-----------------------------\n" ); - printf( "dco = 0x%02x rco = 0x%02x\n", dco, rco ); - return 0; - } - if ( dca != rca ) { - printf( "\n-----------------------------\n" ); - printf( "dca = 0x%02x rca = 0x%02x\n", dca, rca ); - return 0; - } - for ( i = 0 ; i < TEST_COUNT ; i++ ) { - if ( dm[i] != rm[i] ) { - GLfloat *c = source->start; - STRIDE_F(c, source->stride * i); - if (psize == 4 && xyz_close_to_w(c)) { - /* The coordinate is very close to the clip plane. The clipmask - * may vary depending on code path, but that's OK. - */ - continue; - } - printf( "\n-----------------------------\n" ); - printf( "mask[%d] = 0x%02x ref mask[%d] = 0x%02x\n", i, dm[i], i,rm[i] ); - printf(" coord = %f, %f, %f, %f\n", - c[0], c[1], c[2], c[3]); - return 0; - } - } - - /* Only verify output on projected points4 case. FIXME: Do we need - * to test other cases? - */ - if ( np || psize < 4 ) - return 1; - - for ( i = 0 ; i < TEST_COUNT ; i++ ) { - for ( j = 0 ; j < 4 ; j++ ) { - if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) { - printf( "\n-----------------------------\n" ); - printf( "(i = %i, j = %i) dm = 0x%02x rm = 0x%02x\n", - i, j, dm[i], rm[i] ); - printf( "%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][0], r[i][0], r[i][0]-d[i][0], - MAX_PRECISION - significand_match( d[i][0], r[i][0] ) ); - printf( "%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][1], r[i][1], r[i][1]-d[i][1], - MAX_PRECISION - significand_match( d[i][1], r[i][1] ) ); - printf( "%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][2], r[i][2], r[i][2]-d[i][2], - MAX_PRECISION - significand_match( d[i][2], r[i][2] ) ); - printf( "%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][3], r[i][3], r[i][3]-d[i][3], - MAX_PRECISION - significand_match( d[i][3], r[i][3] ) ); - return 0; - } - } - } - - return 1; -} - -void _math_test_all_cliptest_functions( char *description ) -{ - int np, psize; - long benchmark_tab[2][4]; - static int first_time = 1; - - if ( first_time ) { - first_time = 0; - mesa_profile = _mesa_getenv( "MESA_PROFILE" ); - } - -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) { - if ( !counter_overhead ) { - INIT_COUNTER(); - printf( "counter overhead: %ld cycles\n\n", counter_overhead ); - } - printf( "cliptest results after hooking in %s functions:\n", description ); - } -#endif - -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) { - printf( "\n\t" ); - for ( psize = 2 ; psize <= 4 ; psize++ ) { - printf( " p%d\t", psize ); - } - printf( "\n--------------------------------------------------------\n\t" ); - } -#endif - - for ( np = 0 ; np < 2 ; np++ ) { - for ( psize = 2 ; psize <= 4 ; psize++ ) { - clip_func func = clip_tab[np][psize]; - long *cycles = &(benchmark_tab[np][psize-1]); - - if ( test_cliptest_function( func, np, psize, cycles ) == 0 ) { - char buf[100]; - sprintf( buf, "%s[%d] failed test (%s)", - cnames[np], psize, description ); - _mesa_problem( NULL, "%s", buf ); - } -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) - printf( " %li\t", benchmark_tab[np][psize-1] ); -#endif - } -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) - printf( " | [%s]\n\t", cstrings[np] ); -#endif - } -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) - printf( "\n" ); -#endif -} - - -#endif /* DEBUG_MATH */ +/* + * Mesa 3-D graphics library + * Version: 6.1 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Gareth Hughes + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "main/macros.h" +#include "main/imports.h" + +#include "m_matrix.h" +#include "m_xform.h" + +#include "m_debug.h" +#include "m_debug_util.h" + +#ifdef __UNIXOS2__ +/* The linker doesn't like empty files */ +static char dummy; +#endif + +#ifdef DEBUG_MATH /* This code only used for debugging */ + +static clip_func *clip_tab[2] = { + _mesa_clip_tab, + _mesa_clip_np_tab +}; +static char *cnames[2] = { + "_mesa_clip_tab", + "_mesa_clip_np_tab" +}; +#ifdef RUN_DEBUG_BENCHMARK +static char *cstrings[2] = { + "clip, perspective divide", + "clip, no divide" +}; +#endif + + +/* ============================================================= + * Reference cliptests + */ + +static GLvector4f *ref_cliptest_points4( GLvector4f *clip_vec, + GLvector4f *proj_vec, + GLubyte clipMask[], + GLubyte *orMask, + GLubyte *andMask, + GLboolean viewport_z_clip ) +{ + const GLuint stride = clip_vec->stride; + const GLuint count = clip_vec->count; + const GLfloat *from = (GLfloat *)clip_vec->start; + GLuint c = 0; + GLfloat (*vProj)[4] = (GLfloat (*)[4])proj_vec->start; + GLubyte tmpAndMask = *andMask; + GLubyte tmpOrMask = *orMask; + GLuint i; + for ( i = 0 ; i < count ; i++, STRIDE_F(from, stride) ) { + const GLfloat cx = from[0]; + const GLfloat cy = from[1]; + const GLfloat cz = from[2]; + const GLfloat cw = from[3]; + GLubyte mask = 0; + if ( -cx + cw < 0 ) mask |= CLIP_RIGHT_BIT; + if ( cx + cw < 0 ) mask |= CLIP_LEFT_BIT; + if ( -cy + cw < 0 ) mask |= CLIP_TOP_BIT; + if ( cy + cw < 0 ) mask |= CLIP_BOTTOM_BIT; + if (viewport_z_clip) { + if ( -cz + cw < 0 ) mask |= CLIP_FAR_BIT; + if ( cz + cw < 0 ) mask |= CLIP_NEAR_BIT; + } + clipMask[i] = mask; + if ( mask ) { + c++; + tmpAndMask &= mask; + tmpOrMask |= mask; + vProj[i][0] = 0; + vProj[i][1] = 0; + vProj[i][2] = 0; + vProj[i][3] = 1; + } else { + GLfloat oow = 1.0F / cw; + vProj[i][0] = cx * oow; + vProj[i][1] = cy * oow; + vProj[i][2] = cz * oow; + vProj[i][3] = oow; + } + } + + *orMask = tmpOrMask; + *andMask = (GLubyte) (c < count ? 0 : tmpAndMask); + + proj_vec->flags |= VEC_SIZE_4; + proj_vec->size = 4; + proj_vec->count = clip_vec->count; + return proj_vec; +} + +/* Keep these here for now, even though we don't use them... + */ +static GLvector4f *ref_cliptest_points3( GLvector4f *clip_vec, + GLvector4f *proj_vec, + GLubyte clipMask[], + GLubyte *orMask, + GLubyte *andMask, + GLboolean viewport_z_clip ) +{ + const GLuint stride = clip_vec->stride; + const GLuint count = clip_vec->count; + const GLfloat *from = (GLfloat *)clip_vec->start; + + GLubyte tmpOrMask = *orMask; + GLubyte tmpAndMask = *andMask; + GLuint i; + for ( i = 0 ; i < count ; i++, STRIDE_F(from, stride) ) { + const GLfloat cx = from[0], cy = from[1], cz = from[2]; + GLubyte mask = 0; + if ( cx > 1.0 ) mask |= CLIP_RIGHT_BIT; + else if ( cx < -1.0 ) mask |= CLIP_LEFT_BIT; + if ( cy > 1.0 ) mask |= CLIP_TOP_BIT; + else if ( cy < -1.0 ) mask |= CLIP_BOTTOM_BIT; + if (viewport_z_clip) { + if ( cz > 1.0 ) mask |= CLIP_FAR_BIT; + else if ( cz < -1.0 ) mask |= CLIP_NEAR_BIT; + } + clipMask[i] = mask; + tmpOrMask |= mask; + tmpAndMask &= mask; + } + + *orMask = tmpOrMask; + *andMask = tmpAndMask; + return clip_vec; +} + +static GLvector4f * ref_cliptest_points2( GLvector4f *clip_vec, + GLvector4f *proj_vec, + GLubyte clipMask[], + GLubyte *orMask, + GLubyte *andMask, + GLboolean viewport_z_clip ) +{ + const GLuint stride = clip_vec->stride; + const GLuint count = clip_vec->count; + const GLfloat *from = (GLfloat *)clip_vec->start; + + GLubyte tmpOrMask = *orMask; + GLubyte tmpAndMask = *andMask; + GLuint i; + + (void) viewport_z_clip; + + for ( i = 0 ; i < count ; i++, STRIDE_F(from, stride) ) { + const GLfloat cx = from[0], cy = from[1]; + GLubyte mask = 0; + if ( cx > 1.0 ) mask |= CLIP_RIGHT_BIT; + else if ( cx < -1.0 ) mask |= CLIP_LEFT_BIT; + if ( cy > 1.0 ) mask |= CLIP_TOP_BIT; + else if ( cy < -1.0 ) mask |= CLIP_BOTTOM_BIT; + clipMask[i] = mask; + tmpOrMask |= mask; + tmpAndMask &= mask; + } + + *orMask = tmpOrMask; + *andMask = tmpAndMask; + return clip_vec; +} + +static clip_func ref_cliptest[5] = { + 0, + 0, + ref_cliptest_points2, + ref_cliptest_points3, + ref_cliptest_points4 +}; + + +/* ============================================================= + * Cliptest tests + */ + +ALIGN16(static GLfloat, s[TEST_COUNT][4]); +ALIGN16(static GLfloat, d[TEST_COUNT][4]); +ALIGN16(static GLfloat, r[TEST_COUNT][4]); + + +/** + * Check if X, Y or Z component of the coordinate is close to W, in terms + * of the clip test. + */ +static GLboolean +xyz_close_to_w(const GLfloat c[4]) +{ + float k = 0.0001; + return (fabs(c[0] - c[3]) < k || + fabs(c[1] - c[3]) < k || + fabs(c[2] - c[3]) < k || + fabs(-c[0] - c[3]) < k || + fabs(-c[1] - c[3]) < k || + fabs(-c[2] - c[3]) < k); +} + + + +static int test_cliptest_function( clip_func func, int np, + int psize, long *cycles ) +{ + GLvector4f source[1], dest[1], ref[1]; + GLubyte dm[TEST_COUNT], dco, dca; + GLubyte rm[TEST_COUNT], rco, rca; + int i, j; +#ifdef RUN_DEBUG_BENCHMARK + int cycle_i; /* the counter for the benchmarks we run */ +#endif + GLboolean viewport_z_clip = GL_TRUE; + + (void) cycles; + + if ( psize > 4 ) { + _mesa_problem( NULL, "test_cliptest_function called with psize > 4\n" ); + return 0; + } + + for ( i = 0 ; i < TEST_COUNT ; i++) { + ASSIGN_4V( d[i], 0.0, 0.0, 0.0, 1.0 ); + ASSIGN_4V( s[i], 0.0, 0.0, 0.0, 1.0 ); + for ( j = 0 ; j < psize ; j++ ) + s[i][j] = rnd(); + } + + source->data = (GLfloat(*)[4])s; + source->start = (GLfloat *)s; + source->count = TEST_COUNT; + source->stride = sizeof(s[0]); + source->size = 4; + source->flags = 0; + + dest->data = (GLfloat(*)[4])d; + dest->start = (GLfloat *)d; + dest->count = TEST_COUNT; + dest->stride = sizeof(float[4]); + dest->size = 0; + dest->flags = 0; + + ref->data = (GLfloat(*)[4])r; + ref->start = (GLfloat *)r; + ref->count = TEST_COUNT; + ref->stride = sizeof(float[4]); + ref->size = 0; + ref->flags = 0; + + dco = rco = 0; + dca = rca = CLIP_FRUSTUM_BITS; + + ref_cliptest[psize]( source, ref, rm, &rco, &rca, viewport_z_clip ); + + if ( mesa_profile ) { + BEGIN_RACE( *cycles ); + func( source, dest, dm, &dco, &dca, viewport_z_clip ); + END_RACE( *cycles ); + } + else { + func( source, dest, dm, &dco, &dca, viewport_z_clip ); + } + + if ( dco != rco ) { + printf( "\n-----------------------------\n" ); + printf( "dco = 0x%02x rco = 0x%02x\n", dco, rco ); + return 0; + } + if ( dca != rca ) { + printf( "\n-----------------------------\n" ); + printf( "dca = 0x%02x rca = 0x%02x\n", dca, rca ); + return 0; + } + for ( i = 0 ; i < TEST_COUNT ; i++ ) { + if ( dm[i] != rm[i] ) { + GLfloat *c = source->start; + STRIDE_F(c, source->stride * i); + if (psize == 4 && xyz_close_to_w(c)) { + /* The coordinate is very close to the clip plane. The clipmask + * may vary depending on code path, but that's OK. + */ + continue; + } + printf( "\n-----------------------------\n" ); + printf( "mask[%d] = 0x%02x ref mask[%d] = 0x%02x\n", i, dm[i], i,rm[i] ); + printf(" coord = %f, %f, %f, %f\n", + c[0], c[1], c[2], c[3]); + return 0; + } + } + + /* Only verify output on projected points4 case. FIXME: Do we need + * to test other cases? + */ + if ( np || psize < 4 ) + return 1; + + for ( i = 0 ; i < TEST_COUNT ; i++ ) { + for ( j = 0 ; j < 4 ; j++ ) { + if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) { + printf( "\n-----------------------------\n" ); + printf( "(i = %i, j = %i) dm = 0x%02x rm = 0x%02x\n", + i, j, dm[i], rm[i] ); + printf( "%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][0], r[i][0], r[i][0]-d[i][0], + MAX_PRECISION - significand_match( d[i][0], r[i][0] ) ); + printf( "%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][1], r[i][1], r[i][1]-d[i][1], + MAX_PRECISION - significand_match( d[i][1], r[i][1] ) ); + printf( "%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][2], r[i][2], r[i][2]-d[i][2], + MAX_PRECISION - significand_match( d[i][2], r[i][2] ) ); + printf( "%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][3], r[i][3], r[i][3]-d[i][3], + MAX_PRECISION - significand_match( d[i][3], r[i][3] ) ); + return 0; + } + } + } + + return 1; +} + +void _math_test_all_cliptest_functions( char *description ) +{ + int np, psize; + long benchmark_tab[2][4]; + static int first_time = 1; + + if ( first_time ) { + first_time = 0; + mesa_profile = _mesa_getenv( "MESA_PROFILE" ); + } + +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) { + if ( !counter_overhead ) { + INIT_COUNTER(); + printf( "counter overhead: %ld cycles\n\n", counter_overhead ); + } + printf( "cliptest results after hooking in %s functions:\n", description ); + } +#endif + +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) { + printf( "\n\t" ); + for ( psize = 2 ; psize <= 4 ; psize++ ) { + printf( " p%d\t", psize ); + } + printf( "\n--------------------------------------------------------\n\t" ); + } +#endif + + for ( np = 0 ; np < 2 ; np++ ) { + for ( psize = 2 ; psize <= 4 ; psize++ ) { + clip_func func = clip_tab[np][psize]; + long *cycles = &(benchmark_tab[np][psize-1]); + + if ( test_cliptest_function( func, np, psize, cycles ) == 0 ) { + char buf[100]; + sprintf( buf, "%s[%d] failed test (%s)", + cnames[np], psize, description ); + _mesa_problem( NULL, "%s", buf ); + } +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) + printf( " %li\t", benchmark_tab[np][psize-1] ); +#endif + } +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) + printf( " | [%s]\n\t", cstrings[np] ); +#endif + } +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) + printf( "\n" ); +#endif +} + + +#endif /* DEBUG_MATH */ diff --git a/mesalib/src/mesa/math/m_debug_norm.c b/mesalib/src/mesa/math/m_debug_norm.c index eae37c225..02eb1f989 100644 --- a/mesalib/src/mesa/math/m_debug_norm.c +++ b/mesalib/src/mesa/math/m_debug_norm.c @@ -1,383 +1,383 @@ - -/* - * Mesa 3-D graphics library - * Version: 5.1 - * - * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Gareth Hughes - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "main/macros.h" -#include "main/imports.h" - -#include "m_matrix.h" -#include "m_xform.h" - -#include "m_debug.h" -#include "m_debug_util.h" - - -#ifdef __UNIXOS2__ -/* The linker doesn't like empty files */ -static char dummy; -#endif - -#ifdef DEBUG_MATH /* This code only used for debugging */ - - -static int m_norm_identity[16] = { - ONE, NIL, NIL, NIL, - NIL, ONE, NIL, NIL, - NIL, NIL, ONE, NIL, - NIL, NIL, NIL, NIL -}; -static int m_norm_general[16] = { - VAR, VAR, VAR, NIL, - VAR, VAR, VAR, NIL, - VAR, VAR, VAR, NIL, - NIL, NIL, NIL, NIL -}; -static int m_norm_no_rot[16] = { - VAR, NIL, NIL, NIL, - NIL, VAR, NIL, NIL, - NIL, NIL, VAR, NIL, - NIL, NIL, NIL, NIL -}; -static int *norm_templates[8] = { - m_norm_no_rot, - m_norm_no_rot, - m_norm_no_rot, - m_norm_general, - m_norm_general, - m_norm_general, - m_norm_identity, - m_norm_identity -}; -static int norm_types[8] = { - NORM_TRANSFORM_NO_ROT, - NORM_TRANSFORM_NO_ROT | NORM_RESCALE, - NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE, - NORM_TRANSFORM, - NORM_TRANSFORM | NORM_RESCALE, - NORM_TRANSFORM | NORM_NORMALIZE, - NORM_RESCALE, - NORM_NORMALIZE -}; -static int norm_scale_types[8] = { /* rescale factor */ - NIL, /* NIL disables rescaling */ - VAR, - NIL, - NIL, - VAR, - NIL, - VAR, - NIL -}; -static int norm_normalize_types[8] = { /* normalizing ?? (no = 0) */ - 0, - 0, - 1, - 0, - 0, - 1, - 0, - 1 -}; -static char *norm_strings[8] = { - "NORM_TRANSFORM_NO_ROT", - "NORM_TRANSFORM_NO_ROT | NORM_RESCALE", - "NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE", - "NORM_TRANSFORM", - "NORM_TRANSFORM | NORM_RESCALE", - "NORM_TRANSFORM | NORM_NORMALIZE", - "NORM_RESCALE", - "NORM_NORMALIZE" -}; - - -/* ============================================================= - * Reference transformations - */ - -static void ref_norm_transform_rescale( const GLmatrix *mat, - GLfloat scale, - const GLvector4f *in, - const GLfloat *lengths, - GLvector4f *dest ) -{ - GLuint i; - const GLfloat *s = in->start; - const GLfloat *m = mat->inv; - GLfloat (*out)[4] = (GLfloat (*)[4]) dest->start; - - (void) lengths; - - for ( i = 0 ; i < in->count ; i++ ) { - GLfloat t[3]; - - TRANSFORM_NORMAL( t, s, m ); - SCALE_SCALAR_3V( out[i], scale, t ); - - s = (GLfloat *)((char *)s + in->stride); - } -} - -static void ref_norm_transform_normalize( const GLmatrix *mat, - GLfloat scale, - const GLvector4f *in, - const GLfloat *lengths, - GLvector4f *dest ) -{ - GLuint i; - const GLfloat *s = in->start; - const GLfloat *m = mat->inv; - GLfloat (*out)[4] = (GLfloat (*)[4]) dest->start; - - for ( i = 0 ; i < in->count ; i++ ) { - GLfloat t[3]; - - TRANSFORM_NORMAL( t, s, m ); - - if ( !lengths ) { - GLfloat len = LEN_SQUARED_3FV( t ); - if ( len > 1e-20 ) { - /* Hmmm, don't know how we could test the precalculated - * length case... - */ - scale = 1.0 / SQRTF( len ); - SCALE_SCALAR_3V( out[i], scale, t ); - } else { - out[i][0] = out[i][1] = out[i][2] = 0; - } - } else { - scale = lengths[i];; - SCALE_SCALAR_3V( out[i], scale, t ); - } - - s = (GLfloat *)((char *)s + in->stride); - } -} - - -/* ============================================================= - * Normal transformation tests - */ - -static void init_matrix( GLfloat *m ) -{ - m[0] = 63.0; m[4] = 43.0; m[ 8] = 29.0; m[12] = 43.0; - m[1] = 55.0; m[5] = 17.0; m[ 9] = 31.0; m[13] = 7.0; - m[2] = 44.0; m[6] = 9.0; m[10] = 7.0; m[14] = 3.0; - m[3] = 11.0; m[7] = 23.0; m[11] = 91.0; m[15] = 9.0; -} - - -static int test_norm_function( normal_func func, int mtype, long *cycles ) -{ - GLvector4f source[1], dest[1], dest2[1], ref[1], ref2[1]; - GLmatrix mat[1]; - GLfloat s[TEST_COUNT][5], d[TEST_COUNT][4], r[TEST_COUNT][4]; - GLfloat d2[TEST_COUNT][4], r2[TEST_COUNT][4], length[TEST_COUNT]; - GLfloat scale; - GLfloat *m; - int i, j; -#ifdef RUN_DEBUG_BENCHMARK - int cycle_i; /* the counter for the benchmarks we run */ -#endif - - (void) cycles; - - mat->m = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); - mat->inv = m = mat->m; - - init_matrix( m ); - - scale = 1.0F + rnd () * norm_scale_types[mtype]; - - for ( i = 0 ; i < 4 ; i++ ) { - for ( j = 0 ; j < 4 ; j++ ) { - switch ( norm_templates[mtype][i * 4 + j] ) { - case NIL: - m[j * 4 + i] = 0.0; - break; - case ONE: - m[j * 4 + i] = 1.0; - break; - case NEG: - m[j * 4 + i] = -1.0; - break; - case VAR: - break; - default: - exit(1); - } - } - } - - for ( i = 0 ; i < TEST_COUNT ; i++ ) { - ASSIGN_3V( d[i], 0.0, 0.0, 0.0 ); - ASSIGN_3V( s[i], 0.0, 0.0, 0.0 ); - ASSIGN_3V( d2[i], 0.0, 0.0, 0.0 ); - for ( j = 0 ; j < 3 ; j++ ) - s[i][j] = rnd(); - length[i] = 1 / SQRTF( LEN_SQUARED_3FV( s[i] ) ); - } - - source->data = (GLfloat(*)[4]) s; - source->start = (GLfloat *) s; - source->count = TEST_COUNT; - source->stride = sizeof(s[0]); - source->flags = 0; - - dest->data = d; - dest->start = (GLfloat *) d; - dest->count = TEST_COUNT; - dest->stride = sizeof(float[4]); - dest->flags = 0; - - dest2->data = d2; - dest2->start = (GLfloat *) d2; - dest2->count = TEST_COUNT; - dest2->stride = sizeof(float[4]); - dest2->flags = 0; - - ref->data = r; - ref->start = (GLfloat *) r; - ref->count = TEST_COUNT; - ref->stride = sizeof(float[4]); - ref->flags = 0; - - ref2->data = r2; - ref2->start = (GLfloat *) r2; - ref2->count = TEST_COUNT; - ref2->stride = sizeof(float[4]); - ref2->flags = 0; - - if ( norm_normalize_types[mtype] == 0 ) { - ref_norm_transform_rescale( mat, scale, source, NULL, ref ); - } else { - ref_norm_transform_normalize( mat, scale, source, NULL, ref ); - ref_norm_transform_normalize( mat, scale, source, length, ref2 ); - } - - if ( mesa_profile ) { - BEGIN_RACE( *cycles ); - func( mat, scale, source, NULL, dest ); - END_RACE( *cycles ); - func( mat, scale, source, length, dest2 ); - } else { - func( mat, scale, source, NULL, dest ); - func( mat, scale, source, length, dest2 ); - } - - for ( i = 0 ; i < TEST_COUNT ; i++ ) { - for ( j = 0 ; j < 3 ; j++ ) { - if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) { - printf( "-----------------------------\n" ); - printf( "(i = %i, j = %i)\n", i, j ); - printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", - d[i][0], r[i][0], r[i][0]/d[i][0], - MAX_PRECISION - significand_match( d[i][0], r[i][0] ) ); - printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", - d[i][1], r[i][1], r[i][1]/d[i][1], - MAX_PRECISION - significand_match( d[i][1], r[i][1] ) ); - printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", - d[i][2], r[i][2], r[i][2]/d[i][2], - MAX_PRECISION - significand_match( d[i][2], r[i][2] ) ); - return 0; - } - - if ( norm_normalize_types[mtype] != 0 ) { - if ( significand_match( d2[i][j], r2[i][j] ) < REQUIRED_PRECISION ) { - printf( "------------------- precalculated length case ------\n" ); - printf( "(i = %i, j = %i)\n", i, j ); - printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", - d2[i][0], r2[i][0], r2[i][0]/d2[i][0], - MAX_PRECISION - significand_match( d2[i][0], r2[i][0] ) ); - printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", - d2[i][1], r2[i][1], r2[i][1]/d2[i][1], - MAX_PRECISION - significand_match( d2[i][1], r2[i][1] ) ); - printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", - d2[i][2], r2[i][2], r2[i][2]/d2[i][2], - MAX_PRECISION - significand_match( d2[i][2], r2[i][2] ) ); - return 0; - } - } - } - } - - _mesa_align_free( mat->m ); - return 1; -} - -void _math_test_all_normal_transform_functions( char *description ) -{ - int mtype; - long benchmark_tab[0xf]; - static int first_time = 1; - - if ( first_time ) { - first_time = 0; - mesa_profile = _mesa_getenv( "MESA_PROFILE" ); - } - -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) { - if ( !counter_overhead ) { - INIT_COUNTER(); - printf( "counter overhead: %ld cycles\n\n", counter_overhead ); - } - printf( "normal transform results after hooking in %s functions:\n", - description ); - printf( "\n-------------------------------------------------------\n" ); - } -#endif - - for ( mtype = 0 ; mtype < 8 ; mtype++ ) { - normal_func func = _mesa_normal_tab[norm_types[mtype]]; - long *cycles = &benchmark_tab[mtype]; - - if ( test_norm_function( func, mtype, cycles ) == 0 ) { - char buf[100]; - sprintf( buf, "_mesa_normal_tab[0][%s] failed test (%s)", - norm_strings[mtype], description ); - _mesa_problem( NULL, "%s", buf ); - } - -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) { - printf( " %li\t", benchmark_tab[mtype] ); - printf( " | [%s]\n", norm_strings[mtype] ); - } -#endif - } -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) { - printf( "\n" ); - } -#endif -} - - -#endif /* DEBUG_MATH */ + +/* + * Mesa 3-D graphics library + * Version: 5.1 + * + * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Gareth Hughes + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "main/macros.h" +#include "main/imports.h" + +#include "m_matrix.h" +#include "m_xform.h" + +#include "m_debug.h" +#include "m_debug_util.h" + + +#ifdef __UNIXOS2__ +/* The linker doesn't like empty files */ +static char dummy; +#endif + +#ifdef DEBUG_MATH /* This code only used for debugging */ + + +static int m_norm_identity[16] = { + ONE, NIL, NIL, NIL, + NIL, ONE, NIL, NIL, + NIL, NIL, ONE, NIL, + NIL, NIL, NIL, NIL +}; +static int m_norm_general[16] = { + VAR, VAR, VAR, NIL, + VAR, VAR, VAR, NIL, + VAR, VAR, VAR, NIL, + NIL, NIL, NIL, NIL +}; +static int m_norm_no_rot[16] = { + VAR, NIL, NIL, NIL, + NIL, VAR, NIL, NIL, + NIL, NIL, VAR, NIL, + NIL, NIL, NIL, NIL +}; +static int *norm_templates[8] = { + m_norm_no_rot, + m_norm_no_rot, + m_norm_no_rot, + m_norm_general, + m_norm_general, + m_norm_general, + m_norm_identity, + m_norm_identity +}; +static int norm_types[8] = { + NORM_TRANSFORM_NO_ROT, + NORM_TRANSFORM_NO_ROT | NORM_RESCALE, + NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE, + NORM_TRANSFORM, + NORM_TRANSFORM | NORM_RESCALE, + NORM_TRANSFORM | NORM_NORMALIZE, + NORM_RESCALE, + NORM_NORMALIZE +}; +static int norm_scale_types[8] = { /* rescale factor */ + NIL, /* NIL disables rescaling */ + VAR, + NIL, + NIL, + VAR, + NIL, + VAR, + NIL +}; +static int norm_normalize_types[8] = { /* normalizing ?? (no = 0) */ + 0, + 0, + 1, + 0, + 0, + 1, + 0, + 1 +}; +static char *norm_strings[8] = { + "NORM_TRANSFORM_NO_ROT", + "NORM_TRANSFORM_NO_ROT | NORM_RESCALE", + "NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE", + "NORM_TRANSFORM", + "NORM_TRANSFORM | NORM_RESCALE", + "NORM_TRANSFORM | NORM_NORMALIZE", + "NORM_RESCALE", + "NORM_NORMALIZE" +}; + + +/* ============================================================= + * Reference transformations + */ + +static void ref_norm_transform_rescale( const GLmatrix *mat, + GLfloat scale, + const GLvector4f *in, + const GLfloat *lengths, + GLvector4f *dest ) +{ + GLuint i; + const GLfloat *s = in->start; + const GLfloat *m = mat->inv; + GLfloat (*out)[4] = (GLfloat (*)[4]) dest->start; + + (void) lengths; + + for ( i = 0 ; i < in->count ; i++ ) { + GLfloat t[3]; + + TRANSFORM_NORMAL( t, s, m ); + SCALE_SCALAR_3V( out[i], scale, t ); + + s = (GLfloat *)((char *)s + in->stride); + } +} + +static void ref_norm_transform_normalize( const GLmatrix *mat, + GLfloat scale, + const GLvector4f *in, + const GLfloat *lengths, + GLvector4f *dest ) +{ + GLuint i; + const GLfloat *s = in->start; + const GLfloat *m = mat->inv; + GLfloat (*out)[4] = (GLfloat (*)[4]) dest->start; + + for ( i = 0 ; i < in->count ; i++ ) { + GLfloat t[3]; + + TRANSFORM_NORMAL( t, s, m ); + + if ( !lengths ) { + GLfloat len = LEN_SQUARED_3FV( t ); + if ( len > 1e-20 ) { + /* Hmmm, don't know how we could test the precalculated + * length case... + */ + scale = 1.0 / SQRTF( len ); + SCALE_SCALAR_3V( out[i], scale, t ); + } else { + out[i][0] = out[i][1] = out[i][2] = 0; + } + } else { + scale = lengths[i];; + SCALE_SCALAR_3V( out[i], scale, t ); + } + + s = (GLfloat *)((char *)s + in->stride); + } +} + + +/* ============================================================= + * Normal transformation tests + */ + +static void init_matrix( GLfloat *m ) +{ + m[0] = 63.0; m[4] = 43.0; m[ 8] = 29.0; m[12] = 43.0; + m[1] = 55.0; m[5] = 17.0; m[ 9] = 31.0; m[13] = 7.0; + m[2] = 44.0; m[6] = 9.0; m[10] = 7.0; m[14] = 3.0; + m[3] = 11.0; m[7] = 23.0; m[11] = 91.0; m[15] = 9.0; +} + + +static int test_norm_function( normal_func func, int mtype, long *cycles ) +{ + GLvector4f source[1], dest[1], dest2[1], ref[1], ref2[1]; + GLmatrix mat[1]; + GLfloat s[TEST_COUNT][5], d[TEST_COUNT][4], r[TEST_COUNT][4]; + GLfloat d2[TEST_COUNT][4], r2[TEST_COUNT][4], length[TEST_COUNT]; + GLfloat scale; + GLfloat *m; + int i, j; +#ifdef RUN_DEBUG_BENCHMARK + int cycle_i; /* the counter for the benchmarks we run */ +#endif + + (void) cycles; + + mat->m = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); + mat->inv = m = mat->m; + + init_matrix( m ); + + scale = 1.0F + rnd () * norm_scale_types[mtype]; + + for ( i = 0 ; i < 4 ; i++ ) { + for ( j = 0 ; j < 4 ; j++ ) { + switch ( norm_templates[mtype][i * 4 + j] ) { + case NIL: + m[j * 4 + i] = 0.0; + break; + case ONE: + m[j * 4 + i] = 1.0; + break; + case NEG: + m[j * 4 + i] = -1.0; + break; + case VAR: + break; + default: + exit(1); + } + } + } + + for ( i = 0 ; i < TEST_COUNT ; i++ ) { + ASSIGN_3V( d[i], 0.0, 0.0, 0.0 ); + ASSIGN_3V( s[i], 0.0, 0.0, 0.0 ); + ASSIGN_3V( d2[i], 0.0, 0.0, 0.0 ); + for ( j = 0 ; j < 3 ; j++ ) + s[i][j] = rnd(); + length[i] = 1 / SQRTF( LEN_SQUARED_3FV( s[i] ) ); + } + + source->data = (GLfloat(*)[4]) s; + source->start = (GLfloat *) s; + source->count = TEST_COUNT; + source->stride = sizeof(s[0]); + source->flags = 0; + + dest->data = d; + dest->start = (GLfloat *) d; + dest->count = TEST_COUNT; + dest->stride = sizeof(float[4]); + dest->flags = 0; + + dest2->data = d2; + dest2->start = (GLfloat *) d2; + dest2->count = TEST_COUNT; + dest2->stride = sizeof(float[4]); + dest2->flags = 0; + + ref->data = r; + ref->start = (GLfloat *) r; + ref->count = TEST_COUNT; + ref->stride = sizeof(float[4]); + ref->flags = 0; + + ref2->data = r2; + ref2->start = (GLfloat *) r2; + ref2->count = TEST_COUNT; + ref2->stride = sizeof(float[4]); + ref2->flags = 0; + + if ( norm_normalize_types[mtype] == 0 ) { + ref_norm_transform_rescale( mat, scale, source, NULL, ref ); + } else { + ref_norm_transform_normalize( mat, scale, source, NULL, ref ); + ref_norm_transform_normalize( mat, scale, source, length, ref2 ); + } + + if ( mesa_profile ) { + BEGIN_RACE( *cycles ); + func( mat, scale, source, NULL, dest ); + END_RACE( *cycles ); + func( mat, scale, source, length, dest2 ); + } else { + func( mat, scale, source, NULL, dest ); + func( mat, scale, source, length, dest2 ); + } + + for ( i = 0 ; i < TEST_COUNT ; i++ ) { + for ( j = 0 ; j < 3 ; j++ ) { + if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) { + printf( "-----------------------------\n" ); + printf( "(i = %i, j = %i)\n", i, j ); + printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", + d[i][0], r[i][0], r[i][0]/d[i][0], + MAX_PRECISION - significand_match( d[i][0], r[i][0] ) ); + printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", + d[i][1], r[i][1], r[i][1]/d[i][1], + MAX_PRECISION - significand_match( d[i][1], r[i][1] ) ); + printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", + d[i][2], r[i][2], r[i][2]/d[i][2], + MAX_PRECISION - significand_match( d[i][2], r[i][2] ) ); + return 0; + } + + if ( norm_normalize_types[mtype] != 0 ) { + if ( significand_match( d2[i][j], r2[i][j] ) < REQUIRED_PRECISION ) { + printf( "------------------- precalculated length case ------\n" ); + printf( "(i = %i, j = %i)\n", i, j ); + printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", + d2[i][0], r2[i][0], r2[i][0]/d2[i][0], + MAX_PRECISION - significand_match( d2[i][0], r2[i][0] ) ); + printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", + d2[i][1], r2[i][1], r2[i][1]/d2[i][1], + MAX_PRECISION - significand_match( d2[i][1], r2[i][1] ) ); + printf( "%f \t %f \t [ratio = %e - %i bit missed]\n", + d2[i][2], r2[i][2], r2[i][2]/d2[i][2], + MAX_PRECISION - significand_match( d2[i][2], r2[i][2] ) ); + return 0; + } + } + } + } + + _mesa_align_free( mat->m ); + return 1; +} + +void _math_test_all_normal_transform_functions( char *description ) +{ + int mtype; + long benchmark_tab[0xf]; + static int first_time = 1; + + if ( first_time ) { + first_time = 0; + mesa_profile = _mesa_getenv( "MESA_PROFILE" ); + } + +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) { + if ( !counter_overhead ) { + INIT_COUNTER(); + printf( "counter overhead: %ld cycles\n\n", counter_overhead ); + } + printf( "normal transform results after hooking in %s functions:\n", + description ); + printf( "\n-------------------------------------------------------\n" ); + } +#endif + + for ( mtype = 0 ; mtype < 8 ; mtype++ ) { + normal_func func = _mesa_normal_tab[norm_types[mtype]]; + long *cycles = &benchmark_tab[mtype]; + + if ( test_norm_function( func, mtype, cycles ) == 0 ) { + char buf[100]; + sprintf( buf, "_mesa_normal_tab[0][%s] failed test (%s)", + norm_strings[mtype], description ); + _mesa_problem( NULL, "%s", buf ); + } + +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) { + printf( " %li\t", benchmark_tab[mtype] ); + printf( " | [%s]\n", norm_strings[mtype] ); + } +#endif + } +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) { + printf( "\n" ); + } +#endif +} + + +#endif /* DEBUG_MATH */ diff --git a/mesalib/src/mesa/math/m_debug_xform.c b/mesalib/src/mesa/math/m_debug_xform.c index 0de43195c..7d815664a 100644 --- a/mesalib/src/mesa/math/m_debug_xform.c +++ b/mesalib/src/mesa/math/m_debug_xform.c @@ -1,339 +1,339 @@ -/* - * Mesa 3-D graphics library - * Version: 6.1 - * - * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * Updated for P6 architecture by Gareth Hughes. - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "main/macros.h" -#include "main/imports.h" - -#include "m_matrix.h" -#include "m_xform.h" - -#include "m_debug.h" -#include "m_debug_util.h" - -#ifdef __UNIXOS2__ -/* The linker doesn't like empty files */ -static char dummy; -#endif - -#ifdef DEBUG_MATH /* This code only used for debugging */ - - -/* Overhead of profiling counter in cycles. Automatically adjusted to - * your machine at run time - counter initialization should give very - * consistent results. - */ -long counter_overhead = 0; - -/* This is the value of the environment variable MESA_PROFILE, and is - * used to determine if we should benchmark the functions as well as - * verify their correctness. - */ -char *mesa_profile = NULL; - - -static int m_general[16] = { - VAR, VAR, VAR, VAR, - VAR, VAR, VAR, VAR, - VAR, VAR, VAR, VAR, - VAR, VAR, VAR, VAR -}; -static int m_identity[16] = { - ONE, NIL, NIL, NIL, - NIL, ONE, NIL, NIL, - NIL, NIL, ONE, NIL, - NIL, NIL, NIL, ONE -}; -static int m_2d[16] = { - VAR, VAR, NIL, VAR, - VAR, VAR, NIL, VAR, - NIL, NIL, ONE, NIL, - NIL, NIL, NIL, ONE -}; -static int m_2d_no_rot[16] = { - VAR, NIL, NIL, VAR, - NIL, VAR, NIL, VAR, - NIL, NIL, ONE, NIL, - NIL, NIL, NIL, ONE -}; -static int m_3d[16] = { - VAR, VAR, VAR, VAR, - VAR, VAR, VAR, VAR, - VAR, VAR, VAR, VAR, - NIL, NIL, NIL, ONE -}; -static int m_3d_no_rot[16] = { - VAR, NIL, NIL, VAR, - NIL, VAR, NIL, VAR, - NIL, NIL, VAR, VAR, - NIL, NIL, NIL, ONE -}; -static int m_perspective[16] = { - VAR, NIL, VAR, NIL, - NIL, VAR, VAR, NIL, - NIL, NIL, VAR, VAR, - NIL, NIL, NEG, NIL -}; -static int *templates[7] = { - m_general, - m_identity, - m_3d_no_rot, - m_perspective, - m_2d, - m_2d_no_rot, - m_3d -}; -static enum GLmatrixtype mtypes[7] = { - MATRIX_GENERAL, - MATRIX_IDENTITY, - MATRIX_3D_NO_ROT, - MATRIX_PERSPECTIVE, - MATRIX_2D, - MATRIX_2D_NO_ROT, - MATRIX_3D -}; -static char *mstrings[7] = { - "MATRIX_GENERAL", - "MATRIX_IDENTITY", - "MATRIX_3D_NO_ROT", - "MATRIX_PERSPECTIVE", - "MATRIX_2D", - "MATRIX_2D_NO_ROT", - "MATRIX_3D" -}; - - -/* ============================================================= - * Reference transformations - */ - -static void ref_transform( GLvector4f *dst, - const GLmatrix *mat, - const GLvector4f *src ) -{ - GLuint i; - GLfloat *s = (GLfloat *)src->start; - GLfloat (*d)[4] = (GLfloat (*)[4])dst->start; - const GLfloat *m = mat->m; - - for ( i = 0 ; i < src->count ; i++ ) { - TRANSFORM_POINT( d[i], m, s ); - s = (GLfloat *)((char *)s + src->stride); - } -} - - -/* ============================================================= - * Vertex transformation tests - */ - -static void init_matrix( GLfloat *m ) -{ - m[0] = 63.0; m[4] = 43.0; m[ 8] = 29.0; m[12] = 43.0; - m[1] = 55.0; m[5] = 17.0; m[ 9] = 31.0; m[13] = 7.0; - m[2] = 44.0; m[6] = 9.0; m[10] = 7.0; m[14] = 3.0; - m[3] = 11.0; m[7] = 23.0; m[11] = 91.0; m[15] = 9.0; -} - -ALIGN16(static GLfloat, s[TEST_COUNT][4]); -ALIGN16(static GLfloat, d[TEST_COUNT][4]); -ALIGN16(static GLfloat, r[TEST_COUNT][4]); - -static int test_transform_function( transform_func func, int psize, - int mtype, unsigned long *cycles ) -{ - GLvector4f source[1], dest[1], ref[1]; - GLmatrix mat[1]; - GLfloat *m; - int i, j; -#ifdef RUN_DEBUG_BENCHMARK - int cycle_i; /* the counter for the benchmarks we run */ -#endif - - (void) cycles; - - if ( psize > 4 ) { - _mesa_problem( NULL, "test_transform_function called with psize > 4\n" ); - return 0; - } - - mat->m = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); - mat->type = mtypes[mtype]; - - m = mat->m; - ASSERT( ((long)m & 15) == 0 ); - - init_matrix( m ); - - for ( i = 0 ; i < 4 ; i++ ) { - for ( j = 0 ; j < 4 ; j++ ) { - switch ( templates[mtype][i * 4 + j] ) { - case NIL: - m[j * 4 + i] = 0.0; - break; - case ONE: - m[j * 4 + i] = 1.0; - break; - case NEG: - m[j * 4 + i] = -1.0; - break; - case VAR: - break; - default: - ASSERT(0); - return 0; - } - } - } - - for ( i = 0 ; i < TEST_COUNT ; i++) { - ASSIGN_4V( d[i], 0.0, 0.0, 0.0, 1.0 ); - ASSIGN_4V( s[i], 0.0, 0.0, 0.0, 1.0 ); - for ( j = 0 ; j < psize ; j++ ) - s[i][j] = rnd(); - } - - source->data = (GLfloat(*)[4])s; - source->start = (GLfloat *)s; - source->count = TEST_COUNT; - source->stride = sizeof(s[0]); - source->size = 4; - source->flags = 0; - - dest->data = (GLfloat(*)[4])d; - dest->start = (GLfloat *)d; - dest->count = TEST_COUNT; - dest->stride = sizeof(float[4]); - dest->size = 0; - dest->flags = 0; - - ref->data = (GLfloat(*)[4])r; - ref->start = (GLfloat *)r; - ref->count = TEST_COUNT; - ref->stride = sizeof(float[4]); - ref->size = 0; - ref->flags = 0; - - ref_transform( ref, mat, source ); - - if ( mesa_profile ) { - BEGIN_RACE( *cycles ); - func( dest, mat->m, source ); - END_RACE( *cycles ); - } - else { - func( dest, mat->m, source ); - } - - for ( i = 0 ; i < TEST_COUNT ; i++ ) { - for ( j = 0 ; j < 4 ; j++ ) { - if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) { - printf("-----------------------------\n" ); - printf("(i = %i, j = %i)\n", i, j ); - printf("%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][0], r[i][0], r[i][0]-d[i][0], - MAX_PRECISION - significand_match( d[i][0], r[i][0] ) ); - printf("%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][1], r[i][1], r[i][1]-d[i][1], - MAX_PRECISION - significand_match( d[i][1], r[i][1] ) ); - printf("%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][2], r[i][2], r[i][2]-d[i][2], - MAX_PRECISION - significand_match( d[i][2], r[i][2] ) ); - printf("%f \t %f \t [diff = %e - %i bit missed]\n", - d[i][3], r[i][3], r[i][3]-d[i][3], - MAX_PRECISION - significand_match( d[i][3], r[i][3] ) ); - return 0; - } - } - } - - _mesa_align_free( mat->m ); - return 1; -} - -void _math_test_all_transform_functions( char *description ) -{ - int psize, mtype; - unsigned long benchmark_tab[4][7]; - static int first_time = 1; - - if ( first_time ) { - first_time = 0; - mesa_profile = _mesa_getenv( "MESA_PROFILE" ); - } - -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) { - if ( !counter_overhead ) { - INIT_COUNTER(); - printf("counter overhead: %lu cycles\n\n", counter_overhead ); - } - printf("transform results after hooking in %s functions:\n", description ); - } -#endif - -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) { - printf("\n" ); - for ( psize = 1 ; psize <= 4 ; psize++ ) { - printf(" p%d\t", psize ); - } - printf("\n--------------------------------------------------------\n" ); - } -#endif - - for ( mtype = 0 ; mtype < 7 ; mtype++ ) { - for ( psize = 1 ; psize <= 4 ; psize++ ) { - transform_func func = _mesa_transform_tab[psize][mtypes[mtype]]; - unsigned long *cycles = &(benchmark_tab[psize-1][mtype]); - - if ( test_transform_function( func, psize, mtype, cycles ) == 0 ) { - char buf[100]; - sprintf(buf, "_mesa_transform_tab[0][%d][%s] failed test (%s)", - psize, mstrings[mtype], description ); - _mesa_problem( NULL, "%s", buf ); - } -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) - printf(" %li\t", benchmark_tab[psize-1][mtype] ); -#endif - } -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) - printf(" | [%s]\n", mstrings[mtype] ); -#endif - } -#ifdef RUN_DEBUG_BENCHMARK - if ( mesa_profile ) - printf( "\n" ); -#endif -} - - -#endif /* DEBUG_MATH */ +/* + * Mesa 3-D graphics library + * Version: 6.1 + * + * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * Updated for P6 architecture by Gareth Hughes. + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "main/macros.h" +#include "main/imports.h" + +#include "m_matrix.h" +#include "m_xform.h" + +#include "m_debug.h" +#include "m_debug_util.h" + +#ifdef __UNIXOS2__ +/* The linker doesn't like empty files */ +static char dummy; +#endif + +#ifdef DEBUG_MATH /* This code only used for debugging */ + + +/* Overhead of profiling counter in cycles. Automatically adjusted to + * your machine at run time - counter initialization should give very + * consistent results. + */ +long counter_overhead = 0; + +/* This is the value of the environment variable MESA_PROFILE, and is + * used to determine if we should benchmark the functions as well as + * verify their correctness. + */ +char *mesa_profile = NULL; + + +static int m_general[16] = { + VAR, VAR, VAR, VAR, + VAR, VAR, VAR, VAR, + VAR, VAR, VAR, VAR, + VAR, VAR, VAR, VAR +}; +static int m_identity[16] = { + ONE, NIL, NIL, NIL, + NIL, ONE, NIL, NIL, + NIL, NIL, ONE, NIL, + NIL, NIL, NIL, ONE +}; +static int m_2d[16] = { + VAR, VAR, NIL, VAR, + VAR, VAR, NIL, VAR, + NIL, NIL, ONE, NIL, + NIL, NIL, NIL, ONE +}; +static int m_2d_no_rot[16] = { + VAR, NIL, NIL, VAR, + NIL, VAR, NIL, VAR, + NIL, NIL, ONE, NIL, + NIL, NIL, NIL, ONE +}; +static int m_3d[16] = { + VAR, VAR, VAR, VAR, + VAR, VAR, VAR, VAR, + VAR, VAR, VAR, VAR, + NIL, NIL, NIL, ONE +}; +static int m_3d_no_rot[16] = { + VAR, NIL, NIL, VAR, + NIL, VAR, NIL, VAR, + NIL, NIL, VAR, VAR, + NIL, NIL, NIL, ONE +}; +static int m_perspective[16] = { + VAR, NIL, VAR, NIL, + NIL, VAR, VAR, NIL, + NIL, NIL, VAR, VAR, + NIL, NIL, NEG, NIL +}; +static int *templates[7] = { + m_general, + m_identity, + m_3d_no_rot, + m_perspective, + m_2d, + m_2d_no_rot, + m_3d +}; +static enum GLmatrixtype mtypes[7] = { + MATRIX_GENERAL, + MATRIX_IDENTITY, + MATRIX_3D_NO_ROT, + MATRIX_PERSPECTIVE, + MATRIX_2D, + MATRIX_2D_NO_ROT, + MATRIX_3D +}; +static char *mstrings[7] = { + "MATRIX_GENERAL", + "MATRIX_IDENTITY", + "MATRIX_3D_NO_ROT", + "MATRIX_PERSPECTIVE", + "MATRIX_2D", + "MATRIX_2D_NO_ROT", + "MATRIX_3D" +}; + + +/* ============================================================= + * Reference transformations + */ + +static void ref_transform( GLvector4f *dst, + const GLmatrix *mat, + const GLvector4f *src ) +{ + GLuint i; + GLfloat *s = (GLfloat *)src->start; + GLfloat (*d)[4] = (GLfloat (*)[4])dst->start; + const GLfloat *m = mat->m; + + for ( i = 0 ; i < src->count ; i++ ) { + TRANSFORM_POINT( d[i], m, s ); + s = (GLfloat *)((char *)s + src->stride); + } +} + + +/* ============================================================= + * Vertex transformation tests + */ + +static void init_matrix( GLfloat *m ) +{ + m[0] = 63.0; m[4] = 43.0; m[ 8] = 29.0; m[12] = 43.0; + m[1] = 55.0; m[5] = 17.0; m[ 9] = 31.0; m[13] = 7.0; + m[2] = 44.0; m[6] = 9.0; m[10] = 7.0; m[14] = 3.0; + m[3] = 11.0; m[7] = 23.0; m[11] = 91.0; m[15] = 9.0; +} + +ALIGN16(static GLfloat, s[TEST_COUNT][4]); +ALIGN16(static GLfloat, d[TEST_COUNT][4]); +ALIGN16(static GLfloat, r[TEST_COUNT][4]); + +static int test_transform_function( transform_func func, int psize, + int mtype, unsigned long *cycles ) +{ + GLvector4f source[1], dest[1], ref[1]; + GLmatrix mat[1]; + GLfloat *m; + int i, j; +#ifdef RUN_DEBUG_BENCHMARK + int cycle_i; /* the counter for the benchmarks we run */ +#endif + + (void) cycles; + + if ( psize > 4 ) { + _mesa_problem( NULL, "test_transform_function called with psize > 4\n" ); + return 0; + } + + mat->m = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); + mat->type = mtypes[mtype]; + + m = mat->m; + ASSERT( ((long)m & 15) == 0 ); + + init_matrix( m ); + + for ( i = 0 ; i < 4 ; i++ ) { + for ( j = 0 ; j < 4 ; j++ ) { + switch ( templates[mtype][i * 4 + j] ) { + case NIL: + m[j * 4 + i] = 0.0; + break; + case ONE: + m[j * 4 + i] = 1.0; + break; + case NEG: + m[j * 4 + i] = -1.0; + break; + case VAR: + break; + default: + ASSERT(0); + return 0; + } + } + } + + for ( i = 0 ; i < TEST_COUNT ; i++) { + ASSIGN_4V( d[i], 0.0, 0.0, 0.0, 1.0 ); + ASSIGN_4V( s[i], 0.0, 0.0, 0.0, 1.0 ); + for ( j = 0 ; j < psize ; j++ ) + s[i][j] = rnd(); + } + + source->data = (GLfloat(*)[4])s; + source->start = (GLfloat *)s; + source->count = TEST_COUNT; + source->stride = sizeof(s[0]); + source->size = 4; + source->flags = 0; + + dest->data = (GLfloat(*)[4])d; + dest->start = (GLfloat *)d; + dest->count = TEST_COUNT; + dest->stride = sizeof(float[4]); + dest->size = 0; + dest->flags = 0; + + ref->data = (GLfloat(*)[4])r; + ref->start = (GLfloat *)r; + ref->count = TEST_COUNT; + ref->stride = sizeof(float[4]); + ref->size = 0; + ref->flags = 0; + + ref_transform( ref, mat, source ); + + if ( mesa_profile ) { + BEGIN_RACE( *cycles ); + func( dest, mat->m, source ); + END_RACE( *cycles ); + } + else { + func( dest, mat->m, source ); + } + + for ( i = 0 ; i < TEST_COUNT ; i++ ) { + for ( j = 0 ; j < 4 ; j++ ) { + if ( significand_match( d[i][j], r[i][j] ) < REQUIRED_PRECISION ) { + printf("-----------------------------\n" ); + printf("(i = %i, j = %i)\n", i, j ); + printf("%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][0], r[i][0], r[i][0]-d[i][0], + MAX_PRECISION - significand_match( d[i][0], r[i][0] ) ); + printf("%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][1], r[i][1], r[i][1]-d[i][1], + MAX_PRECISION - significand_match( d[i][1], r[i][1] ) ); + printf("%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][2], r[i][2], r[i][2]-d[i][2], + MAX_PRECISION - significand_match( d[i][2], r[i][2] ) ); + printf("%f \t %f \t [diff = %e - %i bit missed]\n", + d[i][3], r[i][3], r[i][3]-d[i][3], + MAX_PRECISION - significand_match( d[i][3], r[i][3] ) ); + return 0; + } + } + } + + _mesa_align_free( mat->m ); + return 1; +} + +void _math_test_all_transform_functions( char *description ) +{ + int psize, mtype; + unsigned long benchmark_tab[4][7]; + static int first_time = 1; + + if ( first_time ) { + first_time = 0; + mesa_profile = _mesa_getenv( "MESA_PROFILE" ); + } + +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) { + if ( !counter_overhead ) { + INIT_COUNTER(); + printf("counter overhead: %lu cycles\n\n", counter_overhead ); + } + printf("transform results after hooking in %s functions:\n", description ); + } +#endif + +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) { + printf("\n" ); + for ( psize = 1 ; psize <= 4 ; psize++ ) { + printf(" p%d\t", psize ); + } + printf("\n--------------------------------------------------------\n" ); + } +#endif + + for ( mtype = 0 ; mtype < 7 ; mtype++ ) { + for ( psize = 1 ; psize <= 4 ; psize++ ) { + transform_func func = _mesa_transform_tab[psize][mtypes[mtype]]; + unsigned long *cycles = &(benchmark_tab[psize-1][mtype]); + + if ( test_transform_function( func, psize, mtype, cycles ) == 0 ) { + char buf[100]; + sprintf(buf, "_mesa_transform_tab[0][%d][%s] failed test (%s)", + psize, mstrings[mtype], description ); + _mesa_problem( NULL, "%s", buf ); + } +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) + printf(" %li\t", benchmark_tab[psize-1][mtype] ); +#endif + } +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) + printf(" | [%s]\n", mstrings[mtype] ); +#endif + } +#ifdef RUN_DEBUG_BENCHMARK + if ( mesa_profile ) + printf( "\n" ); +#endif +} + + +#endif /* DEBUG_MATH */ diff --git a/mesalib/src/mesa/math/m_matrix.c b/mesalib/src/mesa/math/m_matrix.c index 83eb787c7..02aedbad8 100644 --- a/mesalib/src/mesa/math/m_matrix.c +++ b/mesalib/src/mesa/math/m_matrix.c @@ -1,1641 +1,1641 @@ -/* - * Mesa 3-D graphics library - * Version: 6.3 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -/** - * \file m_matrix.c - * Matrix operations. - * - * \note - * -# 4x4 transformation matrices are stored in memory in column major order. - * -# Points/vertices are to be thought of as column vectors. - * -# Transformation of a point p by a matrix M is: p' = M * p - */ - - -#include "main/glheader.h" -#include "main/imports.h" -#include "main/macros.h" - -#include "m_matrix.h" - - -/** - * \defgroup MatFlags MAT_FLAG_XXX-flags - * - * Bitmasks to indicate different kinds of 4x4 matrices in GLmatrix::flags - * It would be nice to make all these flags private to m_matrix.c - */ -/*@{*/ -#define MAT_FLAG_IDENTITY 0 /**< is an identity matrix flag. - * (Not actually used - the identity - * matrix is identified by the absense - * of all other flags.) - */ -#define MAT_FLAG_GENERAL 0x1 /**< is a general matrix flag */ -#define MAT_FLAG_ROTATION 0x2 /**< is a rotation matrix flag */ -#define MAT_FLAG_TRANSLATION 0x4 /**< is a translation matrix flag */ -#define MAT_FLAG_UNIFORM_SCALE 0x8 /**< is an uniform scaling matrix flag */ -#define MAT_FLAG_GENERAL_SCALE 0x10 /**< is a general scaling matrix flag */ -#define MAT_FLAG_GENERAL_3D 0x20 /**< general 3D matrix flag */ -#define MAT_FLAG_PERSPECTIVE 0x40 /**< is a perspective proj matrix flag */ -#define MAT_FLAG_SINGULAR 0x80 /**< is a singular matrix flag */ -#define MAT_DIRTY_TYPE 0x100 /**< matrix type is dirty */ -#define MAT_DIRTY_FLAGS 0x200 /**< matrix flags are dirty */ -#define MAT_DIRTY_INVERSE 0x400 /**< matrix inverse is dirty */ - -/** angle preserving matrix flags mask */ -#define MAT_FLAGS_ANGLE_PRESERVING (MAT_FLAG_ROTATION | \ - MAT_FLAG_TRANSLATION | \ - MAT_FLAG_UNIFORM_SCALE) - -/** geometry related matrix flags mask */ -#define MAT_FLAGS_GEOMETRY (MAT_FLAG_GENERAL | \ - MAT_FLAG_ROTATION | \ - MAT_FLAG_TRANSLATION | \ - MAT_FLAG_UNIFORM_SCALE | \ - MAT_FLAG_GENERAL_SCALE | \ - MAT_FLAG_GENERAL_3D | \ - MAT_FLAG_PERSPECTIVE | \ - MAT_FLAG_SINGULAR) - -/** length preserving matrix flags mask */ -#define MAT_FLAGS_LENGTH_PRESERVING (MAT_FLAG_ROTATION | \ - MAT_FLAG_TRANSLATION) - - -/** 3D (non-perspective) matrix flags mask */ -#define MAT_FLAGS_3D (MAT_FLAG_ROTATION | \ - MAT_FLAG_TRANSLATION | \ - MAT_FLAG_UNIFORM_SCALE | \ - MAT_FLAG_GENERAL_SCALE | \ - MAT_FLAG_GENERAL_3D) - -/** dirty matrix flags mask */ -#define MAT_DIRTY (MAT_DIRTY_TYPE | \ - MAT_DIRTY_FLAGS | \ - MAT_DIRTY_INVERSE) - -/*@}*/ - - -/** - * Test geometry related matrix flags. - * - * \param mat a pointer to a GLmatrix structure. - * \param a flags mask. - * - * \returns non-zero if all geometry related matrix flags are contained within - * the mask, or zero otherwise. - */ -#define TEST_MAT_FLAGS(mat, a) \ - ((MAT_FLAGS_GEOMETRY & (~(a)) & ((mat)->flags) ) == 0) - - - -/** - * Names of the corresponding GLmatrixtype values. - */ -static const char *types[] = { - "MATRIX_GENERAL", - "MATRIX_IDENTITY", - "MATRIX_3D_NO_ROT", - "MATRIX_PERSPECTIVE", - "MATRIX_2D", - "MATRIX_2D_NO_ROT", - "MATRIX_3D" -}; - - -/** - * Identity matrix. - */ -static GLfloat Identity[16] = { - 1.0, 0.0, 0.0, 0.0, - 0.0, 1.0, 0.0, 0.0, - 0.0, 0.0, 1.0, 0.0, - 0.0, 0.0, 0.0, 1.0 -}; - - - -/**********************************************************************/ -/** \name Matrix multiplication */ -/*@{*/ - -#define A(row,col) a[(col<<2)+row] -#define B(row,col) b[(col<<2)+row] -#define P(row,col) product[(col<<2)+row] - -/** - * Perform a full 4x4 matrix multiplication. - * - * \param a matrix. - * \param b matrix. - * \param product will receive the product of \p a and \p b. - * - * \warning Is assumed that \p product != \p b. \p product == \p a is allowed. - * - * \note KW: 4*16 = 64 multiplications - * - * \author This \c matmul was contributed by Thomas Malik - */ -static void matmul4( GLfloat *product, const GLfloat *a, const GLfloat *b ) -{ - GLint i; - for (i = 0; i < 4; i++) { - const GLfloat ai0=A(i,0), ai1=A(i,1), ai2=A(i,2), ai3=A(i,3); - P(i,0) = ai0 * B(0,0) + ai1 * B(1,0) + ai2 * B(2,0) + ai3 * B(3,0); - P(i,1) = ai0 * B(0,1) + ai1 * B(1,1) + ai2 * B(2,1) + ai3 * B(3,1); - P(i,2) = ai0 * B(0,2) + ai1 * B(1,2) + ai2 * B(2,2) + ai3 * B(3,2); - P(i,3) = ai0 * B(0,3) + ai1 * B(1,3) + ai2 * B(2,3) + ai3 * B(3,3); - } -} - -/** - * Multiply two matrices known to occupy only the top three rows, such - * as typical model matrices, and orthogonal matrices. - * - * \param a matrix. - * \param b matrix. - * \param product will receive the product of \p a and \p b. - */ -static void matmul34( GLfloat *product, const GLfloat *a, const GLfloat *b ) -{ - GLint i; - for (i = 0; i < 3; i++) { - const GLfloat ai0=A(i,0), ai1=A(i,1), ai2=A(i,2), ai3=A(i,3); - P(i,0) = ai0 * B(0,0) + ai1 * B(1,0) + ai2 * B(2,0); - P(i,1) = ai0 * B(0,1) + ai1 * B(1,1) + ai2 * B(2,1); - P(i,2) = ai0 * B(0,2) + ai1 * B(1,2) + ai2 * B(2,2); - P(i,3) = ai0 * B(0,3) + ai1 * B(1,3) + ai2 * B(2,3) + ai3; - } - P(3,0) = 0; - P(3,1) = 0; - P(3,2) = 0; - P(3,3) = 1; -} - -#undef A -#undef B -#undef P - -/** - * Multiply a matrix by an array of floats with known properties. - * - * \param mat pointer to a GLmatrix structure containing the left multiplication - * matrix, and that will receive the product result. - * \param m right multiplication matrix array. - * \param flags flags of the matrix \p m. - * - * Joins both flags and marks the type and inverse as dirty. Calls matmul34() - * if both matrices are 3D, or matmul4() otherwise. - */ -static void matrix_multf( GLmatrix *mat, const GLfloat *m, GLuint flags ) -{ - mat->flags |= (flags | MAT_DIRTY_TYPE | MAT_DIRTY_INVERSE); - - if (TEST_MAT_FLAGS(mat, MAT_FLAGS_3D)) - matmul34( mat->m, mat->m, m ); - else - matmul4( mat->m, mat->m, m ); -} - -/** - * Matrix multiplication. - * - * \param dest destination matrix. - * \param a left matrix. - * \param b right matrix. - * - * Joins both flags and marks the type and inverse as dirty. Calls matmul34() - * if both matrices are 3D, or matmul4() otherwise. - */ -void -_math_matrix_mul_matrix( GLmatrix *dest, const GLmatrix *a, const GLmatrix *b ) -{ - dest->flags = (a->flags | - b->flags | - MAT_DIRTY_TYPE | - MAT_DIRTY_INVERSE); - - if (TEST_MAT_FLAGS(dest, MAT_FLAGS_3D)) - matmul34( dest->m, a->m, b->m ); - else - matmul4( dest->m, a->m, b->m ); -} - -/** - * Matrix multiplication. - * - * \param dest left and destination matrix. - * \param m right matrix array. - * - * Marks the matrix flags with general flag, and type and inverse dirty flags. - * Calls matmul4() for the multiplication. - */ -void -_math_matrix_mul_floats( GLmatrix *dest, const GLfloat *m ) -{ - dest->flags |= (MAT_FLAG_GENERAL | - MAT_DIRTY_TYPE | - MAT_DIRTY_INVERSE | - MAT_DIRTY_FLAGS); - - matmul4( dest->m, dest->m, m ); -} - -/*@}*/ - - -/**********************************************************************/ -/** \name Matrix output */ -/*@{*/ - -/** - * Print a matrix array. - * - * \param m matrix array. - * - * Called by _math_matrix_print() to print a matrix or its inverse. - */ -static void print_matrix_floats( const GLfloat m[16] ) -{ - int i; - for (i=0;i<4;i++) { - _mesa_debug(NULL,"\t%f %f %f %f\n", m[i], m[4+i], m[8+i], m[12+i] ); - } -} - -/** - * Dumps the contents of a GLmatrix structure. - * - * \param m pointer to the GLmatrix structure. - */ -void -_math_matrix_print( const GLmatrix *m ) -{ - _mesa_debug(NULL, "Matrix type: %s, flags: %x\n", types[m->type], m->flags); - print_matrix_floats(m->m); - _mesa_debug(NULL, "Inverse: \n"); - if (m->inv) { - GLfloat prod[16]; - print_matrix_floats(m->inv); - matmul4(prod, m->m, m->inv); - _mesa_debug(NULL, "Mat * Inverse:\n"); - print_matrix_floats(prod); - } - else { - _mesa_debug(NULL, " - not available\n"); - } -} - -/*@}*/ - - -/** - * References an element of 4x4 matrix. - * - * \param m matrix array. - * \param c column of the desired element. - * \param r row of the desired element. - * - * \return value of the desired element. - * - * Calculate the linear storage index of the element and references it. - */ -#define MAT(m,r,c) (m)[(c)*4+(r)] - - -/**********************************************************************/ -/** \name Matrix inversion */ -/*@{*/ - -/** - * Swaps the values of two floating pointer variables. - * - * Used by invert_matrix_general() to swap the row pointers. - */ -#define SWAP_ROWS(a, b) { GLfloat *_tmp = a; (a)=(b); (b)=_tmp; } - -/** - * Compute inverse of 4x4 transformation matrix. - * - * \param mat pointer to a GLmatrix structure. The matrix inverse will be - * stored in the GLmatrix::inv attribute. - * - * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). - * - * \author - * Code contributed by Jacques Leroy jle@star.be - * - * Calculates the inverse matrix by performing the gaussian matrix reduction - * with partial pivoting followed by back/substitution with the loops manually - * unrolled. - */ -static GLboolean invert_matrix_general( GLmatrix *mat ) -{ - const GLfloat *m = mat->m; - GLfloat *out = mat->inv; - GLfloat wtmp[4][8]; - GLfloat m0, m1, m2, m3, s; - GLfloat *r0, *r1, *r2, *r3; - - r0 = wtmp[0], r1 = wtmp[1], r2 = wtmp[2], r3 = wtmp[3]; - - r0[0] = MAT(m,0,0), r0[1] = MAT(m,0,1), - r0[2] = MAT(m,0,2), r0[3] = MAT(m,0,3), - r0[4] = 1.0, r0[5] = r0[6] = r0[7] = 0.0, - - r1[0] = MAT(m,1,0), r1[1] = MAT(m,1,1), - r1[2] = MAT(m,1,2), r1[3] = MAT(m,1,3), - r1[5] = 1.0, r1[4] = r1[6] = r1[7] = 0.0, - - r2[0] = MAT(m,2,0), r2[1] = MAT(m,2,1), - r2[2] = MAT(m,2,2), r2[3] = MAT(m,2,3), - r2[6] = 1.0, r2[4] = r2[5] = r2[7] = 0.0, - - r3[0] = MAT(m,3,0), r3[1] = MAT(m,3,1), - r3[2] = MAT(m,3,2), r3[3] = MAT(m,3,3), - r3[7] = 1.0, r3[4] = r3[5] = r3[6] = 0.0; - - /* choose pivot - or die */ - if (FABSF(r3[0])>FABSF(r2[0])) SWAP_ROWS(r3, r2); - if (FABSF(r2[0])>FABSF(r1[0])) SWAP_ROWS(r2, r1); - if (FABSF(r1[0])>FABSF(r0[0])) SWAP_ROWS(r1, r0); - if (0.0 == r0[0]) return GL_FALSE; - - /* eliminate first variable */ - m1 = r1[0]/r0[0]; m2 = r2[0]/r0[0]; m3 = r3[0]/r0[0]; - s = r0[1]; r1[1] -= m1 * s; r2[1] -= m2 * s; r3[1] -= m3 * s; - s = r0[2]; r1[2] -= m1 * s; r2[2] -= m2 * s; r3[2] -= m3 * s; - s = r0[3]; r1[3] -= m1 * s; r2[3] -= m2 * s; r3[3] -= m3 * s; - s = r0[4]; - if (s != 0.0) { r1[4] -= m1 * s; r2[4] -= m2 * s; r3[4] -= m3 * s; } - s = r0[5]; - if (s != 0.0) { r1[5] -= m1 * s; r2[5] -= m2 * s; r3[5] -= m3 * s; } - s = r0[6]; - if (s != 0.0) { r1[6] -= m1 * s; r2[6] -= m2 * s; r3[6] -= m3 * s; } - s = r0[7]; - if (s != 0.0) { r1[7] -= m1 * s; r2[7] -= m2 * s; r3[7] -= m3 * s; } - - /* choose pivot - or die */ - if (FABSF(r3[1])>FABSF(r2[1])) SWAP_ROWS(r3, r2); - if (FABSF(r2[1])>FABSF(r1[1])) SWAP_ROWS(r2, r1); - if (0.0 == r1[1]) return GL_FALSE; - - /* eliminate second variable */ - m2 = r2[1]/r1[1]; m3 = r3[1]/r1[1]; - r2[2] -= m2 * r1[2]; r3[2] -= m3 * r1[2]; - r2[3] -= m2 * r1[3]; r3[3] -= m3 * r1[3]; - s = r1[4]; if (0.0 != s) { r2[4] -= m2 * s; r3[4] -= m3 * s; } - s = r1[5]; if (0.0 != s) { r2[5] -= m2 * s; r3[5] -= m3 * s; } - s = r1[6]; if (0.0 != s) { r2[6] -= m2 * s; r3[6] -= m3 * s; } - s = r1[7]; if (0.0 != s) { r2[7] -= m2 * s; r3[7] -= m3 * s; } - - /* choose pivot - or die */ - if (FABSF(r3[2])>FABSF(r2[2])) SWAP_ROWS(r3, r2); - if (0.0 == r2[2]) return GL_FALSE; - - /* eliminate third variable */ - m3 = r3[2]/r2[2]; - r3[3] -= m3 * r2[3], r3[4] -= m3 * r2[4], - r3[5] -= m3 * r2[5], r3[6] -= m3 * r2[6], - r3[7] -= m3 * r2[7]; - - /* last check */ - if (0.0 == r3[3]) return GL_FALSE; - - s = 1.0F/r3[3]; /* now back substitute row 3 */ - r3[4] *= s; r3[5] *= s; r3[6] *= s; r3[7] *= s; - - m2 = r2[3]; /* now back substitute row 2 */ - s = 1.0F/r2[2]; - r2[4] = s * (r2[4] - r3[4] * m2), r2[5] = s * (r2[5] - r3[5] * m2), - r2[6] = s * (r2[6] - r3[6] * m2), r2[7] = s * (r2[7] - r3[7] * m2); - m1 = r1[3]; - r1[4] -= r3[4] * m1, r1[5] -= r3[5] * m1, - r1[6] -= r3[6] * m1, r1[7] -= r3[7] * m1; - m0 = r0[3]; - r0[4] -= r3[4] * m0, r0[5] -= r3[5] * m0, - r0[6] -= r3[6] * m0, r0[7] -= r3[7] * m0; - - m1 = r1[2]; /* now back substitute row 1 */ - s = 1.0F/r1[1]; - r1[4] = s * (r1[4] - r2[4] * m1), r1[5] = s * (r1[5] - r2[5] * m1), - r1[6] = s * (r1[6] - r2[6] * m1), r1[7] = s * (r1[7] - r2[7] * m1); - m0 = r0[2]; - r0[4] -= r2[4] * m0, r0[5] -= r2[5] * m0, - r0[6] -= r2[6] * m0, r0[7] -= r2[7] * m0; - - m0 = r0[1]; /* now back substitute row 0 */ - s = 1.0F/r0[0]; - r0[4] = s * (r0[4] - r1[4] * m0), r0[5] = s * (r0[5] - r1[5] * m0), - r0[6] = s * (r0[6] - r1[6] * m0), r0[7] = s * (r0[7] - r1[7] * m0); - - MAT(out,0,0) = r0[4]; MAT(out,0,1) = r0[5], - MAT(out,0,2) = r0[6]; MAT(out,0,3) = r0[7], - MAT(out,1,0) = r1[4]; MAT(out,1,1) = r1[5], - MAT(out,1,2) = r1[6]; MAT(out,1,3) = r1[7], - MAT(out,2,0) = r2[4]; MAT(out,2,1) = r2[5], - MAT(out,2,2) = r2[6]; MAT(out,2,3) = r2[7], - MAT(out,3,0) = r3[4]; MAT(out,3,1) = r3[5], - MAT(out,3,2) = r3[6]; MAT(out,3,3) = r3[7]; - - return GL_TRUE; -} -#undef SWAP_ROWS - -/** - * Compute inverse of a general 3d transformation matrix. - * - * \param mat pointer to a GLmatrix structure. The matrix inverse will be - * stored in the GLmatrix::inv attribute. - * - * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). - * - * \author Adapted from graphics gems II. - * - * Calculates the inverse of the upper left by first calculating its - * determinant and multiplying it to the symmetric adjust matrix of each - * element. Finally deals with the translation part by transforming the - * original translation vector using by the calculated submatrix inverse. - */ -static GLboolean invert_matrix_3d_general( GLmatrix *mat ) -{ - const GLfloat *in = mat->m; - GLfloat *out = mat->inv; - GLfloat pos, neg, t; - GLfloat det; - - /* Calculate the determinant of upper left 3x3 submatrix and - * determine if the matrix is singular. - */ - pos = neg = 0.0; - t = MAT(in,0,0) * MAT(in,1,1) * MAT(in,2,2); - if (t >= 0.0) pos += t; else neg += t; - - t = MAT(in,1,0) * MAT(in,2,1) * MAT(in,0,2); - if (t >= 0.0) pos += t; else neg += t; - - t = MAT(in,2,0) * MAT(in,0,1) * MAT(in,1,2); - if (t >= 0.0) pos += t; else neg += t; - - t = -MAT(in,2,0) * MAT(in,1,1) * MAT(in,0,2); - if (t >= 0.0) pos += t; else neg += t; - - t = -MAT(in,1,0) * MAT(in,0,1) * MAT(in,2,2); - if (t >= 0.0) pos += t; else neg += t; - - t = -MAT(in,0,0) * MAT(in,2,1) * MAT(in,1,2); - if (t >= 0.0) pos += t; else neg += t; - - det = pos + neg; - - if (det*det < 1e-25) - return GL_FALSE; - - det = 1.0F / det; - MAT(out,0,0) = ( (MAT(in,1,1)*MAT(in,2,2) - MAT(in,2,1)*MAT(in,1,2) )*det); - MAT(out,0,1) = (- (MAT(in,0,1)*MAT(in,2,2) - MAT(in,2,1)*MAT(in,0,2) )*det); - MAT(out,0,2) = ( (MAT(in,0,1)*MAT(in,1,2) - MAT(in,1,1)*MAT(in,0,2) )*det); - MAT(out,1,0) = (- (MAT(in,1,0)*MAT(in,2,2) - MAT(in,2,0)*MAT(in,1,2) )*det); - MAT(out,1,1) = ( (MAT(in,0,0)*MAT(in,2,2) - MAT(in,2,0)*MAT(in,0,2) )*det); - MAT(out,1,2) = (- (MAT(in,0,0)*MAT(in,1,2) - MAT(in,1,0)*MAT(in,0,2) )*det); - MAT(out,2,0) = ( (MAT(in,1,0)*MAT(in,2,1) - MAT(in,2,0)*MAT(in,1,1) )*det); - MAT(out,2,1) = (- (MAT(in,0,0)*MAT(in,2,1) - MAT(in,2,0)*MAT(in,0,1) )*det); - MAT(out,2,2) = ( (MAT(in,0,0)*MAT(in,1,1) - MAT(in,1,0)*MAT(in,0,1) )*det); - - /* Do the translation part */ - MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0) + - MAT(in,1,3) * MAT(out,0,1) + - MAT(in,2,3) * MAT(out,0,2) ); - MAT(out,1,3) = - (MAT(in,0,3) * MAT(out,1,0) + - MAT(in,1,3) * MAT(out,1,1) + - MAT(in,2,3) * MAT(out,1,2) ); - MAT(out,2,3) = - (MAT(in,0,3) * MAT(out,2,0) + - MAT(in,1,3) * MAT(out,2,1) + - MAT(in,2,3) * MAT(out,2,2) ); - - return GL_TRUE; -} - -/** - * Compute inverse of a 3d transformation matrix. - * - * \param mat pointer to a GLmatrix structure. The matrix inverse will be - * stored in the GLmatrix::inv attribute. - * - * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). - * - * If the matrix is not an angle preserving matrix then calls - * invert_matrix_3d_general for the actual calculation. Otherwise calculates - * the inverse matrix analyzing and inverting each of the scaling, rotation and - * translation parts. - */ -static GLboolean invert_matrix_3d( GLmatrix *mat ) -{ - const GLfloat *in = mat->m; - GLfloat *out = mat->inv; - - if (!TEST_MAT_FLAGS(mat, MAT_FLAGS_ANGLE_PRESERVING)) { - return invert_matrix_3d_general( mat ); - } - - if (mat->flags & MAT_FLAG_UNIFORM_SCALE) { - GLfloat scale = (MAT(in,0,0) * MAT(in,0,0) + - MAT(in,0,1) * MAT(in,0,1) + - MAT(in,0,2) * MAT(in,0,2)); - - if (scale == 0.0) - return GL_FALSE; - - scale = 1.0F / scale; - - /* Transpose and scale the 3 by 3 upper-left submatrix. */ - MAT(out,0,0) = scale * MAT(in,0,0); - MAT(out,1,0) = scale * MAT(in,0,1); - MAT(out,2,0) = scale * MAT(in,0,2); - MAT(out,0,1) = scale * MAT(in,1,0); - MAT(out,1,1) = scale * MAT(in,1,1); - MAT(out,2,1) = scale * MAT(in,1,2); - MAT(out,0,2) = scale * MAT(in,2,0); - MAT(out,1,2) = scale * MAT(in,2,1); - MAT(out,2,2) = scale * MAT(in,2,2); - } - else if (mat->flags & MAT_FLAG_ROTATION) { - /* Transpose the 3 by 3 upper-left submatrix. */ - MAT(out,0,0) = MAT(in,0,0); - MAT(out,1,0) = MAT(in,0,1); - MAT(out,2,0) = MAT(in,0,2); - MAT(out,0,1) = MAT(in,1,0); - MAT(out,1,1) = MAT(in,1,1); - MAT(out,2,1) = MAT(in,1,2); - MAT(out,0,2) = MAT(in,2,0); - MAT(out,1,2) = MAT(in,2,1); - MAT(out,2,2) = MAT(in,2,2); - } - else { - /* pure translation */ - memcpy( out, Identity, sizeof(Identity) ); - MAT(out,0,3) = - MAT(in,0,3); - MAT(out,1,3) = - MAT(in,1,3); - MAT(out,2,3) = - MAT(in,2,3); - return GL_TRUE; - } - - if (mat->flags & MAT_FLAG_TRANSLATION) { - /* Do the translation part */ - MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0) + - MAT(in,1,3) * MAT(out,0,1) + - MAT(in,2,3) * MAT(out,0,2) ); - MAT(out,1,3) = - (MAT(in,0,3) * MAT(out,1,0) + - MAT(in,1,3) * MAT(out,1,1) + - MAT(in,2,3) * MAT(out,1,2) ); - MAT(out,2,3) = - (MAT(in,0,3) * MAT(out,2,0) + - MAT(in,1,3) * MAT(out,2,1) + - MAT(in,2,3) * MAT(out,2,2) ); - } - else { - MAT(out,0,3) = MAT(out,1,3) = MAT(out,2,3) = 0.0; - } - - return GL_TRUE; -} - -/** - * Compute inverse of an identity transformation matrix. - * - * \param mat pointer to a GLmatrix structure. The matrix inverse will be - * stored in the GLmatrix::inv attribute. - * - * \return always GL_TRUE. - * - * Simply copies Identity into GLmatrix::inv. - */ -static GLboolean invert_matrix_identity( GLmatrix *mat ) -{ - memcpy( mat->inv, Identity, sizeof(Identity) ); - return GL_TRUE; -} - -/** - * Compute inverse of a no-rotation 3d transformation matrix. - * - * \param mat pointer to a GLmatrix structure. The matrix inverse will be - * stored in the GLmatrix::inv attribute. - * - * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). - * - * Calculates the - */ -static GLboolean invert_matrix_3d_no_rot( GLmatrix *mat ) -{ - const GLfloat *in = mat->m; - GLfloat *out = mat->inv; - - if (MAT(in,0,0) == 0 || MAT(in,1,1) == 0 || MAT(in,2,2) == 0 ) - return GL_FALSE; - - memcpy( out, Identity, 16 * sizeof(GLfloat) ); - MAT(out,0,0) = 1.0F / MAT(in,0,0); - MAT(out,1,1) = 1.0F / MAT(in,1,1); - MAT(out,2,2) = 1.0F / MAT(in,2,2); - - if (mat->flags & MAT_FLAG_TRANSLATION) { - MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0)); - MAT(out,1,3) = - (MAT(in,1,3) * MAT(out,1,1)); - MAT(out,2,3) = - (MAT(in,2,3) * MAT(out,2,2)); - } - - return GL_TRUE; -} - -/** - * Compute inverse of a no-rotation 2d transformation matrix. - * - * \param mat pointer to a GLmatrix structure. The matrix inverse will be - * stored in the GLmatrix::inv attribute. - * - * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). - * - * Calculates the inverse matrix by applying the inverse scaling and - * translation to the identity matrix. - */ -static GLboolean invert_matrix_2d_no_rot( GLmatrix *mat ) -{ - const GLfloat *in = mat->m; - GLfloat *out = mat->inv; - - if (MAT(in,0,0) == 0 || MAT(in,1,1) == 0) - return GL_FALSE; - - memcpy( out, Identity, 16 * sizeof(GLfloat) ); - MAT(out,0,0) = 1.0F / MAT(in,0,0); - MAT(out,1,1) = 1.0F / MAT(in,1,1); - - if (mat->flags & MAT_FLAG_TRANSLATION) { - MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0)); - MAT(out,1,3) = - (MAT(in,1,3) * MAT(out,1,1)); - } - - return GL_TRUE; -} - -#if 0 -/* broken */ -static GLboolean invert_matrix_perspective( GLmatrix *mat ) -{ - const GLfloat *in = mat->m; - GLfloat *out = mat->inv; - - if (MAT(in,2,3) == 0) - return GL_FALSE; - - memcpy( out, Identity, 16 * sizeof(GLfloat) ); - - MAT(out,0,0) = 1.0F / MAT(in,0,0); - MAT(out,1,1) = 1.0F / MAT(in,1,1); - - MAT(out,0,3) = MAT(in,0,2); - MAT(out,1,3) = MAT(in,1,2); - - MAT(out,2,2) = 0; - MAT(out,2,3) = -1; - - MAT(out,3,2) = 1.0F / MAT(in,2,3); - MAT(out,3,3) = MAT(in,2,2) * MAT(out,3,2); - - return GL_TRUE; -} -#endif - -/** - * Matrix inversion function pointer type. - */ -typedef GLboolean (*inv_mat_func)( GLmatrix *mat ); - -/** - * Table of the matrix inversion functions according to the matrix type. - */ -static inv_mat_func inv_mat_tab[7] = { - invert_matrix_general, - invert_matrix_identity, - invert_matrix_3d_no_rot, -#if 0 - /* Don't use this function for now - it fails when the projection matrix - * is premultiplied by a translation (ala Chromium's tilesort SPU). - */ - invert_matrix_perspective, -#else - invert_matrix_general, -#endif - invert_matrix_3d, /* lazy! */ - invert_matrix_2d_no_rot, - invert_matrix_3d -}; - -/** - * Compute inverse of a transformation matrix. - * - * \param mat pointer to a GLmatrix structure. The matrix inverse will be - * stored in the GLmatrix::inv attribute. - * - * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). - * - * Calls the matrix inversion function in inv_mat_tab corresponding to the - * given matrix type. In case of failure, updates the MAT_FLAG_SINGULAR flag, - * and copies the identity matrix into GLmatrix::inv. - */ -static GLboolean matrix_invert( GLmatrix *mat ) -{ - if (inv_mat_tab[mat->type](mat)) { - mat->flags &= ~MAT_FLAG_SINGULAR; - return GL_TRUE; - } else { - mat->flags |= MAT_FLAG_SINGULAR; - memcpy( mat->inv, Identity, sizeof(Identity) ); - return GL_FALSE; - } -} - -/*@}*/ - - -/**********************************************************************/ -/** \name Matrix generation */ -/*@{*/ - -/** - * Generate a 4x4 transformation matrix from glRotate parameters, and - * post-multiply the input matrix by it. - * - * \author - * This function was contributed by Erich Boleyn (erich@uruk.org). - * Optimizations contributed by Rudolf Opalla (rudi@khm.de). - */ -void -_math_matrix_rotate( GLmatrix *mat, - GLfloat angle, GLfloat x, GLfloat y, GLfloat z ) -{ - GLfloat xx, yy, zz, xy, yz, zx, xs, ys, zs, one_c, s, c; - GLfloat m[16]; - GLboolean optimized; - - s = (GLfloat) sin( angle * DEG2RAD ); - c = (GLfloat) cos( angle * DEG2RAD ); - - memcpy(m, Identity, sizeof(GLfloat)*16); - optimized = GL_FALSE; - -#define M(row,col) m[col*4+row] - - if (x == 0.0F) { - if (y == 0.0F) { - if (z != 0.0F) { - optimized = GL_TRUE; - /* rotate only around z-axis */ - M(0,0) = c; - M(1,1) = c; - if (z < 0.0F) { - M(0,1) = s; - M(1,0) = -s; - } - else { - M(0,1) = -s; - M(1,0) = s; - } - } - } - else if (z == 0.0F) { - optimized = GL_TRUE; - /* rotate only around y-axis */ - M(0,0) = c; - M(2,2) = c; - if (y < 0.0F) { - M(0,2) = -s; - M(2,0) = s; - } - else { - M(0,2) = s; - M(2,0) = -s; - } - } - } - else if (y == 0.0F) { - if (z == 0.0F) { - optimized = GL_TRUE; - /* rotate only around x-axis */ - M(1,1) = c; - M(2,2) = c; - if (x < 0.0F) { - M(1,2) = s; - M(2,1) = -s; - } - else { - M(1,2) = -s; - M(2,1) = s; - } - } - } - - if (!optimized) { - const GLfloat mag = SQRTF(x * x + y * y + z * z); - - if (mag <= 1.0e-4) { - /* no rotation, leave mat as-is */ - return; - } - - x /= mag; - y /= mag; - z /= mag; - - - /* - * Arbitrary axis rotation matrix. - * - * This is composed of 5 matrices, Rz, Ry, T, Ry', Rz', multiplied - * like so: Rz * Ry * T * Ry' * Rz'. T is the final rotation - * (which is about the X-axis), and the two composite transforms - * Ry' * Rz' and Rz * Ry are (respectively) the rotations necessary - * from the arbitrary axis to the X-axis then back. They are - * all elementary rotations. - * - * Rz' is a rotation about the Z-axis, to bring the axis vector - * into the x-z plane. Then Ry' is applied, rotating about the - * Y-axis to bring the axis vector parallel with the X-axis. The - * rotation about the X-axis is then performed. Ry and Rz are - * simply the respective inverse transforms to bring the arbitrary - * axis back to its original orientation. The first transforms - * Rz' and Ry' are considered inverses, since the data from the - * arbitrary axis gives you info on how to get to it, not how - * to get away from it, and an inverse must be applied. - * - * The basic calculation used is to recognize that the arbitrary - * axis vector (x, y, z), since it is of unit length, actually - * represents the sines and cosines of the angles to rotate the - * X-axis to the same orientation, with theta being the angle about - * Z and phi the angle about Y (in the order described above) - * as follows: - * - * cos ( theta ) = x / sqrt ( 1 - z^2 ) - * sin ( theta ) = y / sqrt ( 1 - z^2 ) - * - * cos ( phi ) = sqrt ( 1 - z^2 ) - * sin ( phi ) = z - * - * Note that cos ( phi ) can further be inserted to the above - * formulas: - * - * cos ( theta ) = x / cos ( phi ) - * sin ( theta ) = y / sin ( phi ) - * - * ...etc. Because of those relations and the standard trigonometric - * relations, it is pssible to reduce the transforms down to what - * is used below. It may be that any primary axis chosen will give the - * same results (modulo a sign convention) using thie method. - * - * Particularly nice is to notice that all divisions that might - * have caused trouble when parallel to certain planes or - * axis go away with care paid to reducing the expressions. - * After checking, it does perform correctly under all cases, since - * in all the cases of division where the denominator would have - * been zero, the numerator would have been zero as well, giving - * the expected result. - */ - - xx = x * x; - yy = y * y; - zz = z * z; - xy = x * y; - yz = y * z; - zx = z * x; - xs = x * s; - ys = y * s; - zs = z * s; - one_c = 1.0F - c; - - /* We already hold the identity-matrix so we can skip some statements */ - M(0,0) = (one_c * xx) + c; - M(0,1) = (one_c * xy) - zs; - M(0,2) = (one_c * zx) + ys; -/* M(0,3) = 0.0F; */ - - M(1,0) = (one_c * xy) + zs; - M(1,1) = (one_c * yy) + c; - M(1,2) = (one_c * yz) - xs; -/* M(1,3) = 0.0F; */ - - M(2,0) = (one_c * zx) - ys; - M(2,1) = (one_c * yz) + xs; - M(2,2) = (one_c * zz) + c; -/* M(2,3) = 0.0F; */ - -/* - M(3,0) = 0.0F; - M(3,1) = 0.0F; - M(3,2) = 0.0F; - M(3,3) = 1.0F; -*/ - } -#undef M - - matrix_multf( mat, m, MAT_FLAG_ROTATION ); -} - -/** - * Apply a perspective projection matrix. - * - * \param mat matrix to apply the projection. - * \param left left clipping plane coordinate. - * \param right right clipping plane coordinate. - * \param bottom bottom clipping plane coordinate. - * \param top top clipping plane coordinate. - * \param nearval distance to the near clipping plane. - * \param farval distance to the far clipping plane. - * - * Creates the projection matrix and multiplies it with \p mat, marking the - * MAT_FLAG_PERSPECTIVE flag. - */ -void -_math_matrix_frustum( GLmatrix *mat, - GLfloat left, GLfloat right, - GLfloat bottom, GLfloat top, - GLfloat nearval, GLfloat farval ) -{ - GLfloat x, y, a, b, c, d; - GLfloat m[16]; - - x = (2.0F*nearval) / (right-left); - y = (2.0F*nearval) / (top-bottom); - a = (right+left) / (right-left); - b = (top+bottom) / (top-bottom); - c = -(farval+nearval) / ( farval-nearval); - d = -(2.0F*farval*nearval) / (farval-nearval); /* error? */ - -#define M(row,col) m[col*4+row] - M(0,0) = x; M(0,1) = 0.0F; M(0,2) = a; M(0,3) = 0.0F; - M(1,0) = 0.0F; M(1,1) = y; M(1,2) = b; M(1,3) = 0.0F; - M(2,0) = 0.0F; M(2,1) = 0.0F; M(2,2) = c; M(2,3) = d; - M(3,0) = 0.0F; M(3,1) = 0.0F; M(3,2) = -1.0F; M(3,3) = 0.0F; -#undef M - - matrix_multf( mat, m, MAT_FLAG_PERSPECTIVE ); -} - -/** - * Apply an orthographic projection matrix. - * - * \param mat matrix to apply the projection. - * \param left left clipping plane coordinate. - * \param right right clipping plane coordinate. - * \param bottom bottom clipping plane coordinate. - * \param top top clipping plane coordinate. - * \param nearval distance to the near clipping plane. - * \param farval distance to the far clipping plane. - * - * Creates the projection matrix and multiplies it with \p mat, marking the - * MAT_FLAG_GENERAL_SCALE and MAT_FLAG_TRANSLATION flags. - */ -void -_math_matrix_ortho( GLmatrix *mat, - GLfloat left, GLfloat right, - GLfloat bottom, GLfloat top, - GLfloat nearval, GLfloat farval ) -{ - GLfloat m[16]; - -#define M(row,col) m[col*4+row] - M(0,0) = 2.0F / (right-left); - M(0,1) = 0.0F; - M(0,2) = 0.0F; - M(0,3) = -(right+left) / (right-left); - - M(1,0) = 0.0F; - M(1,1) = 2.0F / (top-bottom); - M(1,2) = 0.0F; - M(1,3) = -(top+bottom) / (top-bottom); - - M(2,0) = 0.0F; - M(2,1) = 0.0F; - M(2,2) = -2.0F / (farval-nearval); - M(2,3) = -(farval+nearval) / (farval-nearval); - - M(3,0) = 0.0F; - M(3,1) = 0.0F; - M(3,2) = 0.0F; - M(3,3) = 1.0F; -#undef M - - matrix_multf( mat, m, (MAT_FLAG_GENERAL_SCALE|MAT_FLAG_TRANSLATION)); -} - -/** - * Multiply a matrix with a general scaling matrix. - * - * \param mat matrix. - * \param x x axis scale factor. - * \param y y axis scale factor. - * \param z z axis scale factor. - * - * Multiplies in-place the elements of \p mat by the scale factors. Checks if - * the scales factors are roughly the same, marking the MAT_FLAG_UNIFORM_SCALE - * flag, or MAT_FLAG_GENERAL_SCALE. Marks the MAT_DIRTY_TYPE and - * MAT_DIRTY_INVERSE dirty flags. - */ -void -_math_matrix_scale( GLmatrix *mat, GLfloat x, GLfloat y, GLfloat z ) -{ - GLfloat *m = mat->m; - m[0] *= x; m[4] *= y; m[8] *= z; - m[1] *= x; m[5] *= y; m[9] *= z; - m[2] *= x; m[6] *= y; m[10] *= z; - m[3] *= x; m[7] *= y; m[11] *= z; - - if (FABSF(x - y) < 1e-8 && FABSF(x - z) < 1e-8) - mat->flags |= MAT_FLAG_UNIFORM_SCALE; - else - mat->flags |= MAT_FLAG_GENERAL_SCALE; - - mat->flags |= (MAT_DIRTY_TYPE | - MAT_DIRTY_INVERSE); -} - -/** - * Multiply a matrix with a translation matrix. - * - * \param mat matrix. - * \param x translation vector x coordinate. - * \param y translation vector y coordinate. - * \param z translation vector z coordinate. - * - * Adds the translation coordinates to the elements of \p mat in-place. Marks - * the MAT_FLAG_TRANSLATION flag, and the MAT_DIRTY_TYPE and MAT_DIRTY_INVERSE - * dirty flags. - */ -void -_math_matrix_translate( GLmatrix *mat, GLfloat x, GLfloat y, GLfloat z ) -{ - GLfloat *m = mat->m; - m[12] = m[0] * x + m[4] * y + m[8] * z + m[12]; - m[13] = m[1] * x + m[5] * y + m[9] * z + m[13]; - m[14] = m[2] * x + m[6] * y + m[10] * z + m[14]; - m[15] = m[3] * x + m[7] * y + m[11] * z + m[15]; - - mat->flags |= (MAT_FLAG_TRANSLATION | - MAT_DIRTY_TYPE | - MAT_DIRTY_INVERSE); -} - - -/** - * Set matrix to do viewport and depthrange mapping. - * Transforms Normalized Device Coords to window/Z values. - */ -void -_math_matrix_viewport(GLmatrix *m, GLint x, GLint y, GLint width, GLint height, - GLfloat zNear, GLfloat zFar, GLfloat depthMax) -{ - m->m[MAT_SX] = (GLfloat) width / 2.0F; - m->m[MAT_TX] = m->m[MAT_SX] + x; - m->m[MAT_SY] = (GLfloat) height / 2.0F; - m->m[MAT_TY] = m->m[MAT_SY] + y; - m->m[MAT_SZ] = depthMax * ((zFar - zNear) / 2.0F); - m->m[MAT_TZ] = depthMax * ((zFar - zNear) / 2.0F + zNear); - m->flags = MAT_FLAG_GENERAL_SCALE | MAT_FLAG_TRANSLATION; - m->type = MATRIX_3D_NO_ROT; -} - - -/** - * Set a matrix to the identity matrix. - * - * \param mat matrix. - * - * Copies ::Identity into \p GLmatrix::m, and into GLmatrix::inv if not NULL. - * Sets the matrix type to identity, and clear the dirty flags. - */ -void -_math_matrix_set_identity( GLmatrix *mat ) -{ - memcpy( mat->m, Identity, 16*sizeof(GLfloat) ); - - if (mat->inv) - memcpy( mat->inv, Identity, 16*sizeof(GLfloat) ); - - mat->type = MATRIX_IDENTITY; - mat->flags &= ~(MAT_DIRTY_FLAGS| - MAT_DIRTY_TYPE| - MAT_DIRTY_INVERSE); -} - -/*@}*/ - - -/**********************************************************************/ -/** \name Matrix analysis */ -/*@{*/ - -#define ZERO(x) (1<m; - GLuint mask = 0; - GLuint i; - - for (i = 0 ; i < 16 ; i++) { - if (m[i] == 0.0) mask |= (1<flags &= ~MAT_FLAGS_GEOMETRY; - - /* Check for translation - no-one really cares - */ - if ((mask & MASK_NO_TRX) != MASK_NO_TRX) - mat->flags |= MAT_FLAG_TRANSLATION; - - /* Do the real work - */ - if (mask == (GLuint) MASK_IDENTITY) { - mat->type = MATRIX_IDENTITY; - } - else if ((mask & MASK_2D_NO_ROT) == (GLuint) MASK_2D_NO_ROT) { - mat->type = MATRIX_2D_NO_ROT; - - if ((mask & MASK_NO_2D_SCALE) != MASK_NO_2D_SCALE) - mat->flags |= MAT_FLAG_GENERAL_SCALE; - } - else if ((mask & MASK_2D) == (GLuint) MASK_2D) { - GLfloat mm = DOT2(m, m); - GLfloat m4m4 = DOT2(m+4,m+4); - GLfloat mm4 = DOT2(m,m+4); - - mat->type = MATRIX_2D; - - /* Check for scale */ - if (SQ(mm-1) > SQ(1e-6) || - SQ(m4m4-1) > SQ(1e-6)) - mat->flags |= MAT_FLAG_GENERAL_SCALE; - - /* Check for rotation */ - if (SQ(mm4) > SQ(1e-6)) - mat->flags |= MAT_FLAG_GENERAL_3D; - else - mat->flags |= MAT_FLAG_ROTATION; - - } - else if ((mask & MASK_3D_NO_ROT) == (GLuint) MASK_3D_NO_ROT) { - mat->type = MATRIX_3D_NO_ROT; - - /* Check for scale */ - if (SQ(m[0]-m[5]) < SQ(1e-6) && - SQ(m[0]-m[10]) < SQ(1e-6)) { - if (SQ(m[0]-1.0) > SQ(1e-6)) { - mat->flags |= MAT_FLAG_UNIFORM_SCALE; - } - } - else { - mat->flags |= MAT_FLAG_GENERAL_SCALE; - } - } - else if ((mask & MASK_3D) == (GLuint) MASK_3D) { - GLfloat c1 = DOT3(m,m); - GLfloat c2 = DOT3(m+4,m+4); - GLfloat c3 = DOT3(m+8,m+8); - GLfloat d1 = DOT3(m, m+4); - GLfloat cp[3]; - - mat->type = MATRIX_3D; - - /* Check for scale */ - if (SQ(c1-c2) < SQ(1e-6) && SQ(c1-c3) < SQ(1e-6)) { - if (SQ(c1-1.0) > SQ(1e-6)) - mat->flags |= MAT_FLAG_UNIFORM_SCALE; - /* else no scale at all */ - } - else { - mat->flags |= MAT_FLAG_GENERAL_SCALE; - } - - /* Check for rotation */ - if (SQ(d1) < SQ(1e-6)) { - CROSS3( cp, m, m+4 ); - SUB_3V( cp, cp, (m+8) ); - if (LEN_SQUARED_3FV(cp) < SQ(1e-6)) - mat->flags |= MAT_FLAG_ROTATION; - else - mat->flags |= MAT_FLAG_GENERAL_3D; - } - else { - mat->flags |= MAT_FLAG_GENERAL_3D; /* shear, etc */ - } - } - else if ((mask & MASK_PERSPECTIVE) == MASK_PERSPECTIVE && m[11]==-1.0F) { - mat->type = MATRIX_PERSPECTIVE; - mat->flags |= MAT_FLAG_GENERAL; - } - else { - mat->type = MATRIX_GENERAL; - mat->flags |= MAT_FLAG_GENERAL; - } -} - -/** - * Analyze a matrix given that its flags are accurate. - * - * This is the more common operation, hopefully. - */ -static void analyse_from_flags( GLmatrix *mat ) -{ - const GLfloat *m = mat->m; - - if (TEST_MAT_FLAGS(mat, 0)) { - mat->type = MATRIX_IDENTITY; - } - else if (TEST_MAT_FLAGS(mat, (MAT_FLAG_TRANSLATION | - MAT_FLAG_UNIFORM_SCALE | - MAT_FLAG_GENERAL_SCALE))) { - if ( m[10]==1.0F && m[14]==0.0F ) { - mat->type = MATRIX_2D_NO_ROT; - } - else { - mat->type = MATRIX_3D_NO_ROT; - } - } - else if (TEST_MAT_FLAGS(mat, MAT_FLAGS_3D)) { - if ( m[ 8]==0.0F - && m[ 9]==0.0F - && m[2]==0.0F && m[6]==0.0F && m[10]==1.0F && m[14]==0.0F) { - mat->type = MATRIX_2D; - } - else { - mat->type = MATRIX_3D; - } - } - else if ( m[4]==0.0F && m[12]==0.0F - && m[1]==0.0F && m[13]==0.0F - && m[2]==0.0F && m[6]==0.0F - && m[3]==0.0F && m[7]==0.0F && m[11]==-1.0F && m[15]==0.0F) { - mat->type = MATRIX_PERSPECTIVE; - } - else { - mat->type = MATRIX_GENERAL; - } -} - -/** - * Analyze and update a matrix. - * - * \param mat matrix. - * - * If the matrix type is dirty then calls either analyse_from_scratch() or - * analyse_from_flags() to determine its type, according to whether the flags - * are dirty or not, respectively. If the matrix has an inverse and it's dirty - * then calls matrix_invert(). Finally clears the dirty flags. - */ -void -_math_matrix_analyse( GLmatrix *mat ) -{ - if (mat->flags & MAT_DIRTY_TYPE) { - if (mat->flags & MAT_DIRTY_FLAGS) - analyse_from_scratch( mat ); - else - analyse_from_flags( mat ); - } - - if (mat->inv && (mat->flags & MAT_DIRTY_INVERSE)) { - matrix_invert( mat ); - mat->flags &= ~MAT_DIRTY_INVERSE; - } - - mat->flags &= ~(MAT_DIRTY_FLAGS | MAT_DIRTY_TYPE); -} - -/*@}*/ - - -/** - * Test if the given matrix preserves vector lengths. - */ -GLboolean -_math_matrix_is_length_preserving( const GLmatrix *m ) -{ - return TEST_MAT_FLAGS( m, MAT_FLAGS_LENGTH_PRESERVING); -} - - -/** - * Test if the given matrix does any rotation. - * (or perhaps if the upper-left 3x3 is non-identity) - */ -GLboolean -_math_matrix_has_rotation( const GLmatrix *m ) -{ - if (m->flags & (MAT_FLAG_GENERAL | - MAT_FLAG_ROTATION | - MAT_FLAG_GENERAL_3D | - MAT_FLAG_PERSPECTIVE)) - return GL_TRUE; - else - return GL_FALSE; -} - - -GLboolean -_math_matrix_is_general_scale( const GLmatrix *m ) -{ - return (m->flags & MAT_FLAG_GENERAL_SCALE) ? GL_TRUE : GL_FALSE; -} - - -GLboolean -_math_matrix_is_dirty( const GLmatrix *m ) -{ - return (m->flags & MAT_DIRTY) ? GL_TRUE : GL_FALSE; -} - - -/**********************************************************************/ -/** \name Matrix setup */ -/*@{*/ - -/** - * Copy a matrix. - * - * \param to destination matrix. - * \param from source matrix. - * - * Copies all fields in GLmatrix, creating an inverse array if necessary. - */ -void -_math_matrix_copy( GLmatrix *to, const GLmatrix *from ) -{ - memcpy( to->m, from->m, sizeof(Identity) ); - to->flags = from->flags; - to->type = from->type; - - if (to->inv != 0) { - if (from->inv == 0) { - matrix_invert( to ); - } - else { - memcpy(to->inv, from->inv, sizeof(GLfloat)*16); - } - } -} - -/** - * Loads a matrix array into GLmatrix. - * - * \param m matrix array. - * \param mat matrix. - * - * Copies \p m into GLmatrix::m and marks the MAT_FLAG_GENERAL and MAT_DIRTY - * flags. - */ -void -_math_matrix_loadf( GLmatrix *mat, const GLfloat *m ) -{ - memcpy( mat->m, m, 16*sizeof(GLfloat) ); - mat->flags = (MAT_FLAG_GENERAL | MAT_DIRTY); -} - -/** - * Matrix constructor. - * - * \param m matrix. - * - * Initialize the GLmatrix fields. - */ -void -_math_matrix_ctr( GLmatrix *m ) -{ - m->m = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); - if (m->m) - memcpy( m->m, Identity, sizeof(Identity) ); - m->inv = NULL; - m->type = MATRIX_IDENTITY; - m->flags = 0; -} - -/** - * Matrix destructor. - * - * \param m matrix. - * - * Frees the data in a GLmatrix. - */ -void -_math_matrix_dtr( GLmatrix *m ) -{ - if (m->m) { - _mesa_align_free( m->m ); - m->m = NULL; - } - if (m->inv) { - _mesa_align_free( m->inv ); - m->inv = NULL; - } -} - -/** - * Allocate a matrix inverse. - * - * \param m matrix. - * - * Allocates the matrix inverse, GLmatrix::inv, and sets it to Identity. - */ -void -_math_matrix_alloc_inv( GLmatrix *m ) -{ - if (!m->inv) { - m->inv = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); - if (m->inv) - memcpy( m->inv, Identity, 16 * sizeof(GLfloat) ); - } -} - -/*@}*/ - - -/**********************************************************************/ -/** \name Matrix transpose */ -/*@{*/ - -/** - * Transpose a GLfloat matrix. - * - * \param to destination array. - * \param from source array. - */ -void -_math_transposef( GLfloat to[16], const GLfloat from[16] ) -{ - to[0] = from[0]; - to[1] = from[4]; - to[2] = from[8]; - to[3] = from[12]; - to[4] = from[1]; - to[5] = from[5]; - to[6] = from[9]; - to[7] = from[13]; - to[8] = from[2]; - to[9] = from[6]; - to[10] = from[10]; - to[11] = from[14]; - to[12] = from[3]; - to[13] = from[7]; - to[14] = from[11]; - to[15] = from[15]; -} - -/** - * Transpose a GLdouble matrix. - * - * \param to destination array. - * \param from source array. - */ -void -_math_transposed( GLdouble to[16], const GLdouble from[16] ) -{ - to[0] = from[0]; - to[1] = from[4]; - to[2] = from[8]; - to[3] = from[12]; - to[4] = from[1]; - to[5] = from[5]; - to[6] = from[9]; - to[7] = from[13]; - to[8] = from[2]; - to[9] = from[6]; - to[10] = from[10]; - to[11] = from[14]; - to[12] = from[3]; - to[13] = from[7]; - to[14] = from[11]; - to[15] = from[15]; -} - -/** - * Transpose a GLdouble matrix and convert to GLfloat. - * - * \param to destination array. - * \param from source array. - */ -void -_math_transposefd( GLfloat to[16], const GLdouble from[16] ) -{ - to[0] = (GLfloat) from[0]; - to[1] = (GLfloat) from[4]; - to[2] = (GLfloat) from[8]; - to[3] = (GLfloat) from[12]; - to[4] = (GLfloat) from[1]; - to[5] = (GLfloat) from[5]; - to[6] = (GLfloat) from[9]; - to[7] = (GLfloat) from[13]; - to[8] = (GLfloat) from[2]; - to[9] = (GLfloat) from[6]; - to[10] = (GLfloat) from[10]; - to[11] = (GLfloat) from[14]; - to[12] = (GLfloat) from[3]; - to[13] = (GLfloat) from[7]; - to[14] = (GLfloat) from[11]; - to[15] = (GLfloat) from[15]; -} - -/*@}*/ - - -/** - * Transform a 4-element row vector (1x4 matrix) by a 4x4 matrix. This - * function is used for transforming clipping plane equations and spotlight - * directions. - * Mathematically, u = v * m. - * Input: v - input vector - * m - transformation matrix - * Output: u - transformed vector - */ -void -_mesa_transform_vector( GLfloat u[4], const GLfloat v[4], const GLfloat m[16] ) -{ - const GLfloat v0 = v[0], v1 = v[1], v2 = v[2], v3 = v[3]; -#define M(row,col) m[row + col*4] - u[0] = v0 * M(0,0) + v1 * M(1,0) + v2 * M(2,0) + v3 * M(3,0); - u[1] = v0 * M(0,1) + v1 * M(1,1) + v2 * M(2,1) + v3 * M(3,1); - u[2] = v0 * M(0,2) + v1 * M(1,2) + v2 * M(2,2) + v3 * M(3,2); - u[3] = v0 * M(0,3) + v1 * M(1,3) + v2 * M(2,3) + v3 * M(3,3); -#undef M -} +/* + * Mesa 3-D graphics library + * Version: 6.3 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +/** + * \file m_matrix.c + * Matrix operations. + * + * \note + * -# 4x4 transformation matrices are stored in memory in column major order. + * -# Points/vertices are to be thought of as column vectors. + * -# Transformation of a point p by a matrix M is: p' = M * p + */ + + +#include "main/glheader.h" +#include "main/imports.h" +#include "main/macros.h" + +#include "m_matrix.h" + + +/** + * \defgroup MatFlags MAT_FLAG_XXX-flags + * + * Bitmasks to indicate different kinds of 4x4 matrices in GLmatrix::flags + * It would be nice to make all these flags private to m_matrix.c + */ +/*@{*/ +#define MAT_FLAG_IDENTITY 0 /**< is an identity matrix flag. + * (Not actually used - the identity + * matrix is identified by the absense + * of all other flags.) + */ +#define MAT_FLAG_GENERAL 0x1 /**< is a general matrix flag */ +#define MAT_FLAG_ROTATION 0x2 /**< is a rotation matrix flag */ +#define MAT_FLAG_TRANSLATION 0x4 /**< is a translation matrix flag */ +#define MAT_FLAG_UNIFORM_SCALE 0x8 /**< is an uniform scaling matrix flag */ +#define MAT_FLAG_GENERAL_SCALE 0x10 /**< is a general scaling matrix flag */ +#define MAT_FLAG_GENERAL_3D 0x20 /**< general 3D matrix flag */ +#define MAT_FLAG_PERSPECTIVE 0x40 /**< is a perspective proj matrix flag */ +#define MAT_FLAG_SINGULAR 0x80 /**< is a singular matrix flag */ +#define MAT_DIRTY_TYPE 0x100 /**< matrix type is dirty */ +#define MAT_DIRTY_FLAGS 0x200 /**< matrix flags are dirty */ +#define MAT_DIRTY_INVERSE 0x400 /**< matrix inverse is dirty */ + +/** angle preserving matrix flags mask */ +#define MAT_FLAGS_ANGLE_PRESERVING (MAT_FLAG_ROTATION | \ + MAT_FLAG_TRANSLATION | \ + MAT_FLAG_UNIFORM_SCALE) + +/** geometry related matrix flags mask */ +#define MAT_FLAGS_GEOMETRY (MAT_FLAG_GENERAL | \ + MAT_FLAG_ROTATION | \ + MAT_FLAG_TRANSLATION | \ + MAT_FLAG_UNIFORM_SCALE | \ + MAT_FLAG_GENERAL_SCALE | \ + MAT_FLAG_GENERAL_3D | \ + MAT_FLAG_PERSPECTIVE | \ + MAT_FLAG_SINGULAR) + +/** length preserving matrix flags mask */ +#define MAT_FLAGS_LENGTH_PRESERVING (MAT_FLAG_ROTATION | \ + MAT_FLAG_TRANSLATION) + + +/** 3D (non-perspective) matrix flags mask */ +#define MAT_FLAGS_3D (MAT_FLAG_ROTATION | \ + MAT_FLAG_TRANSLATION | \ + MAT_FLAG_UNIFORM_SCALE | \ + MAT_FLAG_GENERAL_SCALE | \ + MAT_FLAG_GENERAL_3D) + +/** dirty matrix flags mask */ +#define MAT_DIRTY (MAT_DIRTY_TYPE | \ + MAT_DIRTY_FLAGS | \ + MAT_DIRTY_INVERSE) + +/*@}*/ + + +/** + * Test geometry related matrix flags. + * + * \param mat a pointer to a GLmatrix structure. + * \param a flags mask. + * + * \returns non-zero if all geometry related matrix flags are contained within + * the mask, or zero otherwise. + */ +#define TEST_MAT_FLAGS(mat, a) \ + ((MAT_FLAGS_GEOMETRY & (~(a)) & ((mat)->flags) ) == 0) + + + +/** + * Names of the corresponding GLmatrixtype values. + */ +static const char *types[] = { + "MATRIX_GENERAL", + "MATRIX_IDENTITY", + "MATRIX_3D_NO_ROT", + "MATRIX_PERSPECTIVE", + "MATRIX_2D", + "MATRIX_2D_NO_ROT", + "MATRIX_3D" +}; + + +/** + * Identity matrix. + */ +static GLfloat Identity[16] = { + 1.0, 0.0, 0.0, 0.0, + 0.0, 1.0, 0.0, 0.0, + 0.0, 0.0, 1.0, 0.0, + 0.0, 0.0, 0.0, 1.0 +}; + + + +/**********************************************************************/ +/** \name Matrix multiplication */ +/*@{*/ + +#define A(row,col) a[(col<<2)+row] +#define B(row,col) b[(col<<2)+row] +#define P(row,col) product[(col<<2)+row] + +/** + * Perform a full 4x4 matrix multiplication. + * + * \param a matrix. + * \param b matrix. + * \param product will receive the product of \p a and \p b. + * + * \warning Is assumed that \p product != \p b. \p product == \p a is allowed. + * + * \note KW: 4*16 = 64 multiplications + * + * \author This \c matmul was contributed by Thomas Malik + */ +static void matmul4( GLfloat *product, const GLfloat *a, const GLfloat *b ) +{ + GLint i; + for (i = 0; i < 4; i++) { + const GLfloat ai0=A(i,0), ai1=A(i,1), ai2=A(i,2), ai3=A(i,3); + P(i,0) = ai0 * B(0,0) + ai1 * B(1,0) + ai2 * B(2,0) + ai3 * B(3,0); + P(i,1) = ai0 * B(0,1) + ai1 * B(1,1) + ai2 * B(2,1) + ai3 * B(3,1); + P(i,2) = ai0 * B(0,2) + ai1 * B(1,2) + ai2 * B(2,2) + ai3 * B(3,2); + P(i,3) = ai0 * B(0,3) + ai1 * B(1,3) + ai2 * B(2,3) + ai3 * B(3,3); + } +} + +/** + * Multiply two matrices known to occupy only the top three rows, such + * as typical model matrices, and orthogonal matrices. + * + * \param a matrix. + * \param b matrix. + * \param product will receive the product of \p a and \p b. + */ +static void matmul34( GLfloat *product, const GLfloat *a, const GLfloat *b ) +{ + GLint i; + for (i = 0; i < 3; i++) { + const GLfloat ai0=A(i,0), ai1=A(i,1), ai2=A(i,2), ai3=A(i,3); + P(i,0) = ai0 * B(0,0) + ai1 * B(1,0) + ai2 * B(2,0); + P(i,1) = ai0 * B(0,1) + ai1 * B(1,1) + ai2 * B(2,1); + P(i,2) = ai0 * B(0,2) + ai1 * B(1,2) + ai2 * B(2,2); + P(i,3) = ai0 * B(0,3) + ai1 * B(1,3) + ai2 * B(2,3) + ai3; + } + P(3,0) = 0; + P(3,1) = 0; + P(3,2) = 0; + P(3,3) = 1; +} + +#undef A +#undef B +#undef P + +/** + * Multiply a matrix by an array of floats with known properties. + * + * \param mat pointer to a GLmatrix structure containing the left multiplication + * matrix, and that will receive the product result. + * \param m right multiplication matrix array. + * \param flags flags of the matrix \p m. + * + * Joins both flags and marks the type and inverse as dirty. Calls matmul34() + * if both matrices are 3D, or matmul4() otherwise. + */ +static void matrix_multf( GLmatrix *mat, const GLfloat *m, GLuint flags ) +{ + mat->flags |= (flags | MAT_DIRTY_TYPE | MAT_DIRTY_INVERSE); + + if (TEST_MAT_FLAGS(mat, MAT_FLAGS_3D)) + matmul34( mat->m, mat->m, m ); + else + matmul4( mat->m, mat->m, m ); +} + +/** + * Matrix multiplication. + * + * \param dest destination matrix. + * \param a left matrix. + * \param b right matrix. + * + * Joins both flags and marks the type and inverse as dirty. Calls matmul34() + * if both matrices are 3D, or matmul4() otherwise. + */ +void +_math_matrix_mul_matrix( GLmatrix *dest, const GLmatrix *a, const GLmatrix *b ) +{ + dest->flags = (a->flags | + b->flags | + MAT_DIRTY_TYPE | + MAT_DIRTY_INVERSE); + + if (TEST_MAT_FLAGS(dest, MAT_FLAGS_3D)) + matmul34( dest->m, a->m, b->m ); + else + matmul4( dest->m, a->m, b->m ); +} + +/** + * Matrix multiplication. + * + * \param dest left and destination matrix. + * \param m right matrix array. + * + * Marks the matrix flags with general flag, and type and inverse dirty flags. + * Calls matmul4() for the multiplication. + */ +void +_math_matrix_mul_floats( GLmatrix *dest, const GLfloat *m ) +{ + dest->flags |= (MAT_FLAG_GENERAL | + MAT_DIRTY_TYPE | + MAT_DIRTY_INVERSE | + MAT_DIRTY_FLAGS); + + matmul4( dest->m, dest->m, m ); +} + +/*@}*/ + + +/**********************************************************************/ +/** \name Matrix output */ +/*@{*/ + +/** + * Print a matrix array. + * + * \param m matrix array. + * + * Called by _math_matrix_print() to print a matrix or its inverse. + */ +static void print_matrix_floats( const GLfloat m[16] ) +{ + int i; + for (i=0;i<4;i++) { + _mesa_debug(NULL,"\t%f %f %f %f\n", m[i], m[4+i], m[8+i], m[12+i] ); + } +} + +/** + * Dumps the contents of a GLmatrix structure. + * + * \param m pointer to the GLmatrix structure. + */ +void +_math_matrix_print( const GLmatrix *m ) +{ + _mesa_debug(NULL, "Matrix type: %s, flags: %x\n", types[m->type], m->flags); + print_matrix_floats(m->m); + _mesa_debug(NULL, "Inverse: \n"); + if (m->inv) { + GLfloat prod[16]; + print_matrix_floats(m->inv); + matmul4(prod, m->m, m->inv); + _mesa_debug(NULL, "Mat * Inverse:\n"); + print_matrix_floats(prod); + } + else { + _mesa_debug(NULL, " - not available\n"); + } +} + +/*@}*/ + + +/** + * References an element of 4x4 matrix. + * + * \param m matrix array. + * \param c column of the desired element. + * \param r row of the desired element. + * + * \return value of the desired element. + * + * Calculate the linear storage index of the element and references it. + */ +#define MAT(m,r,c) (m)[(c)*4+(r)] + + +/**********************************************************************/ +/** \name Matrix inversion */ +/*@{*/ + +/** + * Swaps the values of two floating pointer variables. + * + * Used by invert_matrix_general() to swap the row pointers. + */ +#define SWAP_ROWS(a, b) { GLfloat *_tmp = a; (a)=(b); (b)=_tmp; } + +/** + * Compute inverse of 4x4 transformation matrix. + * + * \param mat pointer to a GLmatrix structure. The matrix inverse will be + * stored in the GLmatrix::inv attribute. + * + * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). + * + * \author + * Code contributed by Jacques Leroy jle@star.be + * + * Calculates the inverse matrix by performing the gaussian matrix reduction + * with partial pivoting followed by back/substitution with the loops manually + * unrolled. + */ +static GLboolean invert_matrix_general( GLmatrix *mat ) +{ + const GLfloat *m = mat->m; + GLfloat *out = mat->inv; + GLfloat wtmp[4][8]; + GLfloat m0, m1, m2, m3, s; + GLfloat *r0, *r1, *r2, *r3; + + r0 = wtmp[0], r1 = wtmp[1], r2 = wtmp[2], r3 = wtmp[3]; + + r0[0] = MAT(m,0,0), r0[1] = MAT(m,0,1), + r0[2] = MAT(m,0,2), r0[3] = MAT(m,0,3), + r0[4] = 1.0, r0[5] = r0[6] = r0[7] = 0.0, + + r1[0] = MAT(m,1,0), r1[1] = MAT(m,1,1), + r1[2] = MAT(m,1,2), r1[3] = MAT(m,1,3), + r1[5] = 1.0, r1[4] = r1[6] = r1[7] = 0.0, + + r2[0] = MAT(m,2,0), r2[1] = MAT(m,2,1), + r2[2] = MAT(m,2,2), r2[3] = MAT(m,2,3), + r2[6] = 1.0, r2[4] = r2[5] = r2[7] = 0.0, + + r3[0] = MAT(m,3,0), r3[1] = MAT(m,3,1), + r3[2] = MAT(m,3,2), r3[3] = MAT(m,3,3), + r3[7] = 1.0, r3[4] = r3[5] = r3[6] = 0.0; + + /* choose pivot - or die */ + if (FABSF(r3[0])>FABSF(r2[0])) SWAP_ROWS(r3, r2); + if (FABSF(r2[0])>FABSF(r1[0])) SWAP_ROWS(r2, r1); + if (FABSF(r1[0])>FABSF(r0[0])) SWAP_ROWS(r1, r0); + if (0.0 == r0[0]) return GL_FALSE; + + /* eliminate first variable */ + m1 = r1[0]/r0[0]; m2 = r2[0]/r0[0]; m3 = r3[0]/r0[0]; + s = r0[1]; r1[1] -= m1 * s; r2[1] -= m2 * s; r3[1] -= m3 * s; + s = r0[2]; r1[2] -= m1 * s; r2[2] -= m2 * s; r3[2] -= m3 * s; + s = r0[3]; r1[3] -= m1 * s; r2[3] -= m2 * s; r3[3] -= m3 * s; + s = r0[4]; + if (s != 0.0) { r1[4] -= m1 * s; r2[4] -= m2 * s; r3[4] -= m3 * s; } + s = r0[5]; + if (s != 0.0) { r1[5] -= m1 * s; r2[5] -= m2 * s; r3[5] -= m3 * s; } + s = r0[6]; + if (s != 0.0) { r1[6] -= m1 * s; r2[6] -= m2 * s; r3[6] -= m3 * s; } + s = r0[7]; + if (s != 0.0) { r1[7] -= m1 * s; r2[7] -= m2 * s; r3[7] -= m3 * s; } + + /* choose pivot - or die */ + if (FABSF(r3[1])>FABSF(r2[1])) SWAP_ROWS(r3, r2); + if (FABSF(r2[1])>FABSF(r1[1])) SWAP_ROWS(r2, r1); + if (0.0 == r1[1]) return GL_FALSE; + + /* eliminate second variable */ + m2 = r2[1]/r1[1]; m3 = r3[1]/r1[1]; + r2[2] -= m2 * r1[2]; r3[2] -= m3 * r1[2]; + r2[3] -= m2 * r1[3]; r3[3] -= m3 * r1[3]; + s = r1[4]; if (0.0 != s) { r2[4] -= m2 * s; r3[4] -= m3 * s; } + s = r1[5]; if (0.0 != s) { r2[5] -= m2 * s; r3[5] -= m3 * s; } + s = r1[6]; if (0.0 != s) { r2[6] -= m2 * s; r3[6] -= m3 * s; } + s = r1[7]; if (0.0 != s) { r2[7] -= m2 * s; r3[7] -= m3 * s; } + + /* choose pivot - or die */ + if (FABSF(r3[2])>FABSF(r2[2])) SWAP_ROWS(r3, r2); + if (0.0 == r2[2]) return GL_FALSE; + + /* eliminate third variable */ + m3 = r3[2]/r2[2]; + r3[3] -= m3 * r2[3], r3[4] -= m3 * r2[4], + r3[5] -= m3 * r2[5], r3[6] -= m3 * r2[6], + r3[7] -= m3 * r2[7]; + + /* last check */ + if (0.0 == r3[3]) return GL_FALSE; + + s = 1.0F/r3[3]; /* now back substitute row 3 */ + r3[4] *= s; r3[5] *= s; r3[6] *= s; r3[7] *= s; + + m2 = r2[3]; /* now back substitute row 2 */ + s = 1.0F/r2[2]; + r2[4] = s * (r2[4] - r3[4] * m2), r2[5] = s * (r2[5] - r3[5] * m2), + r2[6] = s * (r2[6] - r3[6] * m2), r2[7] = s * (r2[7] - r3[7] * m2); + m1 = r1[3]; + r1[4] -= r3[4] * m1, r1[5] -= r3[5] * m1, + r1[6] -= r3[6] * m1, r1[7] -= r3[7] * m1; + m0 = r0[3]; + r0[4] -= r3[4] * m0, r0[5] -= r3[5] * m0, + r0[6] -= r3[6] * m0, r0[7] -= r3[7] * m0; + + m1 = r1[2]; /* now back substitute row 1 */ + s = 1.0F/r1[1]; + r1[4] = s * (r1[4] - r2[4] * m1), r1[5] = s * (r1[5] - r2[5] * m1), + r1[6] = s * (r1[6] - r2[6] * m1), r1[7] = s * (r1[7] - r2[7] * m1); + m0 = r0[2]; + r0[4] -= r2[4] * m0, r0[5] -= r2[5] * m0, + r0[6] -= r2[6] * m0, r0[7] -= r2[7] * m0; + + m0 = r0[1]; /* now back substitute row 0 */ + s = 1.0F/r0[0]; + r0[4] = s * (r0[4] - r1[4] * m0), r0[5] = s * (r0[5] - r1[5] * m0), + r0[6] = s * (r0[6] - r1[6] * m0), r0[7] = s * (r0[7] - r1[7] * m0); + + MAT(out,0,0) = r0[4]; MAT(out,0,1) = r0[5], + MAT(out,0,2) = r0[6]; MAT(out,0,3) = r0[7], + MAT(out,1,0) = r1[4]; MAT(out,1,1) = r1[5], + MAT(out,1,2) = r1[6]; MAT(out,1,3) = r1[7], + MAT(out,2,0) = r2[4]; MAT(out,2,1) = r2[5], + MAT(out,2,2) = r2[6]; MAT(out,2,3) = r2[7], + MAT(out,3,0) = r3[4]; MAT(out,3,1) = r3[5], + MAT(out,3,2) = r3[6]; MAT(out,3,3) = r3[7]; + + return GL_TRUE; +} +#undef SWAP_ROWS + +/** + * Compute inverse of a general 3d transformation matrix. + * + * \param mat pointer to a GLmatrix structure. The matrix inverse will be + * stored in the GLmatrix::inv attribute. + * + * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). + * + * \author Adapted from graphics gems II. + * + * Calculates the inverse of the upper left by first calculating its + * determinant and multiplying it to the symmetric adjust matrix of each + * element. Finally deals with the translation part by transforming the + * original translation vector using by the calculated submatrix inverse. + */ +static GLboolean invert_matrix_3d_general( GLmatrix *mat ) +{ + const GLfloat *in = mat->m; + GLfloat *out = mat->inv; + GLfloat pos, neg, t; + GLfloat det; + + /* Calculate the determinant of upper left 3x3 submatrix and + * determine if the matrix is singular. + */ + pos = neg = 0.0; + t = MAT(in,0,0) * MAT(in,1,1) * MAT(in,2,2); + if (t >= 0.0) pos += t; else neg += t; + + t = MAT(in,1,0) * MAT(in,2,1) * MAT(in,0,2); + if (t >= 0.0) pos += t; else neg += t; + + t = MAT(in,2,0) * MAT(in,0,1) * MAT(in,1,2); + if (t >= 0.0) pos += t; else neg += t; + + t = -MAT(in,2,0) * MAT(in,1,1) * MAT(in,0,2); + if (t >= 0.0) pos += t; else neg += t; + + t = -MAT(in,1,0) * MAT(in,0,1) * MAT(in,2,2); + if (t >= 0.0) pos += t; else neg += t; + + t = -MAT(in,0,0) * MAT(in,2,1) * MAT(in,1,2); + if (t >= 0.0) pos += t; else neg += t; + + det = pos + neg; + + if (det*det < 1e-25) + return GL_FALSE; + + det = 1.0F / det; + MAT(out,0,0) = ( (MAT(in,1,1)*MAT(in,2,2) - MAT(in,2,1)*MAT(in,1,2) )*det); + MAT(out,0,1) = (- (MAT(in,0,1)*MAT(in,2,2) - MAT(in,2,1)*MAT(in,0,2) )*det); + MAT(out,0,2) = ( (MAT(in,0,1)*MAT(in,1,2) - MAT(in,1,1)*MAT(in,0,2) )*det); + MAT(out,1,0) = (- (MAT(in,1,0)*MAT(in,2,2) - MAT(in,2,0)*MAT(in,1,2) )*det); + MAT(out,1,1) = ( (MAT(in,0,0)*MAT(in,2,2) - MAT(in,2,0)*MAT(in,0,2) )*det); + MAT(out,1,2) = (- (MAT(in,0,0)*MAT(in,1,2) - MAT(in,1,0)*MAT(in,0,2) )*det); + MAT(out,2,0) = ( (MAT(in,1,0)*MAT(in,2,1) - MAT(in,2,0)*MAT(in,1,1) )*det); + MAT(out,2,1) = (- (MAT(in,0,0)*MAT(in,2,1) - MAT(in,2,0)*MAT(in,0,1) )*det); + MAT(out,2,2) = ( (MAT(in,0,0)*MAT(in,1,1) - MAT(in,1,0)*MAT(in,0,1) )*det); + + /* Do the translation part */ + MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0) + + MAT(in,1,3) * MAT(out,0,1) + + MAT(in,2,3) * MAT(out,0,2) ); + MAT(out,1,3) = - (MAT(in,0,3) * MAT(out,1,0) + + MAT(in,1,3) * MAT(out,1,1) + + MAT(in,2,3) * MAT(out,1,2) ); + MAT(out,2,3) = - (MAT(in,0,3) * MAT(out,2,0) + + MAT(in,1,3) * MAT(out,2,1) + + MAT(in,2,3) * MAT(out,2,2) ); + + return GL_TRUE; +} + +/** + * Compute inverse of a 3d transformation matrix. + * + * \param mat pointer to a GLmatrix structure. The matrix inverse will be + * stored in the GLmatrix::inv attribute. + * + * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). + * + * If the matrix is not an angle preserving matrix then calls + * invert_matrix_3d_general for the actual calculation. Otherwise calculates + * the inverse matrix analyzing and inverting each of the scaling, rotation and + * translation parts. + */ +static GLboolean invert_matrix_3d( GLmatrix *mat ) +{ + const GLfloat *in = mat->m; + GLfloat *out = mat->inv; + + if (!TEST_MAT_FLAGS(mat, MAT_FLAGS_ANGLE_PRESERVING)) { + return invert_matrix_3d_general( mat ); + } + + if (mat->flags & MAT_FLAG_UNIFORM_SCALE) { + GLfloat scale = (MAT(in,0,0) * MAT(in,0,0) + + MAT(in,0,1) * MAT(in,0,1) + + MAT(in,0,2) * MAT(in,0,2)); + + if (scale == 0.0) + return GL_FALSE; + + scale = 1.0F / scale; + + /* Transpose and scale the 3 by 3 upper-left submatrix. */ + MAT(out,0,0) = scale * MAT(in,0,0); + MAT(out,1,0) = scale * MAT(in,0,1); + MAT(out,2,0) = scale * MAT(in,0,2); + MAT(out,0,1) = scale * MAT(in,1,0); + MAT(out,1,1) = scale * MAT(in,1,1); + MAT(out,2,1) = scale * MAT(in,1,2); + MAT(out,0,2) = scale * MAT(in,2,0); + MAT(out,1,2) = scale * MAT(in,2,1); + MAT(out,2,2) = scale * MAT(in,2,2); + } + else if (mat->flags & MAT_FLAG_ROTATION) { + /* Transpose the 3 by 3 upper-left submatrix. */ + MAT(out,0,0) = MAT(in,0,0); + MAT(out,1,0) = MAT(in,0,1); + MAT(out,2,0) = MAT(in,0,2); + MAT(out,0,1) = MAT(in,1,0); + MAT(out,1,1) = MAT(in,1,1); + MAT(out,2,1) = MAT(in,1,2); + MAT(out,0,2) = MAT(in,2,0); + MAT(out,1,2) = MAT(in,2,1); + MAT(out,2,2) = MAT(in,2,2); + } + else { + /* pure translation */ + memcpy( out, Identity, sizeof(Identity) ); + MAT(out,0,3) = - MAT(in,0,3); + MAT(out,1,3) = - MAT(in,1,3); + MAT(out,2,3) = - MAT(in,2,3); + return GL_TRUE; + } + + if (mat->flags & MAT_FLAG_TRANSLATION) { + /* Do the translation part */ + MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0) + + MAT(in,1,3) * MAT(out,0,1) + + MAT(in,2,3) * MAT(out,0,2) ); + MAT(out,1,3) = - (MAT(in,0,3) * MAT(out,1,0) + + MAT(in,1,3) * MAT(out,1,1) + + MAT(in,2,3) * MAT(out,1,2) ); + MAT(out,2,3) = - (MAT(in,0,3) * MAT(out,2,0) + + MAT(in,1,3) * MAT(out,2,1) + + MAT(in,2,3) * MAT(out,2,2) ); + } + else { + MAT(out,0,3) = MAT(out,1,3) = MAT(out,2,3) = 0.0; + } + + return GL_TRUE; +} + +/** + * Compute inverse of an identity transformation matrix. + * + * \param mat pointer to a GLmatrix structure. The matrix inverse will be + * stored in the GLmatrix::inv attribute. + * + * \return always GL_TRUE. + * + * Simply copies Identity into GLmatrix::inv. + */ +static GLboolean invert_matrix_identity( GLmatrix *mat ) +{ + memcpy( mat->inv, Identity, sizeof(Identity) ); + return GL_TRUE; +} + +/** + * Compute inverse of a no-rotation 3d transformation matrix. + * + * \param mat pointer to a GLmatrix structure. The matrix inverse will be + * stored in the GLmatrix::inv attribute. + * + * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). + * + * Calculates the + */ +static GLboolean invert_matrix_3d_no_rot( GLmatrix *mat ) +{ + const GLfloat *in = mat->m; + GLfloat *out = mat->inv; + + if (MAT(in,0,0) == 0 || MAT(in,1,1) == 0 || MAT(in,2,2) == 0 ) + return GL_FALSE; + + memcpy( out, Identity, 16 * sizeof(GLfloat) ); + MAT(out,0,0) = 1.0F / MAT(in,0,0); + MAT(out,1,1) = 1.0F / MAT(in,1,1); + MAT(out,2,2) = 1.0F / MAT(in,2,2); + + if (mat->flags & MAT_FLAG_TRANSLATION) { + MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0)); + MAT(out,1,3) = - (MAT(in,1,3) * MAT(out,1,1)); + MAT(out,2,3) = - (MAT(in,2,3) * MAT(out,2,2)); + } + + return GL_TRUE; +} + +/** + * Compute inverse of a no-rotation 2d transformation matrix. + * + * \param mat pointer to a GLmatrix structure. The matrix inverse will be + * stored in the GLmatrix::inv attribute. + * + * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). + * + * Calculates the inverse matrix by applying the inverse scaling and + * translation to the identity matrix. + */ +static GLboolean invert_matrix_2d_no_rot( GLmatrix *mat ) +{ + const GLfloat *in = mat->m; + GLfloat *out = mat->inv; + + if (MAT(in,0,0) == 0 || MAT(in,1,1) == 0) + return GL_FALSE; + + memcpy( out, Identity, 16 * sizeof(GLfloat) ); + MAT(out,0,0) = 1.0F / MAT(in,0,0); + MAT(out,1,1) = 1.0F / MAT(in,1,1); + + if (mat->flags & MAT_FLAG_TRANSLATION) { + MAT(out,0,3) = - (MAT(in,0,3) * MAT(out,0,0)); + MAT(out,1,3) = - (MAT(in,1,3) * MAT(out,1,1)); + } + + return GL_TRUE; +} + +#if 0 +/* broken */ +static GLboolean invert_matrix_perspective( GLmatrix *mat ) +{ + const GLfloat *in = mat->m; + GLfloat *out = mat->inv; + + if (MAT(in,2,3) == 0) + return GL_FALSE; + + memcpy( out, Identity, 16 * sizeof(GLfloat) ); + + MAT(out,0,0) = 1.0F / MAT(in,0,0); + MAT(out,1,1) = 1.0F / MAT(in,1,1); + + MAT(out,0,3) = MAT(in,0,2); + MAT(out,1,3) = MAT(in,1,2); + + MAT(out,2,2) = 0; + MAT(out,2,3) = -1; + + MAT(out,3,2) = 1.0F / MAT(in,2,3); + MAT(out,3,3) = MAT(in,2,2) * MAT(out,3,2); + + return GL_TRUE; +} +#endif + +/** + * Matrix inversion function pointer type. + */ +typedef GLboolean (*inv_mat_func)( GLmatrix *mat ); + +/** + * Table of the matrix inversion functions according to the matrix type. + */ +static inv_mat_func inv_mat_tab[7] = { + invert_matrix_general, + invert_matrix_identity, + invert_matrix_3d_no_rot, +#if 0 + /* Don't use this function for now - it fails when the projection matrix + * is premultiplied by a translation (ala Chromium's tilesort SPU). + */ + invert_matrix_perspective, +#else + invert_matrix_general, +#endif + invert_matrix_3d, /* lazy! */ + invert_matrix_2d_no_rot, + invert_matrix_3d +}; + +/** + * Compute inverse of a transformation matrix. + * + * \param mat pointer to a GLmatrix structure. The matrix inverse will be + * stored in the GLmatrix::inv attribute. + * + * \return GL_TRUE for success, GL_FALSE for failure (\p singular matrix). + * + * Calls the matrix inversion function in inv_mat_tab corresponding to the + * given matrix type. In case of failure, updates the MAT_FLAG_SINGULAR flag, + * and copies the identity matrix into GLmatrix::inv. + */ +static GLboolean matrix_invert( GLmatrix *mat ) +{ + if (inv_mat_tab[mat->type](mat)) { + mat->flags &= ~MAT_FLAG_SINGULAR; + return GL_TRUE; + } else { + mat->flags |= MAT_FLAG_SINGULAR; + memcpy( mat->inv, Identity, sizeof(Identity) ); + return GL_FALSE; + } +} + +/*@}*/ + + +/**********************************************************************/ +/** \name Matrix generation */ +/*@{*/ + +/** + * Generate a 4x4 transformation matrix from glRotate parameters, and + * post-multiply the input matrix by it. + * + * \author + * This function was contributed by Erich Boleyn (erich@uruk.org). + * Optimizations contributed by Rudolf Opalla (rudi@khm.de). + */ +void +_math_matrix_rotate( GLmatrix *mat, + GLfloat angle, GLfloat x, GLfloat y, GLfloat z ) +{ + GLfloat xx, yy, zz, xy, yz, zx, xs, ys, zs, one_c, s, c; + GLfloat m[16]; + GLboolean optimized; + + s = (GLfloat) sin( angle * DEG2RAD ); + c = (GLfloat) cos( angle * DEG2RAD ); + + memcpy(m, Identity, sizeof(GLfloat)*16); + optimized = GL_FALSE; + +#define M(row,col) m[col*4+row] + + if (x == 0.0F) { + if (y == 0.0F) { + if (z != 0.0F) { + optimized = GL_TRUE; + /* rotate only around z-axis */ + M(0,0) = c; + M(1,1) = c; + if (z < 0.0F) { + M(0,1) = s; + M(1,0) = -s; + } + else { + M(0,1) = -s; + M(1,0) = s; + } + } + } + else if (z == 0.0F) { + optimized = GL_TRUE; + /* rotate only around y-axis */ + M(0,0) = c; + M(2,2) = c; + if (y < 0.0F) { + M(0,2) = -s; + M(2,0) = s; + } + else { + M(0,2) = s; + M(2,0) = -s; + } + } + } + else if (y == 0.0F) { + if (z == 0.0F) { + optimized = GL_TRUE; + /* rotate only around x-axis */ + M(1,1) = c; + M(2,2) = c; + if (x < 0.0F) { + M(1,2) = s; + M(2,1) = -s; + } + else { + M(1,2) = -s; + M(2,1) = s; + } + } + } + + if (!optimized) { + const GLfloat mag = SQRTF(x * x + y * y + z * z); + + if (mag <= 1.0e-4) { + /* no rotation, leave mat as-is */ + return; + } + + x /= mag; + y /= mag; + z /= mag; + + + /* + * Arbitrary axis rotation matrix. + * + * This is composed of 5 matrices, Rz, Ry, T, Ry', Rz', multiplied + * like so: Rz * Ry * T * Ry' * Rz'. T is the final rotation + * (which is about the X-axis), and the two composite transforms + * Ry' * Rz' and Rz * Ry are (respectively) the rotations necessary + * from the arbitrary axis to the X-axis then back. They are + * all elementary rotations. + * + * Rz' is a rotation about the Z-axis, to bring the axis vector + * into the x-z plane. Then Ry' is applied, rotating about the + * Y-axis to bring the axis vector parallel with the X-axis. The + * rotation about the X-axis is then performed. Ry and Rz are + * simply the respective inverse transforms to bring the arbitrary + * axis back to its original orientation. The first transforms + * Rz' and Ry' are considered inverses, since the data from the + * arbitrary axis gives you info on how to get to it, not how + * to get away from it, and an inverse must be applied. + * + * The basic calculation used is to recognize that the arbitrary + * axis vector (x, y, z), since it is of unit length, actually + * represents the sines and cosines of the angles to rotate the + * X-axis to the same orientation, with theta being the angle about + * Z and phi the angle about Y (in the order described above) + * as follows: + * + * cos ( theta ) = x / sqrt ( 1 - z^2 ) + * sin ( theta ) = y / sqrt ( 1 - z^2 ) + * + * cos ( phi ) = sqrt ( 1 - z^2 ) + * sin ( phi ) = z + * + * Note that cos ( phi ) can further be inserted to the above + * formulas: + * + * cos ( theta ) = x / cos ( phi ) + * sin ( theta ) = y / sin ( phi ) + * + * ...etc. Because of those relations and the standard trigonometric + * relations, it is pssible to reduce the transforms down to what + * is used below. It may be that any primary axis chosen will give the + * same results (modulo a sign convention) using thie method. + * + * Particularly nice is to notice that all divisions that might + * have caused trouble when parallel to certain planes or + * axis go away with care paid to reducing the expressions. + * After checking, it does perform correctly under all cases, since + * in all the cases of division where the denominator would have + * been zero, the numerator would have been zero as well, giving + * the expected result. + */ + + xx = x * x; + yy = y * y; + zz = z * z; + xy = x * y; + yz = y * z; + zx = z * x; + xs = x * s; + ys = y * s; + zs = z * s; + one_c = 1.0F - c; + + /* We already hold the identity-matrix so we can skip some statements */ + M(0,0) = (one_c * xx) + c; + M(0,1) = (one_c * xy) - zs; + M(0,2) = (one_c * zx) + ys; +/* M(0,3) = 0.0F; */ + + M(1,0) = (one_c * xy) + zs; + M(1,1) = (one_c * yy) + c; + M(1,2) = (one_c * yz) - xs; +/* M(1,3) = 0.0F; */ + + M(2,0) = (one_c * zx) - ys; + M(2,1) = (one_c * yz) + xs; + M(2,2) = (one_c * zz) + c; +/* M(2,3) = 0.0F; */ + +/* + M(3,0) = 0.0F; + M(3,1) = 0.0F; + M(3,2) = 0.0F; + M(3,3) = 1.0F; +*/ + } +#undef M + + matrix_multf( mat, m, MAT_FLAG_ROTATION ); +} + +/** + * Apply a perspective projection matrix. + * + * \param mat matrix to apply the projection. + * \param left left clipping plane coordinate. + * \param right right clipping plane coordinate. + * \param bottom bottom clipping plane coordinate. + * \param top top clipping plane coordinate. + * \param nearval distance to the near clipping plane. + * \param farval distance to the far clipping plane. + * + * Creates the projection matrix and multiplies it with \p mat, marking the + * MAT_FLAG_PERSPECTIVE flag. + */ +void +_math_matrix_frustum( GLmatrix *mat, + GLfloat left, GLfloat right, + GLfloat bottom, GLfloat top, + GLfloat nearval, GLfloat farval ) +{ + GLfloat x, y, a, b, c, d; + GLfloat m[16]; + + x = (2.0F*nearval) / (right-left); + y = (2.0F*nearval) / (top-bottom); + a = (right+left) / (right-left); + b = (top+bottom) / (top-bottom); + c = -(farval+nearval) / ( farval-nearval); + d = -(2.0F*farval*nearval) / (farval-nearval); /* error? */ + +#define M(row,col) m[col*4+row] + M(0,0) = x; M(0,1) = 0.0F; M(0,2) = a; M(0,3) = 0.0F; + M(1,0) = 0.0F; M(1,1) = y; M(1,2) = b; M(1,3) = 0.0F; + M(2,0) = 0.0F; M(2,1) = 0.0F; M(2,2) = c; M(2,3) = d; + M(3,0) = 0.0F; M(3,1) = 0.0F; M(3,2) = -1.0F; M(3,3) = 0.0F; +#undef M + + matrix_multf( mat, m, MAT_FLAG_PERSPECTIVE ); +} + +/** + * Apply an orthographic projection matrix. + * + * \param mat matrix to apply the projection. + * \param left left clipping plane coordinate. + * \param right right clipping plane coordinate. + * \param bottom bottom clipping plane coordinate. + * \param top top clipping plane coordinate. + * \param nearval distance to the near clipping plane. + * \param farval distance to the far clipping plane. + * + * Creates the projection matrix and multiplies it with \p mat, marking the + * MAT_FLAG_GENERAL_SCALE and MAT_FLAG_TRANSLATION flags. + */ +void +_math_matrix_ortho( GLmatrix *mat, + GLfloat left, GLfloat right, + GLfloat bottom, GLfloat top, + GLfloat nearval, GLfloat farval ) +{ + GLfloat m[16]; + +#define M(row,col) m[col*4+row] + M(0,0) = 2.0F / (right-left); + M(0,1) = 0.0F; + M(0,2) = 0.0F; + M(0,3) = -(right+left) / (right-left); + + M(1,0) = 0.0F; + M(1,1) = 2.0F / (top-bottom); + M(1,2) = 0.0F; + M(1,3) = -(top+bottom) / (top-bottom); + + M(2,0) = 0.0F; + M(2,1) = 0.0F; + M(2,2) = -2.0F / (farval-nearval); + M(2,3) = -(farval+nearval) / (farval-nearval); + + M(3,0) = 0.0F; + M(3,1) = 0.0F; + M(3,2) = 0.0F; + M(3,3) = 1.0F; +#undef M + + matrix_multf( mat, m, (MAT_FLAG_GENERAL_SCALE|MAT_FLAG_TRANSLATION)); +} + +/** + * Multiply a matrix with a general scaling matrix. + * + * \param mat matrix. + * \param x x axis scale factor. + * \param y y axis scale factor. + * \param z z axis scale factor. + * + * Multiplies in-place the elements of \p mat by the scale factors. Checks if + * the scales factors are roughly the same, marking the MAT_FLAG_UNIFORM_SCALE + * flag, or MAT_FLAG_GENERAL_SCALE. Marks the MAT_DIRTY_TYPE and + * MAT_DIRTY_INVERSE dirty flags. + */ +void +_math_matrix_scale( GLmatrix *mat, GLfloat x, GLfloat y, GLfloat z ) +{ + GLfloat *m = mat->m; + m[0] *= x; m[4] *= y; m[8] *= z; + m[1] *= x; m[5] *= y; m[9] *= z; + m[2] *= x; m[6] *= y; m[10] *= z; + m[3] *= x; m[7] *= y; m[11] *= z; + + if (FABSF(x - y) < 1e-8 && FABSF(x - z) < 1e-8) + mat->flags |= MAT_FLAG_UNIFORM_SCALE; + else + mat->flags |= MAT_FLAG_GENERAL_SCALE; + + mat->flags |= (MAT_DIRTY_TYPE | + MAT_DIRTY_INVERSE); +} + +/** + * Multiply a matrix with a translation matrix. + * + * \param mat matrix. + * \param x translation vector x coordinate. + * \param y translation vector y coordinate. + * \param z translation vector z coordinate. + * + * Adds the translation coordinates to the elements of \p mat in-place. Marks + * the MAT_FLAG_TRANSLATION flag, and the MAT_DIRTY_TYPE and MAT_DIRTY_INVERSE + * dirty flags. + */ +void +_math_matrix_translate( GLmatrix *mat, GLfloat x, GLfloat y, GLfloat z ) +{ + GLfloat *m = mat->m; + m[12] = m[0] * x + m[4] * y + m[8] * z + m[12]; + m[13] = m[1] * x + m[5] * y + m[9] * z + m[13]; + m[14] = m[2] * x + m[6] * y + m[10] * z + m[14]; + m[15] = m[3] * x + m[7] * y + m[11] * z + m[15]; + + mat->flags |= (MAT_FLAG_TRANSLATION | + MAT_DIRTY_TYPE | + MAT_DIRTY_INVERSE); +} + + +/** + * Set matrix to do viewport and depthrange mapping. + * Transforms Normalized Device Coords to window/Z values. + */ +void +_math_matrix_viewport(GLmatrix *m, GLint x, GLint y, GLint width, GLint height, + GLfloat zNear, GLfloat zFar, GLfloat depthMax) +{ + m->m[MAT_SX] = (GLfloat) width / 2.0F; + m->m[MAT_TX] = m->m[MAT_SX] + x; + m->m[MAT_SY] = (GLfloat) height / 2.0F; + m->m[MAT_TY] = m->m[MAT_SY] + y; + m->m[MAT_SZ] = depthMax * ((zFar - zNear) / 2.0F); + m->m[MAT_TZ] = depthMax * ((zFar - zNear) / 2.0F + zNear); + m->flags = MAT_FLAG_GENERAL_SCALE | MAT_FLAG_TRANSLATION; + m->type = MATRIX_3D_NO_ROT; +} + + +/** + * Set a matrix to the identity matrix. + * + * \param mat matrix. + * + * Copies ::Identity into \p GLmatrix::m, and into GLmatrix::inv if not NULL. + * Sets the matrix type to identity, and clear the dirty flags. + */ +void +_math_matrix_set_identity( GLmatrix *mat ) +{ + memcpy( mat->m, Identity, 16*sizeof(GLfloat) ); + + if (mat->inv) + memcpy( mat->inv, Identity, 16*sizeof(GLfloat) ); + + mat->type = MATRIX_IDENTITY; + mat->flags &= ~(MAT_DIRTY_FLAGS| + MAT_DIRTY_TYPE| + MAT_DIRTY_INVERSE); +} + +/*@}*/ + + +/**********************************************************************/ +/** \name Matrix analysis */ +/*@{*/ + +#define ZERO(x) (1<m; + GLuint mask = 0; + GLuint i; + + for (i = 0 ; i < 16 ; i++) { + if (m[i] == 0.0) mask |= (1<flags &= ~MAT_FLAGS_GEOMETRY; + + /* Check for translation - no-one really cares + */ + if ((mask & MASK_NO_TRX) != MASK_NO_TRX) + mat->flags |= MAT_FLAG_TRANSLATION; + + /* Do the real work + */ + if (mask == (GLuint) MASK_IDENTITY) { + mat->type = MATRIX_IDENTITY; + } + else if ((mask & MASK_2D_NO_ROT) == (GLuint) MASK_2D_NO_ROT) { + mat->type = MATRIX_2D_NO_ROT; + + if ((mask & MASK_NO_2D_SCALE) != MASK_NO_2D_SCALE) + mat->flags |= MAT_FLAG_GENERAL_SCALE; + } + else if ((mask & MASK_2D) == (GLuint) MASK_2D) { + GLfloat mm = DOT2(m, m); + GLfloat m4m4 = DOT2(m+4,m+4); + GLfloat mm4 = DOT2(m,m+4); + + mat->type = MATRIX_2D; + + /* Check for scale */ + if (SQ(mm-1) > SQ(1e-6) || + SQ(m4m4-1) > SQ(1e-6)) + mat->flags |= MAT_FLAG_GENERAL_SCALE; + + /* Check for rotation */ + if (SQ(mm4) > SQ(1e-6)) + mat->flags |= MAT_FLAG_GENERAL_3D; + else + mat->flags |= MAT_FLAG_ROTATION; + + } + else if ((mask & MASK_3D_NO_ROT) == (GLuint) MASK_3D_NO_ROT) { + mat->type = MATRIX_3D_NO_ROT; + + /* Check for scale */ + if (SQ(m[0]-m[5]) < SQ(1e-6) && + SQ(m[0]-m[10]) < SQ(1e-6)) { + if (SQ(m[0]-1.0) > SQ(1e-6)) { + mat->flags |= MAT_FLAG_UNIFORM_SCALE; + } + } + else { + mat->flags |= MAT_FLAG_GENERAL_SCALE; + } + } + else if ((mask & MASK_3D) == (GLuint) MASK_3D) { + GLfloat c1 = DOT3(m,m); + GLfloat c2 = DOT3(m+4,m+4); + GLfloat c3 = DOT3(m+8,m+8); + GLfloat d1 = DOT3(m, m+4); + GLfloat cp[3]; + + mat->type = MATRIX_3D; + + /* Check for scale */ + if (SQ(c1-c2) < SQ(1e-6) && SQ(c1-c3) < SQ(1e-6)) { + if (SQ(c1-1.0) > SQ(1e-6)) + mat->flags |= MAT_FLAG_UNIFORM_SCALE; + /* else no scale at all */ + } + else { + mat->flags |= MAT_FLAG_GENERAL_SCALE; + } + + /* Check for rotation */ + if (SQ(d1) < SQ(1e-6)) { + CROSS3( cp, m, m+4 ); + SUB_3V( cp, cp, (m+8) ); + if (LEN_SQUARED_3FV(cp) < SQ(1e-6)) + mat->flags |= MAT_FLAG_ROTATION; + else + mat->flags |= MAT_FLAG_GENERAL_3D; + } + else { + mat->flags |= MAT_FLAG_GENERAL_3D; /* shear, etc */ + } + } + else if ((mask & MASK_PERSPECTIVE) == MASK_PERSPECTIVE && m[11]==-1.0F) { + mat->type = MATRIX_PERSPECTIVE; + mat->flags |= MAT_FLAG_GENERAL; + } + else { + mat->type = MATRIX_GENERAL; + mat->flags |= MAT_FLAG_GENERAL; + } +} + +/** + * Analyze a matrix given that its flags are accurate. + * + * This is the more common operation, hopefully. + */ +static void analyse_from_flags( GLmatrix *mat ) +{ + const GLfloat *m = mat->m; + + if (TEST_MAT_FLAGS(mat, 0)) { + mat->type = MATRIX_IDENTITY; + } + else if (TEST_MAT_FLAGS(mat, (MAT_FLAG_TRANSLATION | + MAT_FLAG_UNIFORM_SCALE | + MAT_FLAG_GENERAL_SCALE))) { + if ( m[10]==1.0F && m[14]==0.0F ) { + mat->type = MATRIX_2D_NO_ROT; + } + else { + mat->type = MATRIX_3D_NO_ROT; + } + } + else if (TEST_MAT_FLAGS(mat, MAT_FLAGS_3D)) { + if ( m[ 8]==0.0F + && m[ 9]==0.0F + && m[2]==0.0F && m[6]==0.0F && m[10]==1.0F && m[14]==0.0F) { + mat->type = MATRIX_2D; + } + else { + mat->type = MATRIX_3D; + } + } + else if ( m[4]==0.0F && m[12]==0.0F + && m[1]==0.0F && m[13]==0.0F + && m[2]==0.0F && m[6]==0.0F + && m[3]==0.0F && m[7]==0.0F && m[11]==-1.0F && m[15]==0.0F) { + mat->type = MATRIX_PERSPECTIVE; + } + else { + mat->type = MATRIX_GENERAL; + } +} + +/** + * Analyze and update a matrix. + * + * \param mat matrix. + * + * If the matrix type is dirty then calls either analyse_from_scratch() or + * analyse_from_flags() to determine its type, according to whether the flags + * are dirty or not, respectively. If the matrix has an inverse and it's dirty + * then calls matrix_invert(). Finally clears the dirty flags. + */ +void +_math_matrix_analyse( GLmatrix *mat ) +{ + if (mat->flags & MAT_DIRTY_TYPE) { + if (mat->flags & MAT_DIRTY_FLAGS) + analyse_from_scratch( mat ); + else + analyse_from_flags( mat ); + } + + if (mat->inv && (mat->flags & MAT_DIRTY_INVERSE)) { + matrix_invert( mat ); + mat->flags &= ~MAT_DIRTY_INVERSE; + } + + mat->flags &= ~(MAT_DIRTY_FLAGS | MAT_DIRTY_TYPE); +} + +/*@}*/ + + +/** + * Test if the given matrix preserves vector lengths. + */ +GLboolean +_math_matrix_is_length_preserving( const GLmatrix *m ) +{ + return TEST_MAT_FLAGS( m, MAT_FLAGS_LENGTH_PRESERVING); +} + + +/** + * Test if the given matrix does any rotation. + * (or perhaps if the upper-left 3x3 is non-identity) + */ +GLboolean +_math_matrix_has_rotation( const GLmatrix *m ) +{ + if (m->flags & (MAT_FLAG_GENERAL | + MAT_FLAG_ROTATION | + MAT_FLAG_GENERAL_3D | + MAT_FLAG_PERSPECTIVE)) + return GL_TRUE; + else + return GL_FALSE; +} + + +GLboolean +_math_matrix_is_general_scale( const GLmatrix *m ) +{ + return (m->flags & MAT_FLAG_GENERAL_SCALE) ? GL_TRUE : GL_FALSE; +} + + +GLboolean +_math_matrix_is_dirty( const GLmatrix *m ) +{ + return (m->flags & MAT_DIRTY) ? GL_TRUE : GL_FALSE; +} + + +/**********************************************************************/ +/** \name Matrix setup */ +/*@{*/ + +/** + * Copy a matrix. + * + * \param to destination matrix. + * \param from source matrix. + * + * Copies all fields in GLmatrix, creating an inverse array if necessary. + */ +void +_math_matrix_copy( GLmatrix *to, const GLmatrix *from ) +{ + memcpy( to->m, from->m, sizeof(Identity) ); + to->flags = from->flags; + to->type = from->type; + + if (to->inv != 0) { + if (from->inv == 0) { + matrix_invert( to ); + } + else { + memcpy(to->inv, from->inv, sizeof(GLfloat)*16); + } + } +} + +/** + * Loads a matrix array into GLmatrix. + * + * \param m matrix array. + * \param mat matrix. + * + * Copies \p m into GLmatrix::m and marks the MAT_FLAG_GENERAL and MAT_DIRTY + * flags. + */ +void +_math_matrix_loadf( GLmatrix *mat, const GLfloat *m ) +{ + memcpy( mat->m, m, 16*sizeof(GLfloat) ); + mat->flags = (MAT_FLAG_GENERAL | MAT_DIRTY); +} + +/** + * Matrix constructor. + * + * \param m matrix. + * + * Initialize the GLmatrix fields. + */ +void +_math_matrix_ctr( GLmatrix *m ) +{ + m->m = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); + if (m->m) + memcpy( m->m, Identity, sizeof(Identity) ); + m->inv = NULL; + m->type = MATRIX_IDENTITY; + m->flags = 0; +} + +/** + * Matrix destructor. + * + * \param m matrix. + * + * Frees the data in a GLmatrix. + */ +void +_math_matrix_dtr( GLmatrix *m ) +{ + if (m->m) { + _mesa_align_free( m->m ); + m->m = NULL; + } + if (m->inv) { + _mesa_align_free( m->inv ); + m->inv = NULL; + } +} + +/** + * Allocate a matrix inverse. + * + * \param m matrix. + * + * Allocates the matrix inverse, GLmatrix::inv, and sets it to Identity. + */ +void +_math_matrix_alloc_inv( GLmatrix *m ) +{ + if (!m->inv) { + m->inv = (GLfloat *) _mesa_align_malloc( 16 * sizeof(GLfloat), 16 ); + if (m->inv) + memcpy( m->inv, Identity, 16 * sizeof(GLfloat) ); + } +} + +/*@}*/ + + +/**********************************************************************/ +/** \name Matrix transpose */ +/*@{*/ + +/** + * Transpose a GLfloat matrix. + * + * \param to destination array. + * \param from source array. + */ +void +_math_transposef( GLfloat to[16], const GLfloat from[16] ) +{ + to[0] = from[0]; + to[1] = from[4]; + to[2] = from[8]; + to[3] = from[12]; + to[4] = from[1]; + to[5] = from[5]; + to[6] = from[9]; + to[7] = from[13]; + to[8] = from[2]; + to[9] = from[6]; + to[10] = from[10]; + to[11] = from[14]; + to[12] = from[3]; + to[13] = from[7]; + to[14] = from[11]; + to[15] = from[15]; +} + +/** + * Transpose a GLdouble matrix. + * + * \param to destination array. + * \param from source array. + */ +void +_math_transposed( GLdouble to[16], const GLdouble from[16] ) +{ + to[0] = from[0]; + to[1] = from[4]; + to[2] = from[8]; + to[3] = from[12]; + to[4] = from[1]; + to[5] = from[5]; + to[6] = from[9]; + to[7] = from[13]; + to[8] = from[2]; + to[9] = from[6]; + to[10] = from[10]; + to[11] = from[14]; + to[12] = from[3]; + to[13] = from[7]; + to[14] = from[11]; + to[15] = from[15]; +} + +/** + * Transpose a GLdouble matrix and convert to GLfloat. + * + * \param to destination array. + * \param from source array. + */ +void +_math_transposefd( GLfloat to[16], const GLdouble from[16] ) +{ + to[0] = (GLfloat) from[0]; + to[1] = (GLfloat) from[4]; + to[2] = (GLfloat) from[8]; + to[3] = (GLfloat) from[12]; + to[4] = (GLfloat) from[1]; + to[5] = (GLfloat) from[5]; + to[6] = (GLfloat) from[9]; + to[7] = (GLfloat) from[13]; + to[8] = (GLfloat) from[2]; + to[9] = (GLfloat) from[6]; + to[10] = (GLfloat) from[10]; + to[11] = (GLfloat) from[14]; + to[12] = (GLfloat) from[3]; + to[13] = (GLfloat) from[7]; + to[14] = (GLfloat) from[11]; + to[15] = (GLfloat) from[15]; +} + +/*@}*/ + + +/** + * Transform a 4-element row vector (1x4 matrix) by a 4x4 matrix. This + * function is used for transforming clipping plane equations and spotlight + * directions. + * Mathematically, u = v * m. + * Input: v - input vector + * m - transformation matrix + * Output: u - transformed vector + */ +void +_mesa_transform_vector( GLfloat u[4], const GLfloat v[4], const GLfloat m[16] ) +{ + const GLfloat v0 = v[0], v1 = v[1], v2 = v[2], v3 = v[3]; +#define M(row,col) m[row + col*4] + u[0] = v0 * M(0,0) + v1 * M(1,0) + v2 * M(2,0) + v3 * M(3,0); + u[1] = v0 * M(0,1) + v1 * M(1,1) + v2 * M(2,1) + v3 * M(3,1); + u[2] = v0 * M(0,2) + v1 * M(1,2) + v2 * M(2,2) + v3 * M(3,2); + u[3] = v0 * M(0,3) + v1 * M(1,3) + v2 * M(2,3) + v3 * M(3,3); +#undef M +} diff --git a/mesalib/src/mesa/math/m_vector.c b/mesalib/src/mesa/math/m_vector.c index 7ca08f4c0..4bded31e0 100644 --- a/mesalib/src/mesa/math/m_vector.c +++ b/mesalib/src/mesa/math/m_vector.c @@ -1,184 +1,184 @@ -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * New (3.1) transformation code written by Keith Whitwell. - */ - - -#include "main/glheader.h" -#include "main/imports.h" -#include "main/macros.h" - -#include "m_vector.h" - - - -/** - * Given a vector [count][4] of floats, set all the [][elt] values - * to 0 (if elt = 0, 1, 2) or 1.0 (if elt = 3). - */ -void -_mesa_vector4f_clean_elem( GLvector4f *vec, GLuint count, GLuint elt ) -{ - static const GLubyte elem_bits[4] = { - VEC_DIRTY_0, - VEC_DIRTY_1, - VEC_DIRTY_2, - VEC_DIRTY_3 - }; - static const GLfloat clean[4] = { 0, 0, 0, 1 }; - const GLfloat v = clean[elt]; - GLfloat (*data)[4] = (GLfloat (*)[4])vec->start; - GLuint i; - - for (i = 0; i < count; i++) - data[i][elt] = v; - - vec->flags &= ~elem_bits[elt]; -} - - -static const GLubyte size_bits[5] = { - 0, - VEC_SIZE_1, - VEC_SIZE_2, - VEC_SIZE_3, - VEC_SIZE_4, -}; - - -/** - * Initialize GLvector objects. - * \param v the vector object to initialize. - * \param flags bitwise-OR of VEC_* flags - * \param storage pointer to storage for the vector's data - */ -void -_mesa_vector4f_init( GLvector4f *v, GLbitfield flags, GLfloat (*storage)[4] ) -{ - v->stride = 4 * sizeof(GLfloat); - v->size = 2; /* may change: 2-4 for vertices and 1-4 for texcoords */ - v->data = storage; - v->start = (GLfloat *) storage; - v->count = 0; - v->flags = size_bits[4] | flags; -} - - -/** - * Initialize GLvector objects and allocate storage. - * \param v the vector object - * \param flags bitwise-OR of VEC_* flags - * \param count number of elements to allocate in vector - * \param alignment desired memory alignment for the data (in bytes) - */ -void -_mesa_vector4f_alloc( GLvector4f *v, GLbitfield flags, GLuint count, - GLuint alignment ) -{ - v->stride = 4 * sizeof(GLfloat); - v->size = 2; - v->storage = _mesa_align_malloc( count * 4 * sizeof(GLfloat), alignment ); - v->storage_count = count; - v->start = (GLfloat *) v->storage; - v->data = (GLfloat (*)[4]) v->storage; - v->count = 0; - v->flags = size_bits[4] | flags | VEC_MALLOC; -} - - -/** - * Vector deallocation. Free whatever memory is pointed to by the - * vector's storage field if the VEC_MALLOC flag is set. - * DO NOT free the GLvector object itself, though. - */ -void -_mesa_vector4f_free( GLvector4f *v ) -{ - if (v->flags & VEC_MALLOC) { - _mesa_align_free( v->storage ); - v->data = NULL; - v->start = NULL; - v->storage = NULL; - v->flags &= ~VEC_MALLOC; - } -} - - -/** - * For debugging - */ -void -_mesa_vector4f_print( const GLvector4f *v, const GLubyte *cullmask, - GLboolean culling ) -{ - static const GLfloat c[4] = { 0, 0, 0, 1 }; - static const char *templates[5] = { - "%d:\t0, 0, 0, 1\n", - "%d:\t%f, 0, 0, 1\n", - "%d:\t%f, %f, 0, 1\n", - "%d:\t%f, %f, %f, 1\n", - "%d:\t%f, %f, %f, %f\n" - }; - - const char *t = templates[v->size]; - GLfloat *d = (GLfloat *)v->data; - GLuint j, i = 0, count; - - printf("data-start\n"); - for (; d != v->start; STRIDE_F(d, v->stride), i++) - printf(t, i, d[0], d[1], d[2], d[3]); - - printf("start-count(%u)\n", v->count); - count = i + v->count; - - if (culling) { - for (; i < count; STRIDE_F(d, v->stride), i++) - if (cullmask[i]) - printf(t, i, d[0], d[1], d[2], d[3]); - } - else { - for (; i < count; STRIDE_F(d, v->stride), i++) - printf(t, i, d[0], d[1], d[2], d[3]); - } - - for (j = v->size; j < 4; j++) { - if ((v->flags & (1<data; - i < count && d[j] == c[j]; - i++, STRIDE_F(d, v->stride)) { - /* no-op */ - } - - if (i == count) - printf(" --> ok\n"); - else - printf(" --> Failed at %u ******\n", i); - } - } -} +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * New (3.1) transformation code written by Keith Whitwell. + */ + + +#include "main/glheader.h" +#include "main/imports.h" +#include "main/macros.h" + +#include "m_vector.h" + + + +/** + * Given a vector [count][4] of floats, set all the [][elt] values + * to 0 (if elt = 0, 1, 2) or 1.0 (if elt = 3). + */ +void +_mesa_vector4f_clean_elem( GLvector4f *vec, GLuint count, GLuint elt ) +{ + static const GLubyte elem_bits[4] = { + VEC_DIRTY_0, + VEC_DIRTY_1, + VEC_DIRTY_2, + VEC_DIRTY_3 + }; + static const GLfloat clean[4] = { 0, 0, 0, 1 }; + const GLfloat v = clean[elt]; + GLfloat (*data)[4] = (GLfloat (*)[4])vec->start; + GLuint i; + + for (i = 0; i < count; i++) + data[i][elt] = v; + + vec->flags &= ~elem_bits[elt]; +} + + +static const GLubyte size_bits[5] = { + 0, + VEC_SIZE_1, + VEC_SIZE_2, + VEC_SIZE_3, + VEC_SIZE_4, +}; + + +/** + * Initialize GLvector objects. + * \param v the vector object to initialize. + * \param flags bitwise-OR of VEC_* flags + * \param storage pointer to storage for the vector's data + */ +void +_mesa_vector4f_init( GLvector4f *v, GLbitfield flags, GLfloat (*storage)[4] ) +{ + v->stride = 4 * sizeof(GLfloat); + v->size = 2; /* may change: 2-4 for vertices and 1-4 for texcoords */ + v->data = storage; + v->start = (GLfloat *) storage; + v->count = 0; + v->flags = size_bits[4] | flags; +} + + +/** + * Initialize GLvector objects and allocate storage. + * \param v the vector object + * \param flags bitwise-OR of VEC_* flags + * \param count number of elements to allocate in vector + * \param alignment desired memory alignment for the data (in bytes) + */ +void +_mesa_vector4f_alloc( GLvector4f *v, GLbitfield flags, GLuint count, + GLuint alignment ) +{ + v->stride = 4 * sizeof(GLfloat); + v->size = 2; + v->storage = _mesa_align_malloc( count * 4 * sizeof(GLfloat), alignment ); + v->storage_count = count; + v->start = (GLfloat *) v->storage; + v->data = (GLfloat (*)[4]) v->storage; + v->count = 0; + v->flags = size_bits[4] | flags | VEC_MALLOC; +} + + +/** + * Vector deallocation. Free whatever memory is pointed to by the + * vector's storage field if the VEC_MALLOC flag is set. + * DO NOT free the GLvector object itself, though. + */ +void +_mesa_vector4f_free( GLvector4f *v ) +{ + if (v->flags & VEC_MALLOC) { + _mesa_align_free( v->storage ); + v->data = NULL; + v->start = NULL; + v->storage = NULL; + v->flags &= ~VEC_MALLOC; + } +} + + +/** + * For debugging + */ +void +_mesa_vector4f_print( const GLvector4f *v, const GLubyte *cullmask, + GLboolean culling ) +{ + static const GLfloat c[4] = { 0, 0, 0, 1 }; + static const char *templates[5] = { + "%d:\t0, 0, 0, 1\n", + "%d:\t%f, 0, 0, 1\n", + "%d:\t%f, %f, 0, 1\n", + "%d:\t%f, %f, %f, 1\n", + "%d:\t%f, %f, %f, %f\n" + }; + + const char *t = templates[v->size]; + GLfloat *d = (GLfloat *)v->data; + GLuint j, i = 0, count; + + printf("data-start\n"); + for (; d != v->start; STRIDE_F(d, v->stride), i++) + printf(t, i, d[0], d[1], d[2], d[3]); + + printf("start-count(%u)\n", v->count); + count = i + v->count; + + if (culling) { + for (; i < count; STRIDE_F(d, v->stride), i++) + if (cullmask[i]) + printf(t, i, d[0], d[1], d[2], d[3]); + } + else { + for (; i < count; STRIDE_F(d, v->stride), i++) + printf(t, i, d[0], d[1], d[2], d[3]); + } + + for (j = v->size; j < 4; j++) { + if ((v->flags & (1<data; + i < count && d[j] == c[j]; + i++, STRIDE_F(d, v->stride)) { + /* no-op */ + } + + if (i == count) + printf(" --> ok\n"); + else + printf(" --> Failed at %u ******\n", i); + } + } +} diff --git a/mesalib/src/mesa/program/.gitignore b/mesalib/src/mesa/program/.gitignore new file mode 100644 index 000000000..4c20872e1 --- /dev/null +++ b/mesalib/src/mesa/program/.gitignore @@ -0,0 +1,4 @@ +program_parse.output +lex.yy.c +program_parse.tab.c +program_parse.tab.h diff --git a/mesalib/src/mesa/program/arbprogparse.h b/mesalib/src/mesa/program/arbprogparse.h index a9fe1e941..4c0c30072 100644 --- a/mesalib/src/mesa/program/arbprogparse.h +++ b/mesalib/src/mesa/program/arbprogparse.h @@ -1,45 +1,45 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef ARBPROGPARSE_H -#define ARBPROGPARSE_H - -#include "main/glheader.h" - -struct gl_context; -struct gl_fragment_program; -struct gl_vertex_program; - -extern void -_mesa_parse_arb_vertex_program(struct gl_context *ctx, GLenum target, - const GLvoid *str, GLsizei len, - struct gl_vertex_program *program); - -extern void -_mesa_parse_arb_fragment_program(struct gl_context *ctx, GLenum target, - const GLvoid *str, GLsizei len, - struct gl_fragment_program *program); - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef ARBPROGPARSE_H +#define ARBPROGPARSE_H + +#include "main/glheader.h" + +struct gl_context; +struct gl_fragment_program; +struct gl_vertex_program; + +extern void +_mesa_parse_arb_vertex_program(struct gl_context *ctx, GLenum target, + const GLvoid *str, GLsizei len, + struct gl_vertex_program *program); + +extern void +_mesa_parse_arb_fragment_program(struct gl_context *ctx, GLenum target, + const GLvoid *str, GLsizei len, + struct gl_fragment_program *program); + +#endif diff --git a/mesalib/src/mesa/program/ir_to_mesa.cpp b/mesalib/src/mesa/program/ir_to_mesa.cpp index c021c6956..224540005 100644 --- a/mesalib/src/mesa/program/ir_to_mesa.cpp +++ b/mesalib/src/mesa/program/ir_to_mesa.cpp @@ -1407,6 +1407,11 @@ ir_to_mesa_visitor::visit(ir_expression *ir) emit(ir, OPCODE_SNE, result_dst, op[0], src_reg_for_float(0.0)); break; + case ir_unop_bitcast_f2i: // Ignore these 4, they can't happen here anyway + case ir_unop_bitcast_f2u: + case ir_unop_bitcast_i2f: + case ir_unop_bitcast_u2f: + break; case ir_unop_trunc: emit(ir, OPCODE_TRUNC, result_dst, op[0]); break; diff --git a/mesalib/src/mesa/program/nvfragparse.h b/mesalib/src/mesa/program/nvfragparse.h index a8ab8d6fe..088e7527d 100644 --- a/mesalib/src/mesa/program/nvfragparse.h +++ b/mesalib/src/mesa/program/nvfragparse.h @@ -1,47 +1,47 @@ - -/* - * Mesa 3-D graphics library - * Version: 5.1 - * - * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Brian Paul - */ - - -#ifndef NVFRAGPARSE_H -#define NVFRAGPARSE_H - -#include "main/glheader.h" - -struct gl_context; -struct gl_fragment_program; - -extern void -_mesa_parse_nv_fragment_program(struct gl_context *ctx, GLenum target, - const GLubyte *str, GLsizei len, - struct gl_fragment_program *program); - - -extern const char * -_mesa_nv_fragment_input_register_name(GLuint i); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 5.1 + * + * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Brian Paul + */ + + +#ifndef NVFRAGPARSE_H +#define NVFRAGPARSE_H + +#include "main/glheader.h" + +struct gl_context; +struct gl_fragment_program; + +extern void +_mesa_parse_nv_fragment_program(struct gl_context *ctx, GLenum target, + const GLubyte *str, GLsizei len, + struct gl_fragment_program *program); + + +extern const char * +_mesa_nv_fragment_input_register_name(GLuint i); + +#endif diff --git a/mesalib/src/mesa/program/nvvertparse.h b/mesalib/src/mesa/program/nvvertparse.h index bee6dea3c..7318e1494 100644 --- a/mesalib/src/mesa/program/nvvertparse.h +++ b/mesalib/src/mesa/program/nvvertparse.h @@ -1,49 +1,49 @@ -/* - * Mesa 3-D graphics library - * Version: 5.1 - * - * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Brian Paul - */ - - -#ifndef NVVERTPARSE_H -#define NVVERTPARSE_H - -#include "main/glheader.h" - -struct gl_context; -struct gl_vertex_program; - -extern void -_mesa_parse_nv_vertex_program(struct gl_context *ctx, GLenum target, - const GLubyte *str, GLsizei len, - struct gl_vertex_program *program); - - -extern const char * -_mesa_nv_vertex_input_register_name(GLuint i); - -extern const char * -_mesa_nv_vertex_output_register_name(GLuint i); - -#endif +/* + * Mesa 3-D graphics library + * Version: 5.1 + * + * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Brian Paul + */ + + +#ifndef NVVERTPARSE_H +#define NVVERTPARSE_H + +#include "main/glheader.h" + +struct gl_context; +struct gl_vertex_program; + +extern void +_mesa_parse_nv_vertex_program(struct gl_context *ctx, GLenum target, + const GLubyte *str, GLsizei len, + struct gl_vertex_program *program); + + +extern const char * +_mesa_nv_vertex_input_register_name(GLuint i); + +extern const char * +_mesa_nv_vertex_output_register_name(GLuint i); + +#endif diff --git a/mesalib/src/mesa/program/prog_cache.c b/mesalib/src/mesa/program/prog_cache.c index 398cd5e6d..2ccedb5d7 100644 --- a/mesalib/src/mesa/program/prog_cache.c +++ b/mesalib/src/mesa/program/prog_cache.c @@ -1,251 +1,251 @@ -/************************************************************************** - * - * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#include "main/glheader.h" -#include "main/mtypes.h" -#include "main/imports.h" -#include "main/shaderobj.h" -#include "program/prog_cache.h" -#include "program/program.h" - - -struct cache_item -{ - GLuint hash; - void *key; - struct gl_program *program; - struct cache_item *next; -}; - -struct gl_program_cache -{ - struct cache_item **items; - struct cache_item *last; - GLuint size, n_items; -}; - - - -/** - * Compute hash index from state key. - */ -static GLuint -hash_key(const void *key, GLuint key_size) -{ - const GLuint *ikey = (const GLuint *) key; - GLuint hash = 0, i; - - assert(key_size >= 4); - - /* Make a slightly better attempt at a hash function: - */ - for (i = 0; i < key_size / sizeof(*ikey); i++) - { - hash += ikey[i]; - hash += (hash << 10); - hash ^= (hash >> 6); - } - - return hash; -} - - -/** - * Rebuild/expand the hash table to accomodate more entries - */ -static void -rehash(struct gl_program_cache *cache) -{ - struct cache_item **items; - struct cache_item *c, *next; - GLuint size, i; - - cache->last = NULL; - - size = cache->size * 3; - items = (struct cache_item**) malloc(size * sizeof(*items)); - memset(items, 0, size * sizeof(*items)); - - for (i = 0; i < cache->size; i++) - for (c = cache->items[i]; c; c = next) { - next = c->next; - c->next = items[c->hash % size]; - items[c->hash % size] = c; - } - - free(cache->items); - cache->items = items; - cache->size = size; -} - - -static void -clear_cache(struct gl_context *ctx, struct gl_program_cache *cache, - GLboolean shader) -{ - struct cache_item *c, *next; - GLuint i; - - cache->last = NULL; - - for (i = 0; i < cache->size; i++) { - for (c = cache->items[i]; c; c = next) { - next = c->next; - free(c->key); - if (shader) { - _mesa_reference_shader_program(ctx, - (struct gl_shader_program **)&c->program, - NULL); - } else { - _mesa_reference_program(ctx, &c->program, NULL); - } - free(c); - } - cache->items[i] = NULL; - } - - - cache->n_items = 0; -} - - - -struct gl_program_cache * -_mesa_new_program_cache(void) -{ - struct gl_program_cache *cache = CALLOC_STRUCT(gl_program_cache); - if (cache) { - cache->size = 17; - cache->items = (struct cache_item **) - calloc(1, cache->size * sizeof(struct cache_item)); - if (!cache->items) { - free(cache); - return NULL; - } - } - return cache; -} - - -void -_mesa_delete_program_cache(struct gl_context *ctx, struct gl_program_cache *cache) -{ - clear_cache(ctx, cache, GL_FALSE); - free(cache->items); - free(cache); -} - -void -_mesa_delete_shader_cache(struct gl_context *ctx, - struct gl_program_cache *cache) -{ - clear_cache(ctx, cache, GL_TRUE); - free(cache->items); - free(cache); -} - - -struct gl_program * -_mesa_search_program_cache(struct gl_program_cache *cache, - const void *key, GLuint keysize) -{ - if (cache->last && - memcmp(cache->last->key, key, keysize) == 0) { - return cache->last->program; - } - else { - const GLuint hash = hash_key(key, keysize); - struct cache_item *c; - - for (c = cache->items[hash % cache->size]; c; c = c->next) { - if (c->hash == hash && memcmp(c->key, key, keysize) == 0) { - cache->last = c; - return c->program; - } - } - - return NULL; - } -} - - -void -_mesa_program_cache_insert(struct gl_context *ctx, - struct gl_program_cache *cache, - const void *key, GLuint keysize, - struct gl_program *program) -{ - const GLuint hash = hash_key(key, keysize); - struct cache_item *c = CALLOC_STRUCT(cache_item); - - c->hash = hash; - - c->key = malloc(keysize); - memcpy(c->key, key, keysize); - - c->program = program; /* no refcount change */ - - if (cache->n_items > cache->size * 1.5) { - if (cache->size < 1000) - rehash(cache); - else - clear_cache(ctx, cache, GL_FALSE); - } - - cache->n_items++; - c->next = cache->items[hash % cache->size]; - cache->items[hash % cache->size] = c; -} - -void -_mesa_shader_cache_insert(struct gl_context *ctx, - struct gl_program_cache *cache, - const void *key, GLuint keysize, - struct gl_shader_program *program) -{ - const GLuint hash = hash_key(key, keysize); - struct cache_item *c = CALLOC_STRUCT(cache_item); - - c->hash = hash; - - c->key = malloc(keysize); - memcpy(c->key, key, keysize); - - c->program = (struct gl_program *)program; /* no refcount change */ - - if (cache->n_items > cache->size * 1.5) { - if (cache->size < 1000) - rehash(cache); - else - clear_cache(ctx, cache, GL_TRUE); - } - - cache->n_items++; - c->next = cache->items[hash % cache->size]; - cache->items[hash % cache->size] = c; -} +/************************************************************************** + * + * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#include "main/glheader.h" +#include "main/mtypes.h" +#include "main/imports.h" +#include "main/shaderobj.h" +#include "program/prog_cache.h" +#include "program/program.h" + + +struct cache_item +{ + GLuint hash; + void *key; + struct gl_program *program; + struct cache_item *next; +}; + +struct gl_program_cache +{ + struct cache_item **items; + struct cache_item *last; + GLuint size, n_items; +}; + + + +/** + * Compute hash index from state key. + */ +static GLuint +hash_key(const void *key, GLuint key_size) +{ + const GLuint *ikey = (const GLuint *) key; + GLuint hash = 0, i; + + assert(key_size >= 4); + + /* Make a slightly better attempt at a hash function: + */ + for (i = 0; i < key_size / sizeof(*ikey); i++) + { + hash += ikey[i]; + hash += (hash << 10); + hash ^= (hash >> 6); + } + + return hash; +} + + +/** + * Rebuild/expand the hash table to accomodate more entries + */ +static void +rehash(struct gl_program_cache *cache) +{ + struct cache_item **items; + struct cache_item *c, *next; + GLuint size, i; + + cache->last = NULL; + + size = cache->size * 3; + items = (struct cache_item**) malloc(size * sizeof(*items)); + memset(items, 0, size * sizeof(*items)); + + for (i = 0; i < cache->size; i++) + for (c = cache->items[i]; c; c = next) { + next = c->next; + c->next = items[c->hash % size]; + items[c->hash % size] = c; + } + + free(cache->items); + cache->items = items; + cache->size = size; +} + + +static void +clear_cache(struct gl_context *ctx, struct gl_program_cache *cache, + GLboolean shader) +{ + struct cache_item *c, *next; + GLuint i; + + cache->last = NULL; + + for (i = 0; i < cache->size; i++) { + for (c = cache->items[i]; c; c = next) { + next = c->next; + free(c->key); + if (shader) { + _mesa_reference_shader_program(ctx, + (struct gl_shader_program **)&c->program, + NULL); + } else { + _mesa_reference_program(ctx, &c->program, NULL); + } + free(c); + } + cache->items[i] = NULL; + } + + + cache->n_items = 0; +} + + + +struct gl_program_cache * +_mesa_new_program_cache(void) +{ + struct gl_program_cache *cache = CALLOC_STRUCT(gl_program_cache); + if (cache) { + cache->size = 17; + cache->items = (struct cache_item **) + calloc(1, cache->size * sizeof(struct cache_item)); + if (!cache->items) { + free(cache); + return NULL; + } + } + return cache; +} + + +void +_mesa_delete_program_cache(struct gl_context *ctx, struct gl_program_cache *cache) +{ + clear_cache(ctx, cache, GL_FALSE); + free(cache->items); + free(cache); +} + +void +_mesa_delete_shader_cache(struct gl_context *ctx, + struct gl_program_cache *cache) +{ + clear_cache(ctx, cache, GL_TRUE); + free(cache->items); + free(cache); +} + + +struct gl_program * +_mesa_search_program_cache(struct gl_program_cache *cache, + const void *key, GLuint keysize) +{ + if (cache->last && + memcmp(cache->last->key, key, keysize) == 0) { + return cache->last->program; + } + else { + const GLuint hash = hash_key(key, keysize); + struct cache_item *c; + + for (c = cache->items[hash % cache->size]; c; c = c->next) { + if (c->hash == hash && memcmp(c->key, key, keysize) == 0) { + cache->last = c; + return c->program; + } + } + + return NULL; + } +} + + +void +_mesa_program_cache_insert(struct gl_context *ctx, + struct gl_program_cache *cache, + const void *key, GLuint keysize, + struct gl_program *program) +{ + const GLuint hash = hash_key(key, keysize); + struct cache_item *c = CALLOC_STRUCT(cache_item); + + c->hash = hash; + + c->key = malloc(keysize); + memcpy(c->key, key, keysize); + + c->program = program; /* no refcount change */ + + if (cache->n_items > cache->size * 1.5) { + if (cache->size < 1000) + rehash(cache); + else + clear_cache(ctx, cache, GL_FALSE); + } + + cache->n_items++; + c->next = cache->items[hash % cache->size]; + cache->items[hash % cache->size] = c; +} + +void +_mesa_shader_cache_insert(struct gl_context *ctx, + struct gl_program_cache *cache, + const void *key, GLuint keysize, + struct gl_shader_program *program) +{ + const GLuint hash = hash_key(key, keysize); + struct cache_item *c = CALLOC_STRUCT(cache_item); + + c->hash = hash; + + c->key = malloc(keysize); + memcpy(c->key, key, keysize); + + c->program = (struct gl_program *)program; /* no refcount change */ + + if (cache->n_items > cache->size * 1.5) { + if (cache->size < 1000) + rehash(cache); + else + clear_cache(ctx, cache, GL_TRUE); + } + + cache->n_items++; + c->next = cache->items[hash % cache->size]; + cache->items[hash % cache->size] = c; +} diff --git a/mesalib/src/mesa/program/prog_cache.h b/mesalib/src/mesa/program/prog_cache.h index e7ec8d39a..5d46bfc5c 100644 --- a/mesalib/src/mesa/program/prog_cache.h +++ b/mesalib/src/mesa/program/prog_cache.h @@ -1,68 +1,68 @@ -/************************************************************************** - * - * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef PROG_CACHE_H -#define PROG_CACHE_H - - -#include "main/glheader.h" - -struct gl_context; - -/** Opaque type */ -struct gl_program_cache; - - -extern struct gl_program_cache * -_mesa_new_program_cache(void); - -extern void -_mesa_delete_program_cache(struct gl_context *ctx, struct gl_program_cache *pc); - -extern void -_mesa_delete_shader_cache(struct gl_context *ctx, - struct gl_program_cache *cache); - -extern struct gl_program * -_mesa_search_program_cache(struct gl_program_cache *cache, - const void *key, GLuint keysize); - -extern void -_mesa_program_cache_insert(struct gl_context *ctx, - struct gl_program_cache *cache, - const void *key, GLuint keysize, - struct gl_program *program); - -void -_mesa_shader_cache_insert(struct gl_context *ctx, - struct gl_program_cache *cache, - const void *key, GLuint keysize, - struct gl_shader_program *program); - - -#endif /* PROG_CACHE_H */ +/************************************************************************** + * + * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef PROG_CACHE_H +#define PROG_CACHE_H + + +#include "main/glheader.h" + +struct gl_context; + +/** Opaque type */ +struct gl_program_cache; + + +extern struct gl_program_cache * +_mesa_new_program_cache(void); + +extern void +_mesa_delete_program_cache(struct gl_context *ctx, struct gl_program_cache *pc); + +extern void +_mesa_delete_shader_cache(struct gl_context *ctx, + struct gl_program_cache *cache); + +extern struct gl_program * +_mesa_search_program_cache(struct gl_program_cache *cache, + const void *key, GLuint keysize); + +extern void +_mesa_program_cache_insert(struct gl_context *ctx, + struct gl_program_cache *cache, + const void *key, GLuint keysize, + struct gl_program *program); + +void +_mesa_shader_cache_insert(struct gl_context *ctx, + struct gl_program_cache *cache, + const void *key, GLuint keysize, + struct gl_shader_program *program); + + +#endif /* PROG_CACHE_H */ diff --git a/mesalib/src/mesa/program/sampler.h b/mesalib/src/mesa/program/sampler.h index c76d7077b..22467e990 100644 --- a/mesalib/src/mesa/program/sampler.h +++ b/mesalib/src/mesa/program/sampler.h @@ -1,29 +1,29 @@ -/* - * Copyright (C) 2005-2007 Brian Paul All Rights Reserved. - * Copyright (C) 2008 VMware, Inc. All Rights Reserved. - * Copyright © 2010 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ - -int -_mesa_get_sampler_uniform_value(class ir_dereference *sampler, - struct gl_shader_program *shader_program, - const struct gl_program *prog); +/* + * Copyright (C) 2005-2007 Brian Paul All Rights Reserved. + * Copyright (C) 2008 VMware, Inc. All Rights Reserved. + * Copyright © 2010 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ + +int +_mesa_get_sampler_uniform_value(class ir_dereference *sampler, + struct gl_shader_program *shader_program, + const struct gl_program *prog); diff --git a/mesalib/src/mesa/program/symbol_table.h b/mesalib/src/mesa/program/symbol_table.h index 1de29a8f0..f9d91649b 100644 --- a/mesalib/src/mesa/program/symbol_table.h +++ b/mesalib/src/mesa/program/symbol_table.h @@ -1,62 +1,62 @@ -/* - * Copyright © 2008 Intel Corporation - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - */ -#ifndef MESA_SYMBOL_TABLE_H -#define MESA_SYMBOL_TABLE_H - -struct _mesa_symbol_table; -struct _mesa_symbol_table_iterator; - -extern void _mesa_symbol_table_push_scope(struct _mesa_symbol_table *table); - -extern void _mesa_symbol_table_pop_scope(struct _mesa_symbol_table *table); - -extern int _mesa_symbol_table_add_symbol(struct _mesa_symbol_table *symtab, - int name_space, const char *name, void *declaration); - -extern int _mesa_symbol_table_add_global_symbol( - struct _mesa_symbol_table *symtab, int name_space, const char *name, - void *declaration); - -extern int _mesa_symbol_table_symbol_scope(struct _mesa_symbol_table *table, - int name_space, const char *name); - -extern void *_mesa_symbol_table_find_symbol( - struct _mesa_symbol_table *symtab, int name_space, const char *name); - -extern struct _mesa_symbol_table *_mesa_symbol_table_ctor(void); - -extern void _mesa_symbol_table_dtor(struct _mesa_symbol_table *); - -extern struct _mesa_symbol_table_iterator *_mesa_symbol_table_iterator_ctor( - struct _mesa_symbol_table *table, int name_space, const char *name); - -extern void _mesa_symbol_table_iterator_dtor( - struct _mesa_symbol_table_iterator *); - -extern void *_mesa_symbol_table_iterator_get( - struct _mesa_symbol_table_iterator *iter); - -extern int _mesa_symbol_table_iterator_next( - struct _mesa_symbol_table_iterator *iter); - -#endif /* MESA_SYMBOL_TABLE_H */ +/* + * Copyright © 2008 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + */ +#ifndef MESA_SYMBOL_TABLE_H +#define MESA_SYMBOL_TABLE_H + +struct _mesa_symbol_table; +struct _mesa_symbol_table_iterator; + +extern void _mesa_symbol_table_push_scope(struct _mesa_symbol_table *table); + +extern void _mesa_symbol_table_pop_scope(struct _mesa_symbol_table *table); + +extern int _mesa_symbol_table_add_symbol(struct _mesa_symbol_table *symtab, + int name_space, const char *name, void *declaration); + +extern int _mesa_symbol_table_add_global_symbol( + struct _mesa_symbol_table *symtab, int name_space, const char *name, + void *declaration); + +extern int _mesa_symbol_table_symbol_scope(struct _mesa_symbol_table *table, + int name_space, const char *name); + +extern void *_mesa_symbol_table_find_symbol( + struct _mesa_symbol_table *symtab, int name_space, const char *name); + +extern struct _mesa_symbol_table *_mesa_symbol_table_ctor(void); + +extern void _mesa_symbol_table_dtor(struct _mesa_symbol_table *); + +extern struct _mesa_symbol_table_iterator *_mesa_symbol_table_iterator_ctor( + struct _mesa_symbol_table *table, int name_space, const char *name); + +extern void _mesa_symbol_table_iterator_dtor( + struct _mesa_symbol_table_iterator *); + +extern void *_mesa_symbol_table_iterator_get( + struct _mesa_symbol_table_iterator *iter); + +extern int _mesa_symbol_table_iterator_next( + struct _mesa_symbol_table_iterator *iter); + +#endif /* MESA_SYMBOL_TABLE_H */ diff --git a/mesalib/src/mesa/state_tracker/Makefile b/mesalib/src/mesa/state_tracker/Makefile index 4a38ca8f1..0ab1dc6e6 100644 --- a/mesalib/src/mesa/state_tracker/Makefile +++ b/mesalib/src/mesa/state_tracker/Makefile @@ -1,2 +1,2 @@ -default: +default: cd ../.. ; make \ No newline at end of file diff --git a/mesalib/src/mesa/state_tracker/st_atom.h b/mesalib/src/mesa/state_tracker/st_atom.h index b5f837f78..930a08444 100644 --- a/mesalib/src/mesa/state_tracker/st_atom.h +++ b/mesalib/src/mesa/state_tracker/st_atom.h @@ -1,73 +1,73 @@ -/************************************************************************** - * - * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - /* - * Authors: - * Keith Whitwell - */ - - -#ifndef ST_ATOM_H -#define ST_ATOM_H - -#include "main/glheader.h" - -struct st_context; -struct st_tracked_state; - -void st_init_atoms( struct st_context *st ); -void st_destroy_atoms( struct st_context *st ); - - -void st_validate_state( struct st_context *st ); - - -extern const struct st_tracked_state st_update_framebuffer; -extern const struct st_tracked_state st_update_clip; -extern const struct st_tracked_state st_update_depth_stencil_alpha; -extern const struct st_tracked_state st_update_fp; -extern const struct st_tracked_state st_update_gp; -extern const struct st_tracked_state st_update_vp; -extern const struct st_tracked_state st_update_rasterizer; -extern const struct st_tracked_state st_update_polygon_stipple; -extern const struct st_tracked_state st_update_viewport; -extern const struct st_tracked_state st_update_scissor; -extern const struct st_tracked_state st_update_blend; -extern const struct st_tracked_state st_update_msaa; -extern const struct st_tracked_state st_update_sampler; -extern const struct st_tracked_state st_update_texture; -extern const struct st_tracked_state st_update_vertex_texture; -extern const struct st_tracked_state st_finalize_textures; -extern const struct st_tracked_state st_update_fs_constants; -extern const struct st_tracked_state st_update_gs_constants; -extern const struct st_tracked_state st_update_vs_constants; -extern const struct st_tracked_state st_update_pixel_transfer; - - -GLuint st_compare_func_to_pipe(GLenum func); - -#endif +/************************************************************************** + * + * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + /* + * Authors: + * Keith Whitwell + */ + + +#ifndef ST_ATOM_H +#define ST_ATOM_H + +#include "main/glheader.h" + +struct st_context; +struct st_tracked_state; + +void st_init_atoms( struct st_context *st ); +void st_destroy_atoms( struct st_context *st ); + + +void st_validate_state( struct st_context *st ); + + +extern const struct st_tracked_state st_update_framebuffer; +extern const struct st_tracked_state st_update_clip; +extern const struct st_tracked_state st_update_depth_stencil_alpha; +extern const struct st_tracked_state st_update_fp; +extern const struct st_tracked_state st_update_gp; +extern const struct st_tracked_state st_update_vp; +extern const struct st_tracked_state st_update_rasterizer; +extern const struct st_tracked_state st_update_polygon_stipple; +extern const struct st_tracked_state st_update_viewport; +extern const struct st_tracked_state st_update_scissor; +extern const struct st_tracked_state st_update_blend; +extern const struct st_tracked_state st_update_msaa; +extern const struct st_tracked_state st_update_sampler; +extern const struct st_tracked_state st_update_texture; +extern const struct st_tracked_state st_update_vertex_texture; +extern const struct st_tracked_state st_finalize_textures; +extern const struct st_tracked_state st_update_fs_constants; +extern const struct st_tracked_state st_update_gs_constants; +extern const struct st_tracked_state st_update_vs_constants; +extern const struct st_tracked_state st_update_pixel_transfer; + + +GLuint st_compare_func_to_pipe(GLenum func); + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_atom_constbuf.h b/mesalib/src/mesa/state_tracker/st_atom_constbuf.h index d53654ee6..97b076629 100644 --- a/mesalib/src/mesa/state_tracker/st_atom_constbuf.h +++ b/mesalib/src/mesa/state_tracker/st_atom_constbuf.h @@ -1,41 +1,41 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_ATOM_CONSTBUF_H -#define ST_ATOM_CONSTBUF_H - -struct gl_program_parameter_list; -struct st_context; - - -void st_upload_constants( struct st_context *st, - struct gl_program_parameter_list *params, - unsigned id); - - -#endif /* ST_ATOM_CONSTBUF_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_ATOM_CONSTBUF_H +#define ST_ATOM_CONSTBUF_H + +struct gl_program_parameter_list; +struct st_context; + + +void st_upload_constants( struct st_context *st, + struct gl_program_parameter_list *params, + unsigned id); + + +#endif /* ST_ATOM_CONSTBUF_H */ diff --git a/mesalib/src/mesa/state_tracker/st_atom_msaa.c b/mesalib/src/mesa/state_tracker/st_atom_msaa.c index 73f0ff160..ea9eb9a6d 100644 --- a/mesalib/src/mesa/state_tracker/st_atom_msaa.c +++ b/mesalib/src/mesa/state_tracker/st_atom_msaa.c @@ -1,83 +1,83 @@ -/************************************************************************** - * - * Copyright 2010 VMware, Inc. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#include "st_context.h" -#include "pipe/p_context.h" -#include "st_atom.h" - -#include "cso_cache/cso_context.h" - - -/* Second state atom for user clip planes: - */ -static void update_sample_mask( struct st_context *st ) -{ - unsigned sample_mask = 0xffffffff; - unsigned sample_count = 1; - struct pipe_framebuffer_state *framebuffer = &st->state.framebuffer; - - /* dependency here on bound surface (or rather, sample count) is worrying */ - if (framebuffer->zsbuf) - sample_count = framebuffer->zsbuf->texture->nr_samples; - else if (framebuffer->cbufs[0]) - sample_count = framebuffer->cbufs[0]->texture->nr_samples; - - if (st->ctx->Multisample.Enabled && sample_count > 1) { - /* unlike in gallium/d3d10 the mask is only active if msaa is enabled */ - if (st->ctx->Multisample.SampleCoverage) { - unsigned nr_bits; - nr_bits = st->ctx->Multisample.SampleCoverageValue * (float)sample_count; - /* there's lot of ways how to do this. We just use first few bits, - since we have no knowledge of sample positions here. When - app-supplied mask though is used too might need to be smarter. - Also, there's a interface restriction here in theory it is - encouraged this mask not be the same at each pixel. */ - sample_mask = (1 << nr_bits) - 1; - if (st->ctx->Multisample.SampleCoverageInvert) - sample_mask = ~sample_mask; - } - /* TODO merge with app-supplied sample mask */ - } - - /* mask off unused bits or don't care? */ - - if (sample_mask != st->state.sample_mask) { - st->state.sample_mask = sample_mask; - cso_set_sample_mask(st->cso_context, sample_mask); - } -} - - -const struct st_tracked_state st_update_msaa = { - "st_update_msaa", /* name */ - { /* dirty */ - (_NEW_MULTISAMPLE | _NEW_BUFFERS), /* mesa */ - ST_NEW_FRAMEBUFFER, /* st */ - }, - update_sample_mask /* update */ -}; +/************************************************************************** + * + * Copyright 2010 VMware, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#include "st_context.h" +#include "pipe/p_context.h" +#include "st_atom.h" + +#include "cso_cache/cso_context.h" + + +/* Second state atom for user clip planes: + */ +static void update_sample_mask( struct st_context *st ) +{ + unsigned sample_mask = 0xffffffff; + unsigned sample_count = 1; + struct pipe_framebuffer_state *framebuffer = &st->state.framebuffer; + + /* dependency here on bound surface (or rather, sample count) is worrying */ + if (framebuffer->zsbuf) + sample_count = framebuffer->zsbuf->texture->nr_samples; + else if (framebuffer->cbufs[0]) + sample_count = framebuffer->cbufs[0]->texture->nr_samples; + + if (st->ctx->Multisample.Enabled && sample_count > 1) { + /* unlike in gallium/d3d10 the mask is only active if msaa is enabled */ + if (st->ctx->Multisample.SampleCoverage) { + unsigned nr_bits; + nr_bits = st->ctx->Multisample.SampleCoverageValue * (float)sample_count; + /* there's lot of ways how to do this. We just use first few bits, + since we have no knowledge of sample positions here. When + app-supplied mask though is used too might need to be smarter. + Also, there's a interface restriction here in theory it is + encouraged this mask not be the same at each pixel. */ + sample_mask = (1 << nr_bits) - 1; + if (st->ctx->Multisample.SampleCoverageInvert) + sample_mask = ~sample_mask; + } + /* TODO merge with app-supplied sample mask */ + } + + /* mask off unused bits or don't care? */ + + if (sample_mask != st->state.sample_mask) { + st->state.sample_mask = sample_mask; + cso_set_sample_mask(st->cso_context, sample_mask); + } +} + + +const struct st_tracked_state st_update_msaa = { + "st_update_msaa", /* name */ + { /* dirty */ + (_NEW_MULTISAMPLE | _NEW_BUFFERS), /* mesa */ + ST_NEW_FRAMEBUFFER, /* st */ + }, + update_sample_mask /* update */ +}; diff --git a/mesalib/src/mesa/state_tracker/st_atom_shader.h b/mesalib/src/mesa/state_tracker/st_atom_shader.h index fbb3a4137..56d4c68f4 100644 --- a/mesalib/src/mesa/state_tracker/st_atom_shader.h +++ b/mesalib/src/mesa/state_tracker/st_atom_shader.h @@ -1,41 +1,41 @@ -/************************************************************************** - * - * Copyright 2007-2008 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_ATOM_SHADER_H -#define ST_ATOM_SHADER_H - - -struct st_context; -struct translated_vertex_program; - -extern void -st_free_translated_vertex_programs(struct st_context *st, - struct translated_vertex_program *xvp); - - -#endif /* ST_ATOM_SHADER_H */ +/************************************************************************** + * + * Copyright 2007-2008 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_ATOM_SHADER_H +#define ST_ATOM_SHADER_H + + +struct st_context; +struct translated_vertex_program; + +extern void +st_free_translated_vertex_programs(struct st_context *st, + struct translated_vertex_program *xvp); + + +#endif /* ST_ATOM_SHADER_H */ diff --git a/mesalib/src/mesa/state_tracker/st_atom_viewport.c b/mesalib/src/mesa/state_tracker/st_atom_viewport.c index df26fb4a6..133cc49ea 100644 --- a/mesalib/src/mesa/state_tracker/st_atom_viewport.c +++ b/mesalib/src/mesa/state_tracker/st_atom_viewport.c @@ -1,94 +1,94 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#include "main/context.h" -#include "st_context.h" -#include "st_atom.h" -#include "pipe/p_context.h" -#include "cso_cache/cso_context.h" - -/** - * Update the viewport transformation matrix. Depends on: - * - viewport pos/size - * - depthrange - * - window pos/size or FBO size - */ -static void -update_viewport( struct st_context *st ) -{ - struct gl_context *ctx = st->ctx; - GLfloat yScale, yBias; - - /* _NEW_BUFFERS - */ - if (st_fb_orientation(ctx->DrawBuffer) == Y_0_TOP) { - /* Drawing to a window. The corresponding gallium surface uses - * Y=0=TOP but OpenGL is Y=0=BOTTOM. So we need to invert the viewport. - */ - yScale = -1; - yBias = (GLfloat)ctx->DrawBuffer->Height; - } - else { - /* Drawing to an FBO where Y=0=BOTTOM, like OpenGL - don't invert */ - yScale = 1.0; - yBias = 0.0; - } - - /* _NEW_VIEWPORT - */ - { - GLfloat x = (GLfloat)ctx->Viewport.X; - GLfloat y = (GLfloat)ctx->Viewport.Y; - GLfloat z = ctx->Viewport.Near; - GLfloat half_width = (GLfloat)ctx->Viewport.Width * 0.5f; - GLfloat half_height = (GLfloat)ctx->Viewport.Height * 0.5f; - GLfloat half_depth = (GLfloat)(ctx->Viewport.Far - ctx->Viewport.Near) * 0.5f; - - st->state.viewport.scale[0] = half_width; - st->state.viewport.scale[1] = half_height * yScale; - st->state.viewport.scale[2] = half_depth; - st->state.viewport.scale[3] = 1.0; - - st->state.viewport.translate[0] = half_width + x; - st->state.viewport.translate[1] = (half_height + y) * yScale + yBias; - st->state.viewport.translate[2] = half_depth + z; - st->state.viewport.translate[3] = 0.0; - - cso_set_viewport(st->cso_context, &st->state.viewport); - } -} - - -const struct st_tracked_state st_update_viewport = { - "st_update_viewport", /* name */ - { /* dirty */ - _NEW_BUFFERS | _NEW_VIEWPORT, /* mesa */ - 0, /* st */ - }, - update_viewport /* update */ -}; +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#include "main/context.h" +#include "st_context.h" +#include "st_atom.h" +#include "pipe/p_context.h" +#include "cso_cache/cso_context.h" + +/** + * Update the viewport transformation matrix. Depends on: + * - viewport pos/size + * - depthrange + * - window pos/size or FBO size + */ +static void +update_viewport( struct st_context *st ) +{ + struct gl_context *ctx = st->ctx; + GLfloat yScale, yBias; + + /* _NEW_BUFFERS + */ + if (st_fb_orientation(ctx->DrawBuffer) == Y_0_TOP) { + /* Drawing to a window. The corresponding gallium surface uses + * Y=0=TOP but OpenGL is Y=0=BOTTOM. So we need to invert the viewport. + */ + yScale = -1; + yBias = (GLfloat)ctx->DrawBuffer->Height; + } + else { + /* Drawing to an FBO where Y=0=BOTTOM, like OpenGL - don't invert */ + yScale = 1.0; + yBias = 0.0; + } + + /* _NEW_VIEWPORT + */ + { + GLfloat x = (GLfloat)ctx->Viewport.X; + GLfloat y = (GLfloat)ctx->Viewport.Y; + GLfloat z = ctx->Viewport.Near; + GLfloat half_width = (GLfloat)ctx->Viewport.Width * 0.5f; + GLfloat half_height = (GLfloat)ctx->Viewport.Height * 0.5f; + GLfloat half_depth = (GLfloat)(ctx->Viewport.Far - ctx->Viewport.Near) * 0.5f; + + st->state.viewport.scale[0] = half_width; + st->state.viewport.scale[1] = half_height * yScale; + st->state.viewport.scale[2] = half_depth; + st->state.viewport.scale[3] = 1.0; + + st->state.viewport.translate[0] = half_width + x; + st->state.viewport.translate[1] = (half_height + y) * yScale + yBias; + st->state.viewport.translate[2] = half_depth + z; + st->state.viewport.translate[3] = 0.0; + + cso_set_viewport(st->cso_context, &st->state.viewport); + } +} + + +const struct st_tracked_state st_update_viewport = { + "st_update_viewport", /* name */ + { /* dirty */ + _NEW_BUFFERS | _NEW_VIEWPORT, /* mesa */ + 0, /* st */ + }, + update_viewport /* update */ +}; diff --git a/mesalib/src/mesa/state_tracker/st_cache.h b/mesalib/src/mesa/state_tracker/st_cache.h index 52c1b378d..6d5de7b13 100644 --- a/mesalib/src/mesa/state_tracker/st_cache.h +++ b/mesalib/src/mesa/state_tracker/st_cache.h @@ -1,69 +1,69 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - /* - * Authors: - * Zack Rusin - */ - -#ifndef ST_CACHE_H -#define ST_CACHE_H - -struct pipe_blend_state; -struct pipe_depth_stencil_alpha_state; -struct pipe_rasterizer_state; -struct pipe_sampler_state; -struct pipe_shader_state; -struct st_context; - - -const struct cso_blend * -st_cached_blend_state(struct st_context *st, - const struct pipe_blend_state *blend); - -const struct cso_sampler * -st_cached_sampler_state(struct st_context *st, - const struct pipe_sampler_state *sampler); - -const struct cso_depth_stencil_alpha * -st_cached_depth_stencil_alpha_state(struct st_context *st, - const struct pipe_depth_stencil_alpha_state *depth_stencil); - -const struct cso_rasterizer * -st_cached_rasterizer_state(struct st_context *st, - const struct pipe_rasterizer_state *raster); - -const struct cso_fragment_shader * -st_cached_fs_state(struct st_context *st, - const struct pipe_shader_state *templ); - - -const struct cso_vertex_shader * -st_cached_vs_state(struct st_context *st, - const struct pipe_shader_state *templ); - -#endif +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + /* + * Authors: + * Zack Rusin + */ + +#ifndef ST_CACHE_H +#define ST_CACHE_H + +struct pipe_blend_state; +struct pipe_depth_stencil_alpha_state; +struct pipe_rasterizer_state; +struct pipe_sampler_state; +struct pipe_shader_state; +struct st_context; + + +const struct cso_blend * +st_cached_blend_state(struct st_context *st, + const struct pipe_blend_state *blend); + +const struct cso_sampler * +st_cached_sampler_state(struct st_context *st, + const struct pipe_sampler_state *sampler); + +const struct cso_depth_stencil_alpha * +st_cached_depth_stencil_alpha_state(struct st_context *st, + const struct pipe_depth_stencil_alpha_state *depth_stencil); + +const struct cso_rasterizer * +st_cached_rasterizer_state(struct st_context *st, + const struct pipe_rasterizer_state *raster); + +const struct cso_fragment_shader * +st_cached_fs_state(struct st_context *st, + const struct pipe_shader_state *templ); + + +const struct cso_vertex_shader * +st_cached_vs_state(struct st_context *st, + const struct pipe_shader_state *templ); + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_cb_bitmap.h b/mesalib/src/mesa/state_tracker/st_cb_bitmap.h index cc6d82391..ed1415803 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_bitmap.h +++ b/mesalib/src/mesa/state_tracker/st_cb_bitmap.h @@ -1,96 +1,96 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_CB_BITMAP_H -#define ST_CB_BITMAP_H - - -#include "main/compiler.h" -#include "main/mfeatures.h" - -struct dd_function_table; -struct st_context; -struct gl_fragment_program; -struct st_fragment_program; - -#if FEATURE_drawpix - -extern void -st_init_bitmap_functions(struct dd_function_table *functions); - -extern void -st_init_bitmap(struct st_context *st); - -extern void -st_destroy_bitmap(struct st_context *st); - -extern void -st_make_bitmap_fragment_program(struct st_context *st, - struct gl_fragment_program *fpIn, - struct gl_fragment_program **fpOut, - GLuint *bitmap_sampler); - -extern void -st_flush_bitmap_cache(struct st_context *st); - -/* Flush bitmap cache and release vertex buffer. Needed at end of - * frame to avoid synchronous rendering. - */ -extern void -st_flush_bitmap(struct st_context *st); - -#else - -static INLINE void -st_init_bitmap_functions(struct dd_function_table *functions) -{ -} - -static INLINE void -st_init_bitmap(struct st_context *st) -{ -} - -static INLINE void -st_destroy_bitmap(struct st_context *st) -{ -} - -static INLINE void -st_flush_bitmap_cache(struct st_context *st) -{ -} - -static INLINE void -st_flush_bitmap(struct st_context *st) -{ -} - -#endif /* FEATURE_drawpix */ - -#endif /* ST_CB_BITMAP_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_CB_BITMAP_H +#define ST_CB_BITMAP_H + + +#include "main/compiler.h" +#include "main/mfeatures.h" + +struct dd_function_table; +struct st_context; +struct gl_fragment_program; +struct st_fragment_program; + +#if FEATURE_drawpix + +extern void +st_init_bitmap_functions(struct dd_function_table *functions); + +extern void +st_init_bitmap(struct st_context *st); + +extern void +st_destroy_bitmap(struct st_context *st); + +extern void +st_make_bitmap_fragment_program(struct st_context *st, + struct gl_fragment_program *fpIn, + struct gl_fragment_program **fpOut, + GLuint *bitmap_sampler); + +extern void +st_flush_bitmap_cache(struct st_context *st); + +/* Flush bitmap cache and release vertex buffer. Needed at end of + * frame to avoid synchronous rendering. + */ +extern void +st_flush_bitmap(struct st_context *st); + +#else + +static INLINE void +st_init_bitmap_functions(struct dd_function_table *functions) +{ +} + +static INLINE void +st_init_bitmap(struct st_context *st) +{ +} + +static INLINE void +st_destroy_bitmap(struct st_context *st) +{ +} + +static INLINE void +st_flush_bitmap_cache(struct st_context *st) +{ +} + +static INLINE void +st_flush_bitmap(struct st_context *st) +{ +} + +#endif /* FEATURE_drawpix */ + +#endif /* ST_CB_BITMAP_H */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_blit.h b/mesalib/src/mesa/state_tracker/st_cb_blit.h index e378e6789..c6d9a3679 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_blit.h +++ b/mesalib/src/mesa/state_tracker/st_cb_blit.h @@ -1,59 +1,59 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_BLIT_H -#define ST_CB_BLIT_H - - -#include "main/compiler.h" -#include "main/mfeatures.h" - -struct dd_function_table; -struct st_context; - - -extern void -st_init_blit(struct st_context *st); - -extern void -st_destroy_blit(struct st_context *st); - -#if FEATURE_EXT_framebuffer_blit - -extern void -st_init_blit_functions(struct dd_function_table *functions); - -#else - -static INLINE void -st_init_blit_functions(struct dd_function_table *functions) -{ -} - -#endif /* FEATURE_EXT_framebuffer_blit */ - -#endif /* ST_CB_BLIT_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_BLIT_H +#define ST_CB_BLIT_H + + +#include "main/compiler.h" +#include "main/mfeatures.h" + +struct dd_function_table; +struct st_context; + + +extern void +st_init_blit(struct st_context *st); + +extern void +st_destroy_blit(struct st_context *st); + +#if FEATURE_EXT_framebuffer_blit + +extern void +st_init_blit_functions(struct dd_function_table *functions); + +#else + +static INLINE void +st_init_blit_functions(struct dd_function_table *functions) +{ +} + +#endif /* FEATURE_EXT_framebuffer_blit */ + +#endif /* ST_CB_BLIT_H */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h index 16abaa856..1c991d208 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h +++ b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h @@ -1,68 +1,68 @@ -/************************************************************************** - * - * Copyright 2005 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_BUFFEROBJECTS_H -#define ST_CB_BUFFEROBJECTS_H - -#include "main/compiler.h" -#include "main/mtypes.h" - -struct dd_function_table; -struct pipe_resource; -struct st_context; - -/** - * State_tracker vertex/pixel buffer object, derived from Mesa's - * gl_buffer_object. - */ -struct st_buffer_object -{ - struct gl_buffer_object Base; - struct pipe_resource *buffer; /* GPU storage */ - struct pipe_transfer *transfer; /* In-progress map information */ -}; - - -/** cast wrapper */ -static INLINE struct st_buffer_object * -st_buffer_object(struct gl_buffer_object *obj) -{ - return (struct st_buffer_object *) obj; -} - - -extern void -st_bufferobj_validate_usage(struct st_context *st, - struct st_buffer_object *obj, - unsigned usage); - - -extern void -st_init_bufferobject_functions(struct dd_function_table *functions); - - -#endif +/************************************************************************** + * + * Copyright 2005 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_BUFFEROBJECTS_H +#define ST_CB_BUFFEROBJECTS_H + +#include "main/compiler.h" +#include "main/mtypes.h" + +struct dd_function_table; +struct pipe_resource; +struct st_context; + +/** + * State_tracker vertex/pixel buffer object, derived from Mesa's + * gl_buffer_object. + */ +struct st_buffer_object +{ + struct gl_buffer_object Base; + struct pipe_resource *buffer; /* GPU storage */ + struct pipe_transfer *transfer; /* In-progress map information */ +}; + + +/** cast wrapper */ +static INLINE struct st_buffer_object * +st_buffer_object(struct gl_buffer_object *obj) +{ + return (struct st_buffer_object *) obj; +} + + +extern void +st_bufferobj_validate_usage(struct st_context *st, + struct st_buffer_object *obj, + unsigned usage); + + +extern void +st_init_bufferobject_functions(struct dd_function_table *functions); + + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_cb_condrender.h b/mesalib/src/mesa/state_tracker/st_cb_condrender.h index 24eb09258..79d0db8d0 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_condrender.h +++ b/mesalib/src/mesa/state_tracker/st_cb_condrender.h @@ -1,37 +1,37 @@ -/************************************************************************** - * - * Copyright 2009 VMware, Inc. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_CONDRENDER_H -#define ST_CB_CONDRENDER_H - - -struct dd_function_table; - -extern void st_init_cond_render_functions(struct dd_function_table *functions); - - -#endif +/************************************************************************** + * + * Copyright 2009 VMware, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_CONDRENDER_H +#define ST_CB_CONDRENDER_H + + +struct dd_function_table; + +extern void st_init_cond_render_functions(struct dd_function_table *functions); + + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.h b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.h index 9c02263ea..44d7f5ed1 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.h +++ b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.h @@ -1,70 +1,70 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_CB_DRAWPIXELS_H -#define ST_CB_DRAWPIXELS_H - - -#include "main/compiler.h" -#include "main/mfeatures.h" - -struct dd_function_table; -struct st_context; - -#if FEATURE_drawpix - -extern void st_init_drawpixels_functions(struct dd_function_table *functions); - -extern void -st_destroy_drawpix(struct st_context *st); - -extern void -st_make_drawpix_fragment_program(struct st_context *st, - struct gl_fragment_program *fpIn, - struct gl_fragment_program **fpOut); - -extern struct gl_fragment_program * -st_make_drawpix_z_stencil_program(struct st_context *st, - GLboolean write_depth, - GLboolean write_stencil); - -#else - -static INLINE void -st_init_drawpixels_functions(struct dd_function_table *functions) -{ -} - -static INLINE void -st_destroy_drawpix(struct st_context *st) -{ -} - -#endif /* FEATURE_drawpix */ - -#endif /* ST_CB_DRAWPIXELS_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_CB_DRAWPIXELS_H +#define ST_CB_DRAWPIXELS_H + + +#include "main/compiler.h" +#include "main/mfeatures.h" + +struct dd_function_table; +struct st_context; + +#if FEATURE_drawpix + +extern void st_init_drawpixels_functions(struct dd_function_table *functions); + +extern void +st_destroy_drawpix(struct st_context *st); + +extern void +st_make_drawpix_fragment_program(struct st_context *st, + struct gl_fragment_program *fpIn, + struct gl_fragment_program **fpOut); + +extern struct gl_fragment_program * +st_make_drawpix_z_stencil_program(struct st_context *st, + GLboolean write_depth, + GLboolean write_stencil); + +#else + +static INLINE void +st_init_drawpixels_functions(struct dd_function_table *functions) +{ +} + +static INLINE void +st_destroy_drawpix(struct st_context *st) +{ +} + +#endif /* FEATURE_drawpix */ + +#endif /* ST_CB_DRAWPIXELS_H */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_drawtex.h b/mesalib/src/mesa/state_tracker/st_cb_drawtex.h index 0937b48f8..455da04e6 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_drawtex.h +++ b/mesalib/src/mesa/state_tracker/st_cb_drawtex.h @@ -1,41 +1,41 @@ -/************************************************************************** - * - * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - **************************************************************************/ - - -#ifndef ST_CB_DRAWTEX_H -#define ST_CB_DRAWTEX_H - - -#include "main/compiler.h" -#include "main/mfeatures.h" - -struct dd_function_table; -struct st_context; - -#if FEATURE_OES_draw_texture - -extern void -st_init_drawtex_functions(struct dd_function_table *functions); - -extern void -st_destroy_drawtex(struct st_context *st); - -#else - -static INLINE void -st_init_drawtex_functions(struct dd_function_table *functions) -{ -} - -static INLINE void -st_destroy_drawtex(struct st_context *st) -{ -} - -#endif /* FEATURE_OES_draw_texture */ - -#endif /* ST_CB_DRAWTEX_H */ +/************************************************************************** + * + * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + **************************************************************************/ + + +#ifndef ST_CB_DRAWTEX_H +#define ST_CB_DRAWTEX_H + + +#include "main/compiler.h" +#include "main/mfeatures.h" + +struct dd_function_table; +struct st_context; + +#if FEATURE_OES_draw_texture + +extern void +st_init_drawtex_functions(struct dd_function_table *functions); + +extern void +st_destroy_drawtex(struct st_context *st); + +#else + +static INLINE void +st_init_drawtex_functions(struct dd_function_table *functions) +{ +} + +static INLINE void +st_destroy_drawtex(struct st_context *st) +{ +} + +#endif /* FEATURE_OES_draw_texture */ + +#endif /* ST_CB_DRAWTEX_H */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_eglimage.h b/mesalib/src/mesa/state_tracker/st_cb_eglimage.h index b325b0608..48567ed9d 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_eglimage.h +++ b/mesalib/src/mesa/state_tracker/st_cb_eglimage.h @@ -1,51 +1,51 @@ -/* - * Mesa 3-D graphics library - * Version: 7.9 - * - * Copyright (C) 2010 LunarG Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - * - * Authors: - * Chia-I Wu - */ - -#ifndef ST_CB_EGLIMAGE_H -#define ST_CB_EGLIMAGE_H - -#include "main/compiler.h" -#include "main/mfeatures.h" - -struct dd_function_table; - -#if FEATURE_OES_EGL_image - -extern void -st_init_eglimage_functions(struct dd_function_table *functions); - -#else - -static INLINE void -st_init_eglimage_functions(struct dd_function_table *functions) -{ -} - -#endif - -#endif /* ST_CB_EGLIMAGE_H */ +/* + * Mesa 3-D graphics library + * Version: 7.9 + * + * Copyright (C) 2010 LunarG Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Chia-I Wu + */ + +#ifndef ST_CB_EGLIMAGE_H +#define ST_CB_EGLIMAGE_H + +#include "main/compiler.h" +#include "main/mfeatures.h" + +struct dd_function_table; + +#if FEATURE_OES_EGL_image + +extern void +st_init_eglimage_functions(struct dd_function_table *functions); + +#else + +static INLINE void +st_init_eglimage_functions(struct dd_function_table *functions) +{ +} + +#endif + +#endif /* ST_CB_EGLIMAGE_H */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_feedback.c b/mesalib/src/mesa/state_tracker/st_cb_feedback.c index d390ba192..9b85a39be 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_feedback.c +++ b/mesalib/src/mesa/state_tracker/st_cb_feedback.c @@ -1,309 +1,309 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -/** - * GL_SELECT and GL_FEEDBACK render modes. - * Basically, we use a private instance of the 'draw' module for doing - * selection/feedback. It would be nice to use the transform_feedback - * hardware feature, but it's defined as happening pre-clip and we want - * post-clipped primitives. Also, there's concerns about the efficiency - * of using the hardware for this anyway. - * - * Authors: - * Brian Paul - */ - -#include "main/imports.h" -#include "main/context.h" -#include "main/feedback.h" -#include "main/mfeatures.h" - -#include "vbo/vbo.h" - -#include "st_context.h" -#include "st_draw.h" -#include "st_cb_feedback.h" - -#include "pipe/p_context.h" -#include "pipe/p_defines.h" - -#include "draw/draw_context.h" -#include "draw/draw_pipe.h" - - -#if FEATURE_feedback - -/** - * This is actually used for both feedback and selection. - */ -struct feedback_stage -{ - struct draw_stage stage; /**< Base class */ - struct gl_context *ctx; /**< Rendering context */ - GLboolean reset_stipple_counter; -}; - - -/********************************************************************** - * GL Feedback functions - **********************************************************************/ - -static INLINE struct feedback_stage * -feedback_stage( struct draw_stage *stage ) -{ - return (struct feedback_stage *)stage; -} - - -static void -feedback_vertex(struct gl_context *ctx, const struct draw_context *draw, - const struct vertex_header *v) -{ - const struct st_context *st = st_context(ctx); - GLfloat win[4]; - const GLfloat *color, *texcoord; - GLuint slot; - - /* Recall that Y=0=Top of window for Gallium wincoords */ - win[0] = v->data[0][0]; - win[1] = ctx->DrawBuffer->Height - v->data[0][1]; - win[2] = v->data[0][2]; - win[3] = 1.0F / v->data[0][3]; - - /* XXX - * When we compute vertex layout, save info about position of the - * color and texcoord attribs to use here. - */ - - slot = st->vertex_result_to_slot[VERT_RESULT_COL0]; - if (slot != ~0U) - color = v->data[slot]; - else - color = ctx->Current.Attrib[VERT_ATTRIB_COLOR0]; - - slot = st->vertex_result_to_slot[VERT_RESULT_TEX0]; - if (slot != ~0U) - texcoord = v->data[slot]; - else - texcoord = ctx->Current.Attrib[VERT_ATTRIB_TEX0]; - - _mesa_feedback_vertex(ctx, win, color, texcoord); -} - - -static void -feedback_tri( struct draw_stage *stage, struct prim_header *prim ) -{ - struct feedback_stage *fs = feedback_stage(stage); - struct draw_context *draw = stage->draw; - _mesa_feedback_token(fs->ctx, (GLfloat) GL_POLYGON_TOKEN); - _mesa_feedback_token(fs->ctx, (GLfloat) 3); /* three vertices */ - feedback_vertex(fs->ctx, draw, prim->v[0]); - feedback_vertex(fs->ctx, draw, prim->v[1]); - feedback_vertex(fs->ctx, draw, prim->v[2]); -} - - -static void -feedback_line( struct draw_stage *stage, struct prim_header *prim ) -{ - struct feedback_stage *fs = feedback_stage(stage); - struct draw_context *draw = stage->draw; - if (fs->reset_stipple_counter) { - _mesa_feedback_token(fs->ctx, (GLfloat) GL_LINE_RESET_TOKEN); - fs->reset_stipple_counter = GL_FALSE; - } - else { - _mesa_feedback_token(fs->ctx, (GLfloat) GL_LINE_TOKEN); - } - feedback_vertex(fs->ctx, draw, prim->v[0]); - feedback_vertex(fs->ctx, draw, prim->v[1]); -} - - -static void -feedback_point( struct draw_stage *stage, struct prim_header *prim ) -{ - struct feedback_stage *fs = feedback_stage(stage); - struct draw_context *draw = stage->draw; - _mesa_feedback_token(fs->ctx, (GLfloat) GL_POINT_TOKEN); - feedback_vertex(fs->ctx, draw, prim->v[0]); -} - - -static void -feedback_flush( struct draw_stage *stage, unsigned flags ) -{ - /* no-op */ -} - - -static void -feedback_reset_stipple_counter( struct draw_stage *stage ) -{ - struct feedback_stage *fs = feedback_stage(stage); - fs->reset_stipple_counter = GL_TRUE; -} - - -static void -feedback_destroy( struct draw_stage *stage ) -{ - /* no-op */ -} - -/** - * Create GL feedback drawing stage. - */ -static struct draw_stage * -draw_glfeedback_stage(struct gl_context *ctx, struct draw_context *draw) -{ - struct feedback_stage *fs = ST_CALLOC_STRUCT(feedback_stage); - - fs->stage.draw = draw; - fs->stage.next = NULL; - fs->stage.point = feedback_point; - fs->stage.line = feedback_line; - fs->stage.tri = feedback_tri; - fs->stage.flush = feedback_flush; - fs->stage.reset_stipple_counter = feedback_reset_stipple_counter; - fs->stage.destroy = feedback_destroy; - fs->ctx = ctx; - - return &fs->stage; -} - - - -/********************************************************************** - * GL Selection functions - **********************************************************************/ - -static void -select_tri( struct draw_stage *stage, struct prim_header *prim ) -{ - struct feedback_stage *fs = feedback_stage(stage); - _mesa_update_hitflag( fs->ctx, prim->v[0]->data[0][2] ); - _mesa_update_hitflag( fs->ctx, prim->v[1]->data[0][2] ); - _mesa_update_hitflag( fs->ctx, prim->v[2]->data[0][2] ); -} - -static void -select_line( struct draw_stage *stage, struct prim_header *prim ) -{ - struct feedback_stage *fs = feedback_stage(stage); - _mesa_update_hitflag( fs->ctx, prim->v[0]->data[0][2] ); - _mesa_update_hitflag( fs->ctx, prim->v[1]->data[0][2] ); -} - - -static void -select_point( struct draw_stage *stage, struct prim_header *prim ) -{ - struct feedback_stage *fs = feedback_stage(stage); - _mesa_update_hitflag( fs->ctx, prim->v[0]->data[0][2] ); -} - - -static void -select_flush( struct draw_stage *stage, unsigned flags ) -{ - /* no-op */ -} - - -static void -select_reset_stipple_counter( struct draw_stage *stage ) -{ - /* no-op */ -} - -static void -select_destroy( struct draw_stage *stage ) -{ - /* no-op */ -} - - -/** - * Create GL selection mode drawing stage. - */ -static struct draw_stage * -draw_glselect_stage(struct gl_context *ctx, struct draw_context *draw) -{ - struct feedback_stage *fs = ST_CALLOC_STRUCT(feedback_stage); - - fs->stage.draw = draw; - fs->stage.next = NULL; - fs->stage.point = select_point; - fs->stage.line = select_line; - fs->stage.tri = select_tri; - fs->stage.flush = select_flush; - fs->stage.reset_stipple_counter = select_reset_stipple_counter; - fs->stage.destroy = select_destroy; - fs->ctx = ctx; - - return &fs->stage; -} - - -static void -st_RenderMode(struct gl_context *ctx, GLenum newMode ) -{ - struct st_context *st = st_context(ctx); - struct draw_context *draw = st->draw; - - if (newMode == GL_RENDER) { - /* restore normal VBO draw function */ - vbo_set_draw_func(ctx, st_draw_vbo); - } - else if (newMode == GL_SELECT) { - if (!st->selection_stage) - st->selection_stage = draw_glselect_stage(ctx, draw); - draw_set_rasterize_stage(draw, st->selection_stage); - /* Plug in new vbo draw function */ - vbo_set_draw_func(ctx, st_feedback_draw_vbo); - } - else { - if (!st->feedback_stage) - st->feedback_stage = draw_glfeedback_stage(ctx, draw); - draw_set_rasterize_stage(draw, st->feedback_stage); - /* Plug in new vbo draw function */ - vbo_set_draw_func(ctx, st_feedback_draw_vbo); - /* need to generate/use a vertex program that emits pos/color/tex */ - st->dirty.st |= ST_NEW_VERTEX_PROGRAM; - } -} - - - -void st_init_feedback_functions(struct dd_function_table *functions) -{ - functions->RenderMode = st_RenderMode; -} - -#endif /* FEATURE_feedback */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +/** + * GL_SELECT and GL_FEEDBACK render modes. + * Basically, we use a private instance of the 'draw' module for doing + * selection/feedback. It would be nice to use the transform_feedback + * hardware feature, but it's defined as happening pre-clip and we want + * post-clipped primitives. Also, there's concerns about the efficiency + * of using the hardware for this anyway. + * + * Authors: + * Brian Paul + */ + +#include "main/imports.h" +#include "main/context.h" +#include "main/feedback.h" +#include "main/mfeatures.h" + +#include "vbo/vbo.h" + +#include "st_context.h" +#include "st_draw.h" +#include "st_cb_feedback.h" + +#include "pipe/p_context.h" +#include "pipe/p_defines.h" + +#include "draw/draw_context.h" +#include "draw/draw_pipe.h" + + +#if FEATURE_feedback + +/** + * This is actually used for both feedback and selection. + */ +struct feedback_stage +{ + struct draw_stage stage; /**< Base class */ + struct gl_context *ctx; /**< Rendering context */ + GLboolean reset_stipple_counter; +}; + + +/********************************************************************** + * GL Feedback functions + **********************************************************************/ + +static INLINE struct feedback_stage * +feedback_stage( struct draw_stage *stage ) +{ + return (struct feedback_stage *)stage; +} + + +static void +feedback_vertex(struct gl_context *ctx, const struct draw_context *draw, + const struct vertex_header *v) +{ + const struct st_context *st = st_context(ctx); + GLfloat win[4]; + const GLfloat *color, *texcoord; + GLuint slot; + + /* Recall that Y=0=Top of window for Gallium wincoords */ + win[0] = v->data[0][0]; + win[1] = ctx->DrawBuffer->Height - v->data[0][1]; + win[2] = v->data[0][2]; + win[3] = 1.0F / v->data[0][3]; + + /* XXX + * When we compute vertex layout, save info about position of the + * color and texcoord attribs to use here. + */ + + slot = st->vertex_result_to_slot[VERT_RESULT_COL0]; + if (slot != ~0U) + color = v->data[slot]; + else + color = ctx->Current.Attrib[VERT_ATTRIB_COLOR0]; + + slot = st->vertex_result_to_slot[VERT_RESULT_TEX0]; + if (slot != ~0U) + texcoord = v->data[slot]; + else + texcoord = ctx->Current.Attrib[VERT_ATTRIB_TEX0]; + + _mesa_feedback_vertex(ctx, win, color, texcoord); +} + + +static void +feedback_tri( struct draw_stage *stage, struct prim_header *prim ) +{ + struct feedback_stage *fs = feedback_stage(stage); + struct draw_context *draw = stage->draw; + _mesa_feedback_token(fs->ctx, (GLfloat) GL_POLYGON_TOKEN); + _mesa_feedback_token(fs->ctx, (GLfloat) 3); /* three vertices */ + feedback_vertex(fs->ctx, draw, prim->v[0]); + feedback_vertex(fs->ctx, draw, prim->v[1]); + feedback_vertex(fs->ctx, draw, prim->v[2]); +} + + +static void +feedback_line( struct draw_stage *stage, struct prim_header *prim ) +{ + struct feedback_stage *fs = feedback_stage(stage); + struct draw_context *draw = stage->draw; + if (fs->reset_stipple_counter) { + _mesa_feedback_token(fs->ctx, (GLfloat) GL_LINE_RESET_TOKEN); + fs->reset_stipple_counter = GL_FALSE; + } + else { + _mesa_feedback_token(fs->ctx, (GLfloat) GL_LINE_TOKEN); + } + feedback_vertex(fs->ctx, draw, prim->v[0]); + feedback_vertex(fs->ctx, draw, prim->v[1]); +} + + +static void +feedback_point( struct draw_stage *stage, struct prim_header *prim ) +{ + struct feedback_stage *fs = feedback_stage(stage); + struct draw_context *draw = stage->draw; + _mesa_feedback_token(fs->ctx, (GLfloat) GL_POINT_TOKEN); + feedback_vertex(fs->ctx, draw, prim->v[0]); +} + + +static void +feedback_flush( struct draw_stage *stage, unsigned flags ) +{ + /* no-op */ +} + + +static void +feedback_reset_stipple_counter( struct draw_stage *stage ) +{ + struct feedback_stage *fs = feedback_stage(stage); + fs->reset_stipple_counter = GL_TRUE; +} + + +static void +feedback_destroy( struct draw_stage *stage ) +{ + /* no-op */ +} + +/** + * Create GL feedback drawing stage. + */ +static struct draw_stage * +draw_glfeedback_stage(struct gl_context *ctx, struct draw_context *draw) +{ + struct feedback_stage *fs = ST_CALLOC_STRUCT(feedback_stage); + + fs->stage.draw = draw; + fs->stage.next = NULL; + fs->stage.point = feedback_point; + fs->stage.line = feedback_line; + fs->stage.tri = feedback_tri; + fs->stage.flush = feedback_flush; + fs->stage.reset_stipple_counter = feedback_reset_stipple_counter; + fs->stage.destroy = feedback_destroy; + fs->ctx = ctx; + + return &fs->stage; +} + + + +/********************************************************************** + * GL Selection functions + **********************************************************************/ + +static void +select_tri( struct draw_stage *stage, struct prim_header *prim ) +{ + struct feedback_stage *fs = feedback_stage(stage); + _mesa_update_hitflag( fs->ctx, prim->v[0]->data[0][2] ); + _mesa_update_hitflag( fs->ctx, prim->v[1]->data[0][2] ); + _mesa_update_hitflag( fs->ctx, prim->v[2]->data[0][2] ); +} + +static void +select_line( struct draw_stage *stage, struct prim_header *prim ) +{ + struct feedback_stage *fs = feedback_stage(stage); + _mesa_update_hitflag( fs->ctx, prim->v[0]->data[0][2] ); + _mesa_update_hitflag( fs->ctx, prim->v[1]->data[0][2] ); +} + + +static void +select_point( struct draw_stage *stage, struct prim_header *prim ) +{ + struct feedback_stage *fs = feedback_stage(stage); + _mesa_update_hitflag( fs->ctx, prim->v[0]->data[0][2] ); +} + + +static void +select_flush( struct draw_stage *stage, unsigned flags ) +{ + /* no-op */ +} + + +static void +select_reset_stipple_counter( struct draw_stage *stage ) +{ + /* no-op */ +} + +static void +select_destroy( struct draw_stage *stage ) +{ + /* no-op */ +} + + +/** + * Create GL selection mode drawing stage. + */ +static struct draw_stage * +draw_glselect_stage(struct gl_context *ctx, struct draw_context *draw) +{ + struct feedback_stage *fs = ST_CALLOC_STRUCT(feedback_stage); + + fs->stage.draw = draw; + fs->stage.next = NULL; + fs->stage.point = select_point; + fs->stage.line = select_line; + fs->stage.tri = select_tri; + fs->stage.flush = select_flush; + fs->stage.reset_stipple_counter = select_reset_stipple_counter; + fs->stage.destroy = select_destroy; + fs->ctx = ctx; + + return &fs->stage; +} + + +static void +st_RenderMode(struct gl_context *ctx, GLenum newMode ) +{ + struct st_context *st = st_context(ctx); + struct draw_context *draw = st->draw; + + if (newMode == GL_RENDER) { + /* restore normal VBO draw function */ + vbo_set_draw_func(ctx, st_draw_vbo); + } + else if (newMode == GL_SELECT) { + if (!st->selection_stage) + st->selection_stage = draw_glselect_stage(ctx, draw); + draw_set_rasterize_stage(draw, st->selection_stage); + /* Plug in new vbo draw function */ + vbo_set_draw_func(ctx, st_feedback_draw_vbo); + } + else { + if (!st->feedback_stage) + st->feedback_stage = draw_glfeedback_stage(ctx, draw); + draw_set_rasterize_stage(draw, st->feedback_stage); + /* Plug in new vbo draw function */ + vbo_set_draw_func(ctx, st_feedback_draw_vbo); + /* need to generate/use a vertex program that emits pos/color/tex */ + st->dirty.st |= ST_NEW_VERTEX_PROGRAM; + } +} + + + +void st_init_feedback_functions(struct dd_function_table *functions) +{ + functions->RenderMode = st_RenderMode; +} + +#endif /* FEATURE_feedback */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_feedback.h b/mesalib/src/mesa/state_tracker/st_cb_feedback.h index 3acb5b5b3..02e34e402 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_feedback.h +++ b/mesalib/src/mesa/state_tracker/st_cb_feedback.h @@ -1,52 +1,52 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_CB_FEEDBACK_H -#define ST_CB_FEEDBACK_H - - -#include "main/compiler.h" -#include "main/mfeatures.h" - -struct dd_function_table; - -#if FEATURE_feedback - -extern void -st_init_feedback_functions(struct dd_function_table *functions); - -#else - -static INLINE void -st_init_feedback_functions(struct dd_function_table *functions) -{ -} - -#endif /* FEATURE_feedback */ - -#endif /* ST_CB_FEEDBACK_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_CB_FEEDBACK_H +#define ST_CB_FEEDBACK_H + + +#include "main/compiler.h" +#include "main/mfeatures.h" + +struct dd_function_table; + +#if FEATURE_feedback + +extern void +st_init_feedback_functions(struct dd_function_table *functions); + +#else + +static INLINE void +st_init_feedback_functions(struct dd_function_table *functions) +{ +} + +#endif /* FEATURE_feedback */ + +#endif /* ST_CB_FEEDBACK_H */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_flush.h b/mesalib/src/mesa/state_tracker/st_cb_flush.h index 1fd0c56d7..598536ba0 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_flush.h +++ b/mesalib/src/mesa/state_tracker/st_cb_flush.h @@ -1,51 +1,51 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_CB_FLUSH_H -#define ST_CB_FLUSH_H - - -#include "pipe/p_compiler.h" - -struct dd_function_table; -struct pipe_fence_handle; -struct st_context; - -extern void -st_init_flush_functions(struct dd_function_table *functions); - -extern void -st_flush(struct st_context *st, - struct pipe_fence_handle **fence); - -extern void -st_finish(struct st_context *st); - - -#endif /* ST_CB_FLUSH_H */ - +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_CB_FLUSH_H +#define ST_CB_FLUSH_H + + +#include "pipe/p_compiler.h" + +struct dd_function_table; +struct pipe_fence_handle; +struct st_context; + +extern void +st_init_flush_functions(struct dd_function_table *functions); + +extern void +st_flush(struct st_context *st, + struct pipe_fence_handle **fence); + +extern void +st_finish(struct st_context *st); + + +#endif /* ST_CB_FLUSH_H */ + diff --git a/mesalib/src/mesa/state_tracker/st_cb_program.h b/mesalib/src/mesa/state_tracker/st_cb_program.h index 05d03e1e6..091a4439c 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_program.h +++ b/mesalib/src/mesa/state_tracker/st_cb_program.h @@ -1,38 +1,38 @@ -/************************************************************************** - * - * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_PROGRAM_H -#define ST_CB_PROGRAM_H - - -struct dd_function_table; - -extern void -st_init_program_functions(struct dd_function_table *functions); - - -#endif +/************************************************************************** + * + * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_PROGRAM_H +#define ST_CB_PROGRAM_H + + +struct dd_function_table; + +extern void +st_init_program_functions(struct dd_function_table *functions); + + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_cb_queryobj.h b/mesalib/src/mesa/state_tracker/st_cb_queryobj.h index 638dba9ec..03f0be837 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_queryobj.h +++ b/mesalib/src/mesa/state_tracker/st_cb_queryobj.h @@ -1,70 +1,70 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_QUERYOBJ_H -#define ST_CB_QUERYOBJ_H - - -#include "main/mfeatures.h" -#include "main/mtypes.h" - -/** - * Subclass of gl_query_object - */ -struct st_query_object -{ - struct gl_query_object base; - struct pipe_query *pq; - unsigned type; /**< PIPE_QUERY_x */ -}; - - -/** - * Cast wrapper - */ -static INLINE struct st_query_object * -st_query_object(struct gl_query_object *q) -{ - return (struct st_query_object *) q; -} - - -#if FEATURE_queryobj - -extern void -st_init_query_functions(struct dd_function_table *functions); - -#else - -static INLINE void -st_init_query_functions(struct dd_function_table *functions) -{ -} - -#endif /* FEATURE_queryobj */ - -#endif +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_QUERYOBJ_H +#define ST_CB_QUERYOBJ_H + + +#include "main/mfeatures.h" +#include "main/mtypes.h" + +/** + * Subclass of gl_query_object + */ +struct st_query_object +{ + struct gl_query_object base; + struct pipe_query *pq; + unsigned type; /**< PIPE_QUERY_x */ +}; + + +/** + * Cast wrapper + */ +static INLINE struct st_query_object * +st_query_object(struct gl_query_object *q) +{ + return (struct st_query_object *) q; +} + + +#if FEATURE_queryobj + +extern void +st_init_query_functions(struct dd_function_table *functions); + +#else + +static INLINE void +st_init_query_functions(struct dd_function_table *functions) +{ +} + +#endif /* FEATURE_queryobj */ + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_cb_rasterpos.h b/mesalib/src/mesa/state_tracker/st_cb_rasterpos.h index c8fe9d135..b61411bd2 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_rasterpos.h +++ b/mesalib/src/mesa/state_tracker/st_cb_rasterpos.h @@ -1,50 +1,50 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_RASTERPOS_H -#define ST_CB_RASTERPOS_H - - -#include "main/compiler.h" -#include "main/mfeatures.h" - -struct dd_function_table; - -#if FEATURE_rastpos - -extern void st_init_rasterpos_functions(struct dd_function_table *functions); - -#else - -static INLINE void -st_init_rasterpos_functions(struct dd_function_table *functions) -{ -} - -#endif /* FEATURE_rastpos */ - -#endif /* ST_CB_RASTERPOS_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_RASTERPOS_H +#define ST_CB_RASTERPOS_H + + +#include "main/compiler.h" +#include "main/mfeatures.h" + +struct dd_function_table; + +#if FEATURE_rastpos + +extern void st_init_rasterpos_functions(struct dd_function_table *functions); + +#else + +static INLINE void +st_init_rasterpos_functions(struct dd_function_table *functions) +{ +} + +#endif /* FEATURE_rastpos */ + +#endif /* ST_CB_RASTERPOS_H */ diff --git a/mesalib/src/mesa/state_tracker/st_cb_strings.c b/mesalib/src/mesa/state_tracker/st_cb_strings.c index b382273ea..21323798f 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_strings.c +++ b/mesalib/src/mesa/state_tracker/st_cb_strings.c @@ -1,73 +1,73 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - /* - * Authors: - * Keith Whitwell - * Brian Paul - */ - -#include "main/glheader.h" -#include "main/macros.h" -#include "pipe/p_context.h" -#include "pipe/p_screen.h" -#include "util/u_string.h" -#include "st_context.h" -#include "st_cb_strings.h" - -#define ST_VERSION_STRING "0.4" - -static const GLubyte * -st_get_string(struct gl_context * ctx, GLenum name) -{ - struct st_context *st = st_context(ctx); - struct pipe_screen *screen = st->pipe->screen; - - switch (name) { - case GL_VENDOR: { - const char *vendor = screen->get_vendor( screen ); - util_snprintf(st->vendor, sizeof(st->vendor), "%s", vendor); - return (GLubyte *) st->vendor; - } - - case GL_RENDERER: - util_snprintf(st->renderer, sizeof(st->renderer), "Gallium %s on %s", - ST_VERSION_STRING, - screen->get_name( screen )); - - return (GLubyte *) st->renderer; - - default: - return NULL; - } -} - - -void st_init_string_functions(struct dd_function_table *functions) -{ - functions->GetString = st_get_string; -} +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + /* + * Authors: + * Keith Whitwell + * Brian Paul + */ + +#include "main/glheader.h" +#include "main/macros.h" +#include "pipe/p_context.h" +#include "pipe/p_screen.h" +#include "util/u_string.h" +#include "st_context.h" +#include "st_cb_strings.h" + +#define ST_VERSION_STRING "0.4" + +static const GLubyte * +st_get_string(struct gl_context * ctx, GLenum name) +{ + struct st_context *st = st_context(ctx); + struct pipe_screen *screen = st->pipe->screen; + + switch (name) { + case GL_VENDOR: { + const char *vendor = screen->get_vendor( screen ); + util_snprintf(st->vendor, sizeof(st->vendor), "%s", vendor); + return (GLubyte *) st->vendor; + } + + case GL_RENDERER: + util_snprintf(st->renderer, sizeof(st->renderer), "Gallium %s on %s", + ST_VERSION_STRING, + screen->get_name( screen )); + + return (GLubyte *) st->renderer; + + default: + return NULL; + } +} + + +void st_init_string_functions(struct dd_function_table *functions) +{ + functions->GetString = st_get_string; +} diff --git a/mesalib/src/mesa/state_tracker/st_cb_strings.h b/mesalib/src/mesa/state_tracker/st_cb_strings.h index 5b4fc2a72..92d5d2d9b 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_strings.h +++ b/mesalib/src/mesa/state_tracker/st_cb_strings.h @@ -1,40 +1,40 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_CB_STRINGS_H -#define ST_CB_STRINGS_H - - -struct dd_function_table; - -extern void -st_init_string_functions(struct dd_function_table *functions); - - -#endif /* ST_CB_CLEAR_H */ - +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_CB_STRINGS_H +#define ST_CB_STRINGS_H + + +struct dd_function_table; + +extern void +st_init_string_functions(struct dd_function_table *functions); + + +#endif /* ST_CB_CLEAR_H */ + diff --git a/mesalib/src/mesa/state_tracker/st_cb_syncobj.c b/mesalib/src/mesa/state_tracker/st_cb_syncobj.c index 7e243561a..d575a8497 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_syncobj.c +++ b/mesalib/src/mesa/state_tracker/st_cb_syncobj.c @@ -1,122 +1,122 @@ -/************************************************************************** - * - * Copyright 2011 Marek Olšák - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - /* - * Authors: - * Marek Olšák - */ - -#include "main/glheader.h" -#include "main/macros.h" -#include "pipe/p_context.h" -#include "pipe/p_screen.h" -#include "st_context.h" -#include "st_cb_syncobj.h" - -struct st_sync_object { - struct gl_sync_object b; - - struct pipe_fence_handle *fence; -}; - - -static struct gl_sync_object * st_new_sync_object(struct gl_context *ctx, - GLenum type) -{ - if (type == GL_SYNC_FENCE) - return (struct gl_sync_object*)CALLOC_STRUCT(st_sync_object); - else - return NULL; -} - -static void st_delete_sync_object(struct gl_context *ctx, - struct gl_sync_object *obj) -{ - struct pipe_screen *screen = st_context(ctx)->pipe->screen; - struct st_sync_object *so = (struct st_sync_object*)obj; - - screen->fence_reference(screen, &so->fence, NULL); - FREE(so); -} - -static void st_fence_sync(struct gl_context *ctx, struct gl_sync_object *obj, - GLenum condition, GLbitfield flags) -{ - struct pipe_context *pipe = st_context(ctx)->pipe; - struct st_sync_object *so = (struct st_sync_object*)obj; - - assert(condition == GL_SYNC_GPU_COMMANDS_COMPLETE && flags == 0); - assert(so->fence == NULL); - - pipe->flush(pipe, &so->fence); -} - -static void st_check_sync(struct gl_context *ctx, struct gl_sync_object *obj) -{ - struct pipe_screen *screen = st_context(ctx)->pipe->screen; - struct st_sync_object *so = (struct st_sync_object*)obj; - - if (so->fence && screen->fence_signalled(screen, so->fence)) { - screen->fence_reference(screen, &so->fence, NULL); - so->b.StatusFlag = GL_TRUE; - } -} - -static void st_client_wait_sync(struct gl_context *ctx, - struct gl_sync_object *obj, - GLbitfield flags, GLuint64 timeout) -{ - struct pipe_screen *screen = st_context(ctx)->pipe->screen; - struct st_sync_object *so = (struct st_sync_object*)obj; - - /* We don't care about GL_SYNC_FLUSH_COMMANDS_BIT, because flush is - * already called when creating a fence. */ - - if (so->fence && - screen->fence_finish(screen, so->fence, timeout)) { - screen->fence_reference(screen, &so->fence, NULL); - so->b.StatusFlag = GL_TRUE; - } -} - -static void st_server_wait_sync(struct gl_context *ctx, - struct gl_sync_object *obj, - GLbitfield flags, GLuint64 timeout) -{ - /* NO-OP. - * Neither Gallium nor DRM interfaces support blocking on the GPU. */ -} - -void st_init_syncobj_functions(struct dd_function_table *functions) -{ - functions->NewSyncObject = st_new_sync_object; - functions->FenceSync = st_fence_sync; - functions->DeleteSyncObject = st_delete_sync_object; - functions->CheckSync = st_check_sync; - functions->ClientWaitSync = st_client_wait_sync; - functions->ServerWaitSync = st_server_wait_sync; -} +/************************************************************************** + * + * Copyright 2011 Marek Olšák + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + /* + * Authors: + * Marek Olšák + */ + +#include "main/glheader.h" +#include "main/macros.h" +#include "pipe/p_context.h" +#include "pipe/p_screen.h" +#include "st_context.h" +#include "st_cb_syncobj.h" + +struct st_sync_object { + struct gl_sync_object b; + + struct pipe_fence_handle *fence; +}; + + +static struct gl_sync_object * st_new_sync_object(struct gl_context *ctx, + GLenum type) +{ + if (type == GL_SYNC_FENCE) + return (struct gl_sync_object*)CALLOC_STRUCT(st_sync_object); + else + return NULL; +} + +static void st_delete_sync_object(struct gl_context *ctx, + struct gl_sync_object *obj) +{ + struct pipe_screen *screen = st_context(ctx)->pipe->screen; + struct st_sync_object *so = (struct st_sync_object*)obj; + + screen->fence_reference(screen, &so->fence, NULL); + FREE(so); +} + +static void st_fence_sync(struct gl_context *ctx, struct gl_sync_object *obj, + GLenum condition, GLbitfield flags) +{ + struct pipe_context *pipe = st_context(ctx)->pipe; + struct st_sync_object *so = (struct st_sync_object*)obj; + + assert(condition == GL_SYNC_GPU_COMMANDS_COMPLETE && flags == 0); + assert(so->fence == NULL); + + pipe->flush(pipe, &so->fence); +} + +static void st_check_sync(struct gl_context *ctx, struct gl_sync_object *obj) +{ + struct pipe_screen *screen = st_context(ctx)->pipe->screen; + struct st_sync_object *so = (struct st_sync_object*)obj; + + if (so->fence && screen->fence_signalled(screen, so->fence)) { + screen->fence_reference(screen, &so->fence, NULL); + so->b.StatusFlag = GL_TRUE; + } +} + +static void st_client_wait_sync(struct gl_context *ctx, + struct gl_sync_object *obj, + GLbitfield flags, GLuint64 timeout) +{ + struct pipe_screen *screen = st_context(ctx)->pipe->screen; + struct st_sync_object *so = (struct st_sync_object*)obj; + + /* We don't care about GL_SYNC_FLUSH_COMMANDS_BIT, because flush is + * already called when creating a fence. */ + + if (so->fence && + screen->fence_finish(screen, so->fence, timeout)) { + screen->fence_reference(screen, &so->fence, NULL); + so->b.StatusFlag = GL_TRUE; + } +} + +static void st_server_wait_sync(struct gl_context *ctx, + struct gl_sync_object *obj, + GLbitfield flags, GLuint64 timeout) +{ + /* NO-OP. + * Neither Gallium nor DRM interfaces support blocking on the GPU. */ +} + +void st_init_syncobj_functions(struct dd_function_table *functions) +{ + functions->NewSyncObject = st_new_sync_object; + functions->FenceSync = st_fence_sync; + functions->DeleteSyncObject = st_delete_sync_object; + functions->CheckSync = st_check_sync; + functions->ClientWaitSync = st_client_wait_sync; + functions->ServerWaitSync = st_server_wait_sync; +} diff --git a/mesalib/src/mesa/state_tracker/st_cb_texture.c b/mesalib/src/mesa/state_tracker/st_cb_texture.c index b24f9a1ce..9d1b7f672 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_texture.c +++ b/mesalib/src/mesa/state_tracker/st_cb_texture.c @@ -530,62 +530,31 @@ prep_teximage(struct gl_context *ctx, struct gl_texture_image *texImage, static void -st_TexImage3D(struct gl_context * ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint depth, - GLint border, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *unpack) +st_TexImage(struct gl_context * ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLint width, GLint height, GLint depth, GLint border, + GLenum format, GLenum type, const void *pixels, + const struct gl_pixelstore_attrib *unpack) { prep_teximage(ctx, texImage, internalFormat, width, height, depth, border, format, type); - _mesa_store_teximage3d(ctx, texImage, internalFormat, width, height, depth, - border, format, type, pixels, unpack); -} - - -static void -st_TexImage2D(struct gl_context * ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint border, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *unpack) -{ - prep_teximage(ctx, texImage, internalFormat, width, height, 1, border, - format, type); - _mesa_store_teximage2d(ctx, texImage, internalFormat, width, height, - border, format, type, pixels, unpack); -} - - -static void -st_TexImage1D(struct gl_context * ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint border, - GLenum format, GLenum type, const void *pixels, - const struct gl_pixelstore_attrib *unpack) -{ - prep_teximage(ctx, texImage, internalFormat, width, 1, 1, border, - format, type); - _mesa_store_teximage1d(ctx, texImage, internalFormat, width, - border, format, type, pixels, unpack); + _mesa_store_teximage(ctx, dims, texImage, internalFormat, width, height, depth, + border, format, type, pixels, unpack); } static void -st_CompressedTexImage2D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint internalFormat, - GLint width, GLint height, GLint border, - GLsizei imageSize, const GLvoid *data) +st_CompressedTexImage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint internalFormat, + GLint width, GLint height, GLint border, GLint depth, + GLsizei imageSize, const GLvoid *data) { - prep_teximage(ctx, texImage, internalFormat, width, 1, 1, border, + prep_teximage(ctx, texImage, internalFormat, width, height, depth, border, GL_NONE, GL_NONE); - _mesa_store_compressed_teximage2d(ctx, texImage, internalFormat, width, - height, border, imageSize, data); + _mesa_store_compressed_teximage(ctx, dims, texImage, internalFormat, width, + height, depth, border, imageSize, data); } @@ -959,12 +928,11 @@ compatible_src_dst_formats(struct gl_context *ctx, * Note: srcY=0=Bottom of renderbuffer (GL convention) */ static void -st_copy_texsubimage(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint destX, GLint destY, GLint destZ, - struct gl_renderbuffer *rb, - GLint srcX, GLint srcY, - GLsizei width, GLsizei height) +st_CopyTexSubImage(struct gl_context *ctx, GLuint dims, + struct gl_texture_image *texImage, + GLint destX, GLint destY, GLint destZ, + struct gl_renderbuffer *rb, + GLint srcX, GLint srcY, GLsizei width, GLsizei height) { struct st_texture_image *stImage = st_texture_image(texImage); const GLenum texBaseFormat = texImage->_BaseFormat; @@ -1127,49 +1095,6 @@ fallback: } - -static void -st_CopyTexSubImage1D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, GLsizei width) -{ - const GLint yoffset = 0, zoffset = 0; - const GLsizei height = 1; - st_copy_texsubimage(ctx, texImage, - xoffset, yoffset, zoffset, /* destX,Y,Z */ - rb, x, y, width, height); /* src X, Y, size */ -} - - -static void -st_CopyTexSubImage2D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, GLsizei width, GLsizei height) -{ - const GLint zoffset = 0; - st_copy_texsubimage(ctx, texImage, - xoffset, yoffset, zoffset, /* destX,Y,Z */ - rb, x, y, width, height); /* src X, Y, size */ -} - - -static void -st_CopyTexSubImage3D(struct gl_context *ctx, - struct gl_texture_image *texImage, - GLint xoffset, GLint yoffset, GLint zoffset, - struct gl_renderbuffer *rb, - GLint x, GLint y, GLsizei width, GLsizei height) -{ - st_copy_texsubimage(ctx, texImage, - xoffset, yoffset, zoffset, /* destX,Y,Z */ - rb, x, y, width, height); /* src X, Y, size */ -} - - /** * Copy image data from stImage into the texture object 'stObj' at level * 'dstLevel'. @@ -1427,24 +1352,16 @@ void st_init_texture_functions(struct dd_function_table *functions) { functions->ChooseTextureFormat = st_ChooseTextureFormat; - functions->TexImage1D = st_TexImage1D; - functions->TexImage2D = st_TexImage2D; - functions->TexImage3D = st_TexImage3D; - functions->TexSubImage1D = _mesa_store_texsubimage1d; - functions->TexSubImage2D = _mesa_store_texsubimage2d; - functions->TexSubImage3D = _mesa_store_texsubimage3d; - functions->CompressedTexSubImage1D = _mesa_store_compressed_texsubimage1d; - functions->CompressedTexSubImage2D = _mesa_store_compressed_texsubimage2d; - functions->CompressedTexSubImage3D = _mesa_store_compressed_texsubimage3d; - functions->CopyTexSubImage1D = st_CopyTexSubImage1D; - functions->CopyTexSubImage2D = st_CopyTexSubImage2D; - functions->CopyTexSubImage3D = st_CopyTexSubImage3D; + functions->TexImage = st_TexImage; + functions->TexSubImage = _mesa_store_texsubimage; + functions->CompressedTexSubImage = _mesa_store_compressed_texsubimage; + functions->CopyTexSubImage = st_CopyTexSubImage; functions->GenerateMipmap = st_generate_mipmap; functions->GetTexImage = st_GetTexImage; /* compressed texture functions */ - functions->CompressedTexImage2D = st_CompressedTexImage2D; + functions->CompressedTexImage = st_CompressedTexImage; functions->GetCompressedTexImage = _mesa_get_compressed_teximage; functions->NewTextureObject = st_NewTextureObject; diff --git a/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c b/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c index 36668d343..2f1a22b58 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c +++ b/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.c @@ -1,60 +1,60 @@ -/************************************************************************** - * - * Copyright 2011 Marek Olšák - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -/** - * glTextureBarrierNV function - * - * \author Marek Olšák - */ - - -#include "main/imports.h" -#include "main/context.h" - -#include "pipe/p_context.h" -#include "pipe/p_defines.h" -#include "st_context.h" -#include "st_cb_texturebarrier.h" - - -/** - * Called via ctx->Driver.TextureBarrier() - */ -static void -st_TextureBarrier(struct gl_context *ctx) -{ - struct pipe_context *pipe = st_context(ctx)->pipe; - - pipe->texture_barrier(pipe); -} - - -void st_init_texture_barrier_functions(struct dd_function_table *functions) -{ - functions->TextureBarrier = st_TextureBarrier; -} +/************************************************************************** + * + * Copyright 2011 Marek Olšák + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +/** + * glTextureBarrierNV function + * + * \author Marek Olšák + */ + + +#include "main/imports.h" +#include "main/context.h" + +#include "pipe/p_context.h" +#include "pipe/p_defines.h" +#include "st_context.h" +#include "st_cb_texturebarrier.h" + + +/** + * Called via ctx->Driver.TextureBarrier() + */ +static void +st_TextureBarrier(struct gl_context *ctx) +{ + struct pipe_context *pipe = st_context(ctx)->pipe; + + pipe->texture_barrier(pipe); +} + + +void st_init_texture_barrier_functions(struct dd_function_table *functions) +{ + functions->TextureBarrier = st_TextureBarrier; +} diff --git a/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.h b/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.h index dc94cc443..3b7d3776f 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.h +++ b/mesalib/src/mesa/state_tracker/st_cb_texturebarrier.h @@ -1,37 +1,37 @@ -/************************************************************************** - * - * Copyright 2011 Marek Olšák - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_TEXTUREBARRIER_H -#define ST_CB_TEXTUREBARRIER_H - - -struct dd_function_table; - -extern void st_init_texture_barrier_functions(struct dd_function_table *functions); - - -#endif +/************************************************************************** + * + * Copyright 2011 Marek Olšák + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL THE AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_TEXTUREBARRIER_H +#define ST_CB_TEXTUREBARRIER_H + + +struct dd_function_table; + +extern void st_init_texture_barrier_functions(struct dd_function_table *functions); + + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_cb_viewport.h b/mesalib/src/mesa/state_tracker/st_cb_viewport.h index 7307aba1b..bcfd7cb68 100644 --- a/mesalib/src/mesa/state_tracker/st_cb_viewport.h +++ b/mesalib/src/mesa/state_tracker/st_cb_viewport.h @@ -1,36 +1,36 @@ -/************************************************************************** - * - * Copyright 2009 VMware, Inc. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#ifndef ST_CB_VIEWPORT_H -#define ST_CB_VIEWPORT_H - -struct dd_function_table; - -extern void -st_init_viewport_functions(struct dd_function_table *functions); - -#endif /* ST_CB_VIEW_PORT_H */ +/************************************************************************** + * + * Copyright 2009 VMware, Inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#ifndef ST_CB_VIEWPORT_H +#define ST_CB_VIEWPORT_H + +struct dd_function_table; + +extern void +st_init_viewport_functions(struct dd_function_table *functions); + +#endif /* ST_CB_VIEW_PORT_H */ diff --git a/mesalib/src/mesa/state_tracker/st_debug.h b/mesalib/src/mesa/state_tracker/st_debug.h index 07864845f..4a060d775 100644 --- a/mesalib/src/mesa/state_tracker/st_debug.h +++ b/mesalib/src/mesa/state_tracker/st_debug.h @@ -1,72 +1,72 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_DEBUG_H -#define ST_DEBUG_H - -#include "pipe/p_compiler.h" -#include "util/u_debug.h" - -extern void -st_print_current(void); - - -#define DEBUG_MESA 0x1 -#define DEBUG_TGSI 0x2 -#define DEBUG_CONSTANTS 0x4 -#define DEBUG_PIPE 0x8 -#define DEBUG_TEX 0x10 -#define DEBUG_FALLBACK 0x20 -#define DEBUG_QUERY 0x40 -#define DEBUG_SCREEN 0x80 - -#ifdef DEBUG -extern int ST_DEBUG; -#define DBSTR(x) x -#else -#define ST_DEBUG 0 -#define DBSTR(x) "" -#endif - -void st_debug_init( void ); - -static INLINE void -ST_DBG( unsigned flag, const char *fmt, ... ) -{ - if (ST_DEBUG & flag) - { - va_list args; - - va_start( args, fmt ); - debug_vprintf( fmt, args ); - va_end( args ); - } -} - - -#endif /* ST_DEBUG_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_DEBUG_H +#define ST_DEBUG_H + +#include "pipe/p_compiler.h" +#include "util/u_debug.h" + +extern void +st_print_current(void); + + +#define DEBUG_MESA 0x1 +#define DEBUG_TGSI 0x2 +#define DEBUG_CONSTANTS 0x4 +#define DEBUG_PIPE 0x8 +#define DEBUG_TEX 0x10 +#define DEBUG_FALLBACK 0x20 +#define DEBUG_QUERY 0x40 +#define DEBUG_SCREEN 0x80 + +#ifdef DEBUG +extern int ST_DEBUG; +#define DBSTR(x) x +#else +#define ST_DEBUG 0 +#define DBSTR(x) "" +#endif + +void st_debug_init( void ); + +static INLINE void +ST_DBG( unsigned flag, const char *fmt, ... ) +{ + if (ST_DEBUG & flag) + { + va_list args; + + va_start( args, fmt ); + debug_vprintf( fmt, args ); + va_end( args ); + } +} + + +#endif /* ST_DEBUG_H */ diff --git a/mesalib/src/mesa/state_tracker/st_extensions.c b/mesalib/src/mesa/state_tracker/st_extensions.c index 953155f36..a9071f58c 100644 --- a/mesalib/src/mesa/state_tracker/st_extensions.c +++ b/mesalib/src/mesa/state_tracker/st_extensions.c @@ -592,6 +592,10 @@ void st_init_extensions(struct st_context *st) ctx->Extensions.S3_s3tc = GL_FALSE; } + if (ctx->Const.NativeIntegers) { + ctx->Extensions.ARB_shader_bit_encoding = GL_TRUE; + } + if (screen->get_shader_param(screen, PIPE_SHADER_GEOMETRY, PIPE_SHADER_CAP_MAX_INSTRUCTIONS) > 0) { #if 0 /* XXX re-enable when GLSL compiler again supports geometry shaders */ diff --git a/mesalib/src/mesa/state_tracker/st_extensions.h b/mesalib/src/mesa/state_tracker/st_extensions.h index f29f9e1cf..aa9b2b2b9 100644 --- a/mesalib/src/mesa/state_tracker/st_extensions.h +++ b/mesalib/src/mesa/state_tracker/st_extensions.h @@ -1,40 +1,40 @@ -/************************************************************************** - * - * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_EXTENSIONS_H -#define ST_EXTENSIONS_H - - -struct st_context; - -extern void st_init_limits(struct st_context *st); - -extern void st_init_extensions(struct st_context *st); - - -#endif /* ST_EXTENSIONS_H */ +/************************************************************************** + * + * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_EXTENSIONS_H +#define ST_EXTENSIONS_H + + +struct st_context; + +extern void st_init_limits(struct st_context *st); + +extern void st_init_extensions(struct st_context *st); + + +#endif /* ST_EXTENSIONS_H */ diff --git a/mesalib/src/mesa/state_tracker/st_gen_mipmap.h b/mesalib/src/mesa/state_tracker/st_gen_mipmap.h index 8a361a1f6..815c6a516 100644 --- a/mesalib/src/mesa/state_tracker/st_gen_mipmap.h +++ b/mesalib/src/mesa/state_tracker/st_gen_mipmap.h @@ -1,52 +1,52 @@ -/************************************************************************** - * - * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - - -#ifndef ST_GEN_MIPMAP_H -#define ST_GEN_MIPMAP_H - - -#include "main/glheader.h" - -struct gl_context; -struct gl_texture_object; -struct st_context; - -extern void -st_init_generate_mipmap(struct st_context *st); - - -extern void -st_destroy_generate_mipmap(struct st_context *st); - - -extern void -st_generate_mipmap(struct gl_context *ctx, GLenum target, - struct gl_texture_object *texObj); - - -#endif /* ST_GEN_MIPMAP_H */ +/************************************************************************** + * + * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + + +#ifndef ST_GEN_MIPMAP_H +#define ST_GEN_MIPMAP_H + + +#include "main/glheader.h" + +struct gl_context; +struct gl_texture_object; +struct st_context; + +extern void +st_init_generate_mipmap(struct st_context *st); + + +extern void +st_destroy_generate_mipmap(struct st_context *st); + + +extern void +st_generate_mipmap(struct gl_context *ctx, GLenum target, + struct gl_texture_object *texObj); + + +#endif /* ST_GEN_MIPMAP_H */ diff --git a/mesalib/src/mesa/state_tracker/st_gl_api.h b/mesalib/src/mesa/state_tracker/st_gl_api.h index 59782d0d5..55ad778c9 100644 --- a/mesalib/src/mesa/state_tracker/st_gl_api.h +++ b/mesalib/src/mesa/state_tracker/st_gl_api.h @@ -1,7 +1,7 @@ - -#ifndef ST_GL_API_H -#define ST_GL_API_H - -struct st_api *st_gl_api_create(void); - -#endif + +#ifndef ST_GL_API_H +#define ST_GL_API_H + +struct st_api *st_gl_api_create(void); + +#endif diff --git a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp index e8d60f80d..347a22f19 100644 --- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp +++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp @@ -1762,6 +1762,12 @@ glsl_to_tgsi_visitor::visit(ir_expression *ir) else emit(ir, TGSI_OPCODE_TRUNC, result_dst, op[0]); break; + case ir_unop_bitcast_f2i: + case ir_unop_bitcast_f2u: + case ir_unop_bitcast_i2f: + case ir_unop_bitcast_u2f: + result_src = op[0]; + break; case ir_unop_f2b: emit(ir, TGSI_OPCODE_SNE, result_dst, op[0], st_src_reg_for_float(0.0)); break; diff --git a/mesalib/src/mesa/state_tracker/st_manager.h b/mesalib/src/mesa/state_tracker/st_manager.h index ffe394cc4..6a9497839 100644 --- a/mesalib/src/mesa/state_tracker/st_manager.h +++ b/mesalib/src/mesa/state_tracker/st_manager.h @@ -1,52 +1,52 @@ -/* - * Mesa 3-D graphics library - * Version: 7.9 - * - * Copyright (C) 2010 LunarG Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER - * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING - * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER - * DEALINGS IN THE SOFTWARE. - * - * Authors: - * Chia-I Wu - */ - -#ifndef ST_MANAGER_H -#define ST_MANAGER_H - -#include "main/mtypes.h" - -#include "pipe/p_compiler.h" - -struct st_context; - -struct pipe_surface * -st_manager_get_egl_image_surface(struct st_context *st, - void *eglimg, unsigned usage); - -void -st_manager_flush_frontbuffer(struct st_context *st); - -void -st_manager_validate_framebuffers(struct st_context *st); - -boolean -st_manager_add_color_renderbuffer(struct st_context *st, struct gl_framebuffer *fb, - gl_buffer_index idx); - -#endif /* ST_MANAGER_H */ +/* + * Mesa 3-D graphics library + * Version: 7.9 + * + * Copyright (C) 2010 LunarG Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER + * DEALINGS IN THE SOFTWARE. + * + * Authors: + * Chia-I Wu + */ + +#ifndef ST_MANAGER_H +#define ST_MANAGER_H + +#include "main/mtypes.h" + +#include "pipe/p_compiler.h" + +struct st_context; + +struct pipe_surface * +st_manager_get_egl_image_surface(struct st_context *st, + void *eglimg, unsigned usage); + +void +st_manager_flush_frontbuffer(struct st_context *st); + +void +st_manager_validate_framebuffers(struct st_context *st); + +boolean +st_manager_add_color_renderbuffer(struct st_context *st, struct gl_framebuffer *fb, + gl_buffer_index idx); + +#endif /* ST_MANAGER_H */ diff --git a/mesalib/src/mesa/swrast/NOTES b/mesalib/src/mesa/swrast/NOTES index d6f991ee8..ea373aa12 100644 --- a/mesalib/src/mesa/swrast/NOTES +++ b/mesalib/src/mesa/swrast/NOTES @@ -1,55 +1,55 @@ -INTRODUCTION - -Mesa's native software rasterizer. This module provides the fallback -paths for rasterization operations and states that aren't accelerated -in hardware drivers, and as the full rasterization engine in software -drivers. - -The swrast module 'stands alone', relying only on interfaces to core -mesa and it's own driver interface. It knows nothing about the tnl or -other modules, allowing it to be used for fallback paths in future tnl -schemes without modification. - -As well as providing triangle/line/point rasterization functionality, -the module provides implementations of the pixel operations -(ReadPixels, etc), and texture operations (CopyTexSubImage) which may -be plugged in to the core Mesa driver interface where accelerated -versions of these operations are unavailable. - - -STATE - -To create and destroy the module: - - GLboolean _swrast_CreateContext( struct gl_context *ctx ); - void _swrast_DestroyContext( struct gl_context *ctx ); - -This module tracks state changes internally and maintains derived -values based on the current state. For this to work, the driver -ensure the following funciton is called whenever the state changes and -the swsetup module is 'awake': - - void _swrast_InvalidateState( struct gl_context *ctx, GLuint new_state ); - -There is no explicit call to put the swrast module to sleep. - - -CUSTOMIZATION - - void (*choose_point)( struct gl_context * ); - void (*choose_line)( struct gl_context * ); - void (*choose_triangle)( struct gl_context * ); - -Drivers may add additional triangle/line/point functions to swrast by -overriding these functions. It is necessary for the driver to be very -careful that it doesn't return an inappropriate function, eg a -rasterization function in feedback mode. See the X11 driver for -examples. - -DRIVER INTERFACE - -The swrast device driver provides swrast primarily with span- and -pixel- level interfaces to a framebuffer, with a few additional hooks -for locking and setting the read buffer. - +INTRODUCTION + +Mesa's native software rasterizer. This module provides the fallback +paths for rasterization operations and states that aren't accelerated +in hardware drivers, and as the full rasterization engine in software +drivers. + +The swrast module 'stands alone', relying only on interfaces to core +mesa and it's own driver interface. It knows nothing about the tnl or +other modules, allowing it to be used for fallback paths in future tnl +schemes without modification. + +As well as providing triangle/line/point rasterization functionality, +the module provides implementations of the pixel operations +(ReadPixels, etc), and texture operations (CopyTexSubImage) which may +be plugged in to the core Mesa driver interface where accelerated +versions of these operations are unavailable. + + +STATE + +To create and destroy the module: + + GLboolean _swrast_CreateContext( struct gl_context *ctx ); + void _swrast_DestroyContext( struct gl_context *ctx ); + +This module tracks state changes internally and maintains derived +values based on the current state. For this to work, the driver +ensure the following funciton is called whenever the state changes and +the swsetup module is 'awake': + + void _swrast_InvalidateState( struct gl_context *ctx, GLuint new_state ); + +There is no explicit call to put the swrast module to sleep. + + +CUSTOMIZATION + + void (*choose_point)( struct gl_context * ); + void (*choose_line)( struct gl_context * ); + void (*choose_triangle)( struct gl_context * ); + +Drivers may add additional triangle/line/point functions to swrast by +overriding these functions. It is necessary for the driver to be very +careful that it doesn't return an inappropriate function, eg a +rasterization function in feedback mode. See the X11 driver for +examples. + +DRIVER INTERFACE + +The swrast device driver provides swrast primarily with span- and +pixel- level interfaces to a framebuffer, with a few additional hooks +for locking and setting the read buffer. + See the definition of struct swrast_device_driver in swrast.h. \ No newline at end of file diff --git a/mesalib/src/mesa/swrast/s_aaline.h b/mesalib/src/mesa/swrast/s_aaline.h index 993c12eab..74d5518e1 100644 --- a/mesalib/src/mesa/swrast/s_aaline.h +++ b/mesalib/src/mesa/swrast/s_aaline.h @@ -1,38 +1,38 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_AALINE_H -#define S_AALINE_H - - -struct gl_context; - - -extern void -_swrast_choose_aa_line_function(struct gl_context *ctx); - - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_AALINE_H +#define S_AALINE_H + + +struct gl_context; + + +extern void +_swrast_choose_aa_line_function(struct gl_context *ctx); + + +#endif diff --git a/mesalib/src/mesa/swrast/s_aatriangle.h b/mesalib/src/mesa/swrast/s_aatriangle.h index dec637920..e40efb198 100644 --- a/mesalib/src/mesa/swrast/s_aatriangle.h +++ b/mesalib/src/mesa/swrast/s_aatriangle.h @@ -1,38 +1,38 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_AATRIANGLE_H -#define S_AATRIANGLE_H - - -struct gl_context; - - -extern void -_swrast_set_aa_triangle_function(struct gl_context *ctx); - - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_AATRIANGLE_H +#define S_AATRIANGLE_H + + +struct gl_context; + + +extern void +_swrast_set_aa_triangle_function(struct gl_context *ctx); + + +#endif diff --git a/mesalib/src/mesa/swrast/s_alpha.c b/mesalib/src/mesa/swrast/s_alpha.c index 798b506e4..df2181ba9 100644 --- a/mesalib/src/mesa/swrast/s_alpha.c +++ b/mesalib/src/mesa/swrast/s_alpha.c @@ -1,160 +1,160 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.2 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file swrast/s_alpha.c - * \brief Functions to apply alpha test. - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "main/colormac.h" -#include "main/macros.h" - -#include "s_alpha.h" -#include "s_context.h" - - -#define ALPHA_TEST(ALPHA, LOOP_CODE) \ -do { \ - switch (ctx->Color.AlphaFunc) { \ - case GL_LESS: \ - for (i = 0; i < n; i++) { \ - mask[i] &= (ALPHA < ref); \ - LOOP_CODE; \ - } \ - break; \ - case GL_LEQUAL: \ - for (i = 0; i < n; i++) { \ - mask[i] &= (ALPHA <= ref); \ - LOOP_CODE; \ - } \ - break; \ - case GL_GEQUAL: \ - for (i = 0; i < n; i++) { \ - mask[i] &= (ALPHA >= ref); \ - LOOP_CODE; \ - } \ - break; \ - case GL_GREATER: \ - for (i = 0; i < n; i++) { \ - mask[i] &= (ALPHA > ref); \ - LOOP_CODE; \ - } \ - break; \ - case GL_NOTEQUAL: \ - for (i = 0; i < n; i++) { \ - mask[i] &= (ALPHA != ref); \ - LOOP_CODE; \ - } \ - break; \ - case GL_EQUAL: \ - for (i = 0; i < n; i++) { \ - mask[i] &= (ALPHA == ref); \ - LOOP_CODE; \ - } \ - break; \ - default: \ - _mesa_problem(ctx, "Invalid alpha test in _swrast_alpha_test" ); \ - return 0; \ - } \ -} while (0) - - - -/** - * Perform the alpha test for an array of pixels. - * For pixels that fail the test, mask[i] will be set to 0. - * \return 0 if all pixels in the span failed the alpha test, - * 1 if one or more pixels passed the alpha test. - */ -GLint -_swrast_alpha_test(const struct gl_context *ctx, SWspan *span) -{ - const GLuint n = span->end; - GLubyte *mask = span->array->mask; - GLuint i; - - if (ctx->Color.AlphaFunc == GL_ALWAYS) { - /* do nothing */ - return 1; - } - else if (ctx->Color.AlphaFunc == GL_NEVER) { - /* All pixels failed - caller should check for this return value and - * act accordingly. - */ - span->writeAll = GL_FALSE; - return 0; - } - - if (span->arrayMask & SPAN_RGBA) { - /* Use array's alpha values */ - if (span->array->ChanType == GL_UNSIGNED_BYTE) { - GLubyte (*rgba)[4] = span->array->rgba8; - GLubyte ref; - CLAMPED_FLOAT_TO_UBYTE(ref, ctx->Color.AlphaRef); - ALPHA_TEST(rgba[i][ACOMP], ;); - } - else if (span->array->ChanType == GL_UNSIGNED_SHORT) { - GLushort (*rgba)[4] = span->array->rgba16; - GLushort ref; - CLAMPED_FLOAT_TO_USHORT(ref, ctx->Color.AlphaRef); - ALPHA_TEST(rgba[i][ACOMP], ;); - } - else { - GLfloat (*rgba)[4] = span->array->attribs[FRAG_ATTRIB_COL0]; - const GLfloat ref = ctx->Color.AlphaRef; - ALPHA_TEST(rgba[i][ACOMP], ;); - } - } - else { - /* Interpolate alpha values */ - ASSERT(span->interpMask & SPAN_RGBA); - if (span->array->ChanType == GL_UNSIGNED_BYTE) { - const GLfixed alphaStep = span->alphaStep; - GLfixed alpha = span->alpha; - GLubyte ref; - CLAMPED_FLOAT_TO_UBYTE(ref, ctx->Color.AlphaRef); - ALPHA_TEST(FixedToInt(alpha), alpha += alphaStep); - } - else if (span->array->ChanType == GL_UNSIGNED_SHORT) { - const GLfixed alphaStep = span->alphaStep; - GLfixed alpha = span->alpha; - GLushort ref; - CLAMPED_FLOAT_TO_USHORT(ref, ctx->Color.AlphaRef); - ALPHA_TEST(FixedToInt(alpha), alpha += alphaStep); - } - else { - const GLfloat alphaStep = FixedToFloat(span->alphaStep); - GLfloat alpha = FixedToFloat(span->alpha); - const GLfloat ref = ctx->Color.AlphaRef; - ALPHA_TEST(alpha, alpha += alphaStep); - } - } - - span->writeAll = GL_FALSE; - - /* XXX examine mask[] values? */ - return 1; -} +/* + * Mesa 3-D graphics library + * Version: 6.5.2 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file swrast/s_alpha.c + * \brief Functions to apply alpha test. + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "main/colormac.h" +#include "main/macros.h" + +#include "s_alpha.h" +#include "s_context.h" + + +#define ALPHA_TEST(ALPHA, LOOP_CODE) \ +do { \ + switch (ctx->Color.AlphaFunc) { \ + case GL_LESS: \ + for (i = 0; i < n; i++) { \ + mask[i] &= (ALPHA < ref); \ + LOOP_CODE; \ + } \ + break; \ + case GL_LEQUAL: \ + for (i = 0; i < n; i++) { \ + mask[i] &= (ALPHA <= ref); \ + LOOP_CODE; \ + } \ + break; \ + case GL_GEQUAL: \ + for (i = 0; i < n; i++) { \ + mask[i] &= (ALPHA >= ref); \ + LOOP_CODE; \ + } \ + break; \ + case GL_GREATER: \ + for (i = 0; i < n; i++) { \ + mask[i] &= (ALPHA > ref); \ + LOOP_CODE; \ + } \ + break; \ + case GL_NOTEQUAL: \ + for (i = 0; i < n; i++) { \ + mask[i] &= (ALPHA != ref); \ + LOOP_CODE; \ + } \ + break; \ + case GL_EQUAL: \ + for (i = 0; i < n; i++) { \ + mask[i] &= (ALPHA == ref); \ + LOOP_CODE; \ + } \ + break; \ + default: \ + _mesa_problem(ctx, "Invalid alpha test in _swrast_alpha_test" ); \ + return 0; \ + } \ +} while (0) + + + +/** + * Perform the alpha test for an array of pixels. + * For pixels that fail the test, mask[i] will be set to 0. + * \return 0 if all pixels in the span failed the alpha test, + * 1 if one or more pixels passed the alpha test. + */ +GLint +_swrast_alpha_test(const struct gl_context *ctx, SWspan *span) +{ + const GLuint n = span->end; + GLubyte *mask = span->array->mask; + GLuint i; + + if (ctx->Color.AlphaFunc == GL_ALWAYS) { + /* do nothing */ + return 1; + } + else if (ctx->Color.AlphaFunc == GL_NEVER) { + /* All pixels failed - caller should check for this return value and + * act accordingly. + */ + span->writeAll = GL_FALSE; + return 0; + } + + if (span->arrayMask & SPAN_RGBA) { + /* Use array's alpha values */ + if (span->array->ChanType == GL_UNSIGNED_BYTE) { + GLubyte (*rgba)[4] = span->array->rgba8; + GLubyte ref; + CLAMPED_FLOAT_TO_UBYTE(ref, ctx->Color.AlphaRef); + ALPHA_TEST(rgba[i][ACOMP], ;); + } + else if (span->array->ChanType == GL_UNSIGNED_SHORT) { + GLushort (*rgba)[4] = span->array->rgba16; + GLushort ref; + CLAMPED_FLOAT_TO_USHORT(ref, ctx->Color.AlphaRef); + ALPHA_TEST(rgba[i][ACOMP], ;); + } + else { + GLfloat (*rgba)[4] = span->array->attribs[FRAG_ATTRIB_COL0]; + const GLfloat ref = ctx->Color.AlphaRef; + ALPHA_TEST(rgba[i][ACOMP], ;); + } + } + else { + /* Interpolate alpha values */ + ASSERT(span->interpMask & SPAN_RGBA); + if (span->array->ChanType == GL_UNSIGNED_BYTE) { + const GLfixed alphaStep = span->alphaStep; + GLfixed alpha = span->alpha; + GLubyte ref; + CLAMPED_FLOAT_TO_UBYTE(ref, ctx->Color.AlphaRef); + ALPHA_TEST(FixedToInt(alpha), alpha += alphaStep); + } + else if (span->array->ChanType == GL_UNSIGNED_SHORT) { + const GLfixed alphaStep = span->alphaStep; + GLfixed alpha = span->alpha; + GLushort ref; + CLAMPED_FLOAT_TO_USHORT(ref, ctx->Color.AlphaRef); + ALPHA_TEST(FixedToInt(alpha), alpha += alphaStep); + } + else { + const GLfloat alphaStep = FixedToFloat(span->alphaStep); + GLfloat alpha = FixedToFloat(span->alpha); + const GLfloat ref = ctx->Color.AlphaRef; + ALPHA_TEST(alpha, alpha += alphaStep); + } + } + + span->writeAll = GL_FALSE; + + /* XXX examine mask[] values? */ + return 1; +} diff --git a/mesalib/src/mesa/swrast/s_alpha.h b/mesalib/src/mesa/swrast/s_alpha.h index ac5cd8dab..fca209a44 100644 --- a/mesalib/src/mesa/swrast/s_alpha.h +++ b/mesalib/src/mesa/swrast/s_alpha.h @@ -1,39 +1,39 @@ -/* - * Mesa 3-D graphics library - * Version: 4.1 - * - * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_ALPHA_H -#define S_ALPHA_H - - -#include "main/glheader.h" -#include "s_span.h" - -struct gl_context; - -extern GLint -_swrast_alpha_test( const struct gl_context *ctx, SWspan *span ); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 4.1 + * + * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_ALPHA_H +#define S_ALPHA_H + + +#include "main/glheader.h" +#include "s_span.h" + +struct gl_context; + +extern GLint +_swrast_alpha_test( const struct gl_context *ctx, SWspan *span ); + + +#endif diff --git a/mesalib/src/mesa/swrast/s_atifragshader.c b/mesalib/src/mesa/swrast/s_atifragshader.c index 6019eea5e..1eb026e00 100644 --- a/mesalib/src/mesa/swrast/s_atifragshader.c +++ b/mesalib/src/mesa/swrast/s_atifragshader.c @@ -1,604 +1,604 @@ -/* - * Copyright (C) 2004 David Airlie All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * DAVID AIRLIE BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/macros.h" -#include "main/atifragshader.h" -#include "swrast/s_atifragshader.h" -#include "swrast/s_context.h" - - -/** - * State for executing ATI fragment shader. - */ -struct atifs_machine -{ - GLfloat Registers[6][4]; /** six temporary registers */ - GLfloat PrevPassRegisters[6][4]; - GLfloat Inputs[2][4]; /** Primary, secondary input colors */ -}; - - - -/** - * Fetch a texel. - */ -static void -fetch_texel(struct gl_context * ctx, const GLfloat texcoord[4], GLfloat lambda, - GLuint unit, GLfloat color[4]) -{ - SWcontext *swrast = SWRAST_CONTEXT(ctx); - - /* XXX use a float-valued TextureSample routine here!!! */ - swrast->TextureSample[unit](ctx, ctx->Texture.Unit[unit]._Current, - 1, (const GLfloat(*)[4]) texcoord, - &lambda, (GLfloat (*)[4]) color); -} - -static void -apply_swizzle(GLfloat values[4], GLuint swizzle) -{ - GLfloat s, t, r, q; - - s = values[0]; - t = values[1]; - r = values[2]; - q = values[3]; - - switch (swizzle) { - case GL_SWIZZLE_STR_ATI: - values[0] = s; - values[1] = t; - values[2] = r; - break; - case GL_SWIZZLE_STQ_ATI: - values[0] = s; - values[1] = t; - values[2] = q; - break; - case GL_SWIZZLE_STR_DR_ATI: - values[0] = s / r; - values[1] = t / r; - values[2] = 1 / r; - break; - case GL_SWIZZLE_STQ_DQ_ATI: -/* make sure q is not 0 to avoid problems later with infinite values (texture lookup)? */ - if (q == 0.0F) - q = 0.000000001F; - values[0] = s / q; - values[1] = t / q; - values[2] = 1.0F / q; - break; - } - values[3] = 0.0; -} - -static void -apply_src_rep(GLint optype, GLuint rep, GLfloat * val) -{ - GLint i; - GLint start, end; - if (!rep) - return; - - start = optype ? 3 : 0; - end = 4; - - for (i = start; i < end; i++) { - switch (rep) { - case GL_RED: - val[i] = val[0]; - break; - case GL_GREEN: - val[i] = val[1]; - break; - case GL_BLUE: - val[i] = val[2]; - break; - case GL_ALPHA: - val[i] = val[3]; - break; - } - } -} - -static void -apply_src_mod(GLint optype, GLuint mod, GLfloat * val) -{ - GLint i; - GLint start, end; - - if (!mod) - return; - - start = optype ? 3 : 0; - end = 4; - - for (i = start; i < end; i++) { - if (mod & GL_COMP_BIT_ATI) - val[i] = 1 - val[i]; - - if (mod & GL_BIAS_BIT_ATI) - val[i] = val[i] - 0.5F; - - if (mod & GL_2X_BIT_ATI) - val[i] = 2 * val[i]; - - if (mod & GL_NEGATE_BIT_ATI) - val[i] = -val[i]; - } -} - -static void -apply_dst_mod(GLuint optype, GLuint mod, GLfloat * val) -{ - GLint i; - GLint has_sat = mod & GL_SATURATE_BIT_ATI; - GLint start, end; - - mod &= ~GL_SATURATE_BIT_ATI; - - start = optype ? 3 : 0; - end = optype ? 4 : 3; - - for (i = start; i < end; i++) { - switch (mod) { - case GL_2X_BIT_ATI: - val[i] = 2 * val[i]; - break; - case GL_4X_BIT_ATI: - val[i] = 4 * val[i]; - break; - case GL_8X_BIT_ATI: - val[i] = 8 * val[i]; - break; - case GL_HALF_BIT_ATI: - val[i] = val[i] * 0.5F; - break; - case GL_QUARTER_BIT_ATI: - val[i] = val[i] * 0.25F; - break; - case GL_EIGHTH_BIT_ATI: - val[i] = val[i] * 0.125F; - break; - } - - if (has_sat) { - if (val[i] < 0.0F) - val[i] = 0.0F; - else if (val[i] > 1.0F) - val[i] = 1.0F; - } - else { - if (val[i] < -8.0F) - val[i] = -8.0F; - else if (val[i] > 8.0F) - val[i] = 8.0F; - } - } -} - - -static void -write_dst_addr(GLuint optype, GLuint mod, GLuint mask, GLfloat * src, - GLfloat * dst) -{ - GLint i; - apply_dst_mod(optype, mod, src); - - if (optype == ATI_FRAGMENT_SHADER_COLOR_OP) { - if (mask) { - if (mask & GL_RED_BIT_ATI) - dst[0] = src[0]; - - if (mask & GL_GREEN_BIT_ATI) - dst[1] = src[1]; - - if (mask & GL_BLUE_BIT_ATI) - dst[2] = src[2]; - } - else { - for (i = 0; i < 3; i++) - dst[i] = src[i]; - } - } - else - dst[3] = src[3]; -} - -static void -finish_pass(struct atifs_machine *machine) -{ - GLint i; - - for (i = 0; i < 6; i++) { - COPY_4V(machine->PrevPassRegisters[i], machine->Registers[i]); - } -} - -struct ati_fs_opcode_st ati_fs_opcodes[] = { - {GL_ADD_ATI, 2}, - {GL_SUB_ATI, 2}, - {GL_MUL_ATI, 2}, - {GL_MAD_ATI, 3}, - {GL_LERP_ATI, 3}, - {GL_MOV_ATI, 1}, - {GL_CND_ATI, 3}, - {GL_CND0_ATI, 3}, - {GL_DOT2_ADD_ATI, 3}, - {GL_DOT3_ATI, 2}, - {GL_DOT4_ATI, 2} -}; - - - -static void -handle_pass_op(struct atifs_machine *machine, struct atifs_setupinst *texinst, - const SWspan *span, GLuint column, GLuint idx) -{ - GLuint swizzle = texinst->swizzle; - GLuint pass_tex = texinst->src; - - if (pass_tex >= GL_TEXTURE0_ARB && pass_tex <= GL_TEXTURE7_ARB) { - pass_tex -= GL_TEXTURE0_ARB; - COPY_4V(machine->Registers[idx], - span->array->attribs[FRAG_ATTRIB_TEX0 + pass_tex][column]); - } - else if (pass_tex >= GL_REG_0_ATI && pass_tex <= GL_REG_5_ATI) { - pass_tex -= GL_REG_0_ATI; - COPY_4V(machine->Registers[idx], machine->PrevPassRegisters[pass_tex]); - } - apply_swizzle(machine->Registers[idx], swizzle); - -} - -static void -handle_sample_op(struct gl_context * ctx, struct atifs_machine *machine, - struct atifs_setupinst *texinst, const SWspan *span, - GLuint column, GLuint idx) -{ -/* sample from unit idx using texinst->src as coords */ - GLuint swizzle = texinst->swizzle; - GLuint coord_source = texinst->src; - GLfloat tex_coords[4] = { 0 }; - - if (coord_source >= GL_TEXTURE0_ARB && coord_source <= GL_TEXTURE7_ARB) { - coord_source -= GL_TEXTURE0_ARB; - COPY_4V(tex_coords, - span->array->attribs[FRAG_ATTRIB_TEX0 + coord_source][column]); - } - else if (coord_source >= GL_REG_0_ATI && coord_source <= GL_REG_5_ATI) { - coord_source -= GL_REG_0_ATI; - COPY_4V(tex_coords, machine->PrevPassRegisters[coord_source]); - } - apply_swizzle(tex_coords, swizzle); - fetch_texel(ctx, tex_coords, 0.0F, idx, machine->Registers[idx]); -} - -#define SETUP_SRC_REG(optype, i, x) \ -do { \ - COPY_4V(src[optype][i], x); \ -} while (0) - - - -/** - * Execute the given fragment shader. - * NOTE: we do everything in single-precision floating point - * \param ctx - rendering context - * \param shader - the shader to execute - * \param machine - virtual machine state - * \param span - the SWspan we're operating on - * \param column - which pixel [i] we're operating on in the span - */ -static void -execute_shader(struct gl_context *ctx, const struct ati_fragment_shader *shader, - struct atifs_machine *machine, const SWspan *span, - GLuint column) -{ - GLuint pc; - struct atifs_instruction *inst; - struct atifs_setupinst *texinst; - GLint optype; - GLuint i; - GLint j, pass; - GLint dstreg; - GLfloat src[2][3][4]; - GLfloat zeros[4] = { 0.0, 0.0, 0.0, 0.0 }; - GLfloat ones[4] = { 1.0, 1.0, 1.0, 1.0 }; - GLfloat dst[2][4], *dstp; - - for (pass = 0; pass < shader->NumPasses; pass++) { - if (pass > 0) - finish_pass(machine); - for (j = 0; j < MAX_NUM_FRAGMENT_REGISTERS_ATI; j++) { - texinst = &shader->SetupInst[pass][j]; - if (texinst->Opcode == ATI_FRAGMENT_SHADER_PASS_OP) - handle_pass_op(machine, texinst, span, column, j); - else if (texinst->Opcode == ATI_FRAGMENT_SHADER_SAMPLE_OP) - handle_sample_op(ctx, machine, texinst, span, column, j); - } - - for (pc = 0; pc < shader->numArithInstr[pass]; pc++) { - inst = &shader->Instructions[pass][pc]; - - /* setup the source registers for color and alpha ops */ - for (optype = 0; optype < 2; optype++) { - for (i = 0; i < inst->ArgCount[optype]; i++) { - GLint index = inst->SrcReg[optype][i].Index; - - if (index >= GL_REG_0_ATI && index <= GL_REG_5_ATI) - SETUP_SRC_REG(optype, i, - machine->Registers[index - GL_REG_0_ATI]); - else if (index >= GL_CON_0_ATI && index <= GL_CON_7_ATI) { - if (shader->LocalConstDef & (1 << (index - GL_CON_0_ATI))) { - SETUP_SRC_REG(optype, i, - shader->Constants[index - GL_CON_0_ATI]); - } else { - SETUP_SRC_REG(optype, i, - ctx->ATIFragmentShader.GlobalConstants[index - GL_CON_0_ATI]); - } - } - else if (index == GL_ONE) - SETUP_SRC_REG(optype, i, ones); - else if (index == GL_ZERO) - SETUP_SRC_REG(optype, i, zeros); - else if (index == GL_PRIMARY_COLOR_EXT) - SETUP_SRC_REG(optype, i, - machine->Inputs[ATI_FS_INPUT_PRIMARY]); - else if (index == GL_SECONDARY_INTERPOLATOR_ATI) - SETUP_SRC_REG(optype, i, - machine->Inputs[ATI_FS_INPUT_SECONDARY]); - - apply_src_rep(optype, inst->SrcReg[optype][i].argRep, - src[optype][i]); - apply_src_mod(optype, inst->SrcReg[optype][i].argMod, - src[optype][i]); - } - } - - /* Execute the operations - color then alpha */ - for (optype = 0; optype < 2; optype++) { - if (inst->Opcode[optype]) { - switch (inst->Opcode[optype]) { - case GL_ADD_ATI: - if (!optype) - for (i = 0; i < 3; i++) { - dst[optype][i] = - src[optype][0][i] + src[optype][1][i]; - } - else - dst[optype][3] = src[optype][0][3] + src[optype][1][3]; - break; - case GL_SUB_ATI: - if (!optype) - for (i = 0; i < 3; i++) { - dst[optype][i] = - src[optype][0][i] - src[optype][1][i]; - } - else - dst[optype][3] = src[optype][0][3] - src[optype][1][3]; - break; - case GL_MUL_ATI: - if (!optype) - for (i = 0; i < 3; i++) { - dst[optype][i] = - src[optype][0][i] * src[optype][1][i]; - } - else - dst[optype][3] = src[optype][0][3] * src[optype][1][3]; - break; - case GL_MAD_ATI: - if (!optype) - for (i = 0; i < 3; i++) { - dst[optype][i] = - src[optype][0][i] * src[optype][1][i] + - src[optype][2][i]; - } - else - dst[optype][3] = - src[optype][0][3] * src[optype][1][3] + - src[optype][2][3]; - break; - case GL_LERP_ATI: - if (!optype) - for (i = 0; i < 3; i++) { - dst[optype][i] = - src[optype][0][i] * src[optype][1][i] + (1 - - src - [optype] - [0][i]) * - src[optype][2][i]; - } - else - dst[optype][3] = - src[optype][0][3] * src[optype][1][3] + (1 - - src[optype] - [0][3]) * - src[optype][2][3]; - break; - - case GL_MOV_ATI: - if (!optype) - for (i = 0; i < 3; i++) { - dst[optype][i] = src[optype][0][i]; - } - else - dst[optype][3] = src[optype][0][3]; - break; - case GL_CND_ATI: - if (!optype) { - for (i = 0; i < 3; i++) { - dst[optype][i] = - (src[optype][2][i] > - 0.5) ? src[optype][0][i] : src[optype][1][i]; - } - } - else { - dst[optype][3] = - (src[optype][2][3] > - 0.5) ? src[optype][0][3] : src[optype][1][3]; - } - break; - - case GL_CND0_ATI: - if (!optype) - for (i = 0; i < 3; i++) { - dst[optype][i] = - (src[optype][2][i] >= - 0) ? src[optype][0][i] : src[optype][1][i]; - } - else { - dst[optype][3] = - (src[optype][2][3] >= - 0) ? src[optype][0][3] : src[optype][1][3]; - } - break; - case GL_DOT2_ADD_ATI: - { - GLfloat result; - - /* DOT 2 always uses the source from the color op */ - /* could save recalculation of dot products for alpha inst */ - result = src[0][0][0] * src[0][1][0] + - src[0][0][1] * src[0][1][1] + src[0][2][2]; - if (!optype) { - for (i = 0; i < 3; i++) { - dst[optype][i] = result; - } - } - else - dst[optype][3] = result; - } - break; - case GL_DOT3_ATI: - { - GLfloat result; - - /* DOT 3 always uses the source from the color op */ - result = src[0][0][0] * src[0][1][0] + - src[0][0][1] * src[0][1][1] + - src[0][0][2] * src[0][1][2]; - - if (!optype) { - for (i = 0; i < 3; i++) { - dst[optype][i] = result; - } - } - else - dst[optype][3] = result; - } - break; - case GL_DOT4_ATI: - { - GLfloat result; - - /* DOT 4 always uses the source from the color op */ - result = src[0][0][0] * src[0][1][0] + - src[0][0][1] * src[0][1][1] + - src[0][0][2] * src[0][1][2] + - src[0][0][3] * src[0][1][3]; - if (!optype) { - for (i = 0; i < 3; i++) { - dst[optype][i] = result; - } - } - else - dst[optype][3] = result; - } - break; - - } - } - } - - /* write out the destination registers */ - for (optype = 0; optype < 2; optype++) { - if (inst->Opcode[optype]) { - dstreg = inst->DstReg[optype].Index; - dstp = machine->Registers[dstreg - GL_REG_0_ATI]; - - if ((optype == 0) || ((inst->Opcode[1] != GL_DOT2_ADD_ATI) && - (inst->Opcode[1] != GL_DOT3_ATI) && (inst->Opcode[1] != GL_DOT4_ATI))) - write_dst_addr(optype, inst->DstReg[optype].dstMod, - inst->DstReg[optype].dstMask, dst[optype], - dstp); - else - write_dst_addr(1, inst->DstReg[0].dstMod, 0, dst[1], dstp); - } - } - } - } -} - - -/** - * Init fragment shader virtual machine state. - */ -static void -init_machine(struct gl_context * ctx, struct atifs_machine *machine, - const struct ati_fragment_shader *shader, - const SWspan *span, GLuint col) -{ - GLfloat (*inputs)[4] = machine->Inputs; - GLint i, j; - - for (i = 0; i < 6; i++) { - for (j = 0; j < 4; j++) - machine->Registers[i][j] = 0.0; - } - - COPY_4V(inputs[ATI_FS_INPUT_PRIMARY], span->array->attribs[FRAG_ATTRIB_COL0][col]); - COPY_4V(inputs[ATI_FS_INPUT_SECONDARY], span->array->attribs[FRAG_ATTRIB_COL1][col]); -} - - - -/** - * Execute the current ATI shader program, operating on the given span. - */ -void -_swrast_exec_fragment_shader(struct gl_context * ctx, SWspan *span) -{ - const struct ati_fragment_shader *shader = ctx->ATIFragmentShader.Current; - struct atifs_machine machine; - GLuint i; - - /* incoming colors should be floats */ - ASSERT(span->array->ChanType == GL_FLOAT); - - for (i = 0; i < span->end; i++) { - if (span->array->mask[i]) { - init_machine(ctx, &machine, shader, span, i); - - execute_shader(ctx, shader, &machine, span, i); - - /* store result color */ - { - const GLfloat *colOut = machine.Registers[0]; - /*fprintf(stderr,"outputs %f %f %f %f\n", - colOut[0], colOut[1], colOut[2], colOut[3]); */ - COPY_4V(span->array->attribs[FRAG_ATTRIB_COL0][i], colOut); - } - } - } -} +/* + * Copyright (C) 2004 David Airlie All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * DAVID AIRLIE BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/macros.h" +#include "main/atifragshader.h" +#include "swrast/s_atifragshader.h" +#include "swrast/s_context.h" + + +/** + * State for executing ATI fragment shader. + */ +struct atifs_machine +{ + GLfloat Registers[6][4]; /** six temporary registers */ + GLfloat PrevPassRegisters[6][4]; + GLfloat Inputs[2][4]; /** Primary, secondary input colors */ +}; + + + +/** + * Fetch a texel. + */ +static void +fetch_texel(struct gl_context * ctx, const GLfloat texcoord[4], GLfloat lambda, + GLuint unit, GLfloat color[4]) +{ + SWcontext *swrast = SWRAST_CONTEXT(ctx); + + /* XXX use a float-valued TextureSample routine here!!! */ + swrast->TextureSample[unit](ctx, ctx->Texture.Unit[unit]._Current, + 1, (const GLfloat(*)[4]) texcoord, + &lambda, (GLfloat (*)[4]) color); +} + +static void +apply_swizzle(GLfloat values[4], GLuint swizzle) +{ + GLfloat s, t, r, q; + + s = values[0]; + t = values[1]; + r = values[2]; + q = values[3]; + + switch (swizzle) { + case GL_SWIZZLE_STR_ATI: + values[0] = s; + values[1] = t; + values[2] = r; + break; + case GL_SWIZZLE_STQ_ATI: + values[0] = s; + values[1] = t; + values[2] = q; + break; + case GL_SWIZZLE_STR_DR_ATI: + values[0] = s / r; + values[1] = t / r; + values[2] = 1 / r; + break; + case GL_SWIZZLE_STQ_DQ_ATI: +/* make sure q is not 0 to avoid problems later with infinite values (texture lookup)? */ + if (q == 0.0F) + q = 0.000000001F; + values[0] = s / q; + values[1] = t / q; + values[2] = 1.0F / q; + break; + } + values[3] = 0.0; +} + +static void +apply_src_rep(GLint optype, GLuint rep, GLfloat * val) +{ + GLint i; + GLint start, end; + if (!rep) + return; + + start = optype ? 3 : 0; + end = 4; + + for (i = start; i < end; i++) { + switch (rep) { + case GL_RED: + val[i] = val[0]; + break; + case GL_GREEN: + val[i] = val[1]; + break; + case GL_BLUE: + val[i] = val[2]; + break; + case GL_ALPHA: + val[i] = val[3]; + break; + } + } +} + +static void +apply_src_mod(GLint optype, GLuint mod, GLfloat * val) +{ + GLint i; + GLint start, end; + + if (!mod) + return; + + start = optype ? 3 : 0; + end = 4; + + for (i = start; i < end; i++) { + if (mod & GL_COMP_BIT_ATI) + val[i] = 1 - val[i]; + + if (mod & GL_BIAS_BIT_ATI) + val[i] = val[i] - 0.5F; + + if (mod & GL_2X_BIT_ATI) + val[i] = 2 * val[i]; + + if (mod & GL_NEGATE_BIT_ATI) + val[i] = -val[i]; + } +} + +static void +apply_dst_mod(GLuint optype, GLuint mod, GLfloat * val) +{ + GLint i; + GLint has_sat = mod & GL_SATURATE_BIT_ATI; + GLint start, end; + + mod &= ~GL_SATURATE_BIT_ATI; + + start = optype ? 3 : 0; + end = optype ? 4 : 3; + + for (i = start; i < end; i++) { + switch (mod) { + case GL_2X_BIT_ATI: + val[i] = 2 * val[i]; + break; + case GL_4X_BIT_ATI: + val[i] = 4 * val[i]; + break; + case GL_8X_BIT_ATI: + val[i] = 8 * val[i]; + break; + case GL_HALF_BIT_ATI: + val[i] = val[i] * 0.5F; + break; + case GL_QUARTER_BIT_ATI: + val[i] = val[i] * 0.25F; + break; + case GL_EIGHTH_BIT_ATI: + val[i] = val[i] * 0.125F; + break; + } + + if (has_sat) { + if (val[i] < 0.0F) + val[i] = 0.0F; + else if (val[i] > 1.0F) + val[i] = 1.0F; + } + else { + if (val[i] < -8.0F) + val[i] = -8.0F; + else if (val[i] > 8.0F) + val[i] = 8.0F; + } + } +} + + +static void +write_dst_addr(GLuint optype, GLuint mod, GLuint mask, GLfloat * src, + GLfloat * dst) +{ + GLint i; + apply_dst_mod(optype, mod, src); + + if (optype == ATI_FRAGMENT_SHADER_COLOR_OP) { + if (mask) { + if (mask & GL_RED_BIT_ATI) + dst[0] = src[0]; + + if (mask & GL_GREEN_BIT_ATI) + dst[1] = src[1]; + + if (mask & GL_BLUE_BIT_ATI) + dst[2] = src[2]; + } + else { + for (i = 0; i < 3; i++) + dst[i] = src[i]; + } + } + else + dst[3] = src[3]; +} + +static void +finish_pass(struct atifs_machine *machine) +{ + GLint i; + + for (i = 0; i < 6; i++) { + COPY_4V(machine->PrevPassRegisters[i], machine->Registers[i]); + } +} + +struct ati_fs_opcode_st ati_fs_opcodes[] = { + {GL_ADD_ATI, 2}, + {GL_SUB_ATI, 2}, + {GL_MUL_ATI, 2}, + {GL_MAD_ATI, 3}, + {GL_LERP_ATI, 3}, + {GL_MOV_ATI, 1}, + {GL_CND_ATI, 3}, + {GL_CND0_ATI, 3}, + {GL_DOT2_ADD_ATI, 3}, + {GL_DOT3_ATI, 2}, + {GL_DOT4_ATI, 2} +}; + + + +static void +handle_pass_op(struct atifs_machine *machine, struct atifs_setupinst *texinst, + const SWspan *span, GLuint column, GLuint idx) +{ + GLuint swizzle = texinst->swizzle; + GLuint pass_tex = texinst->src; + + if (pass_tex >= GL_TEXTURE0_ARB && pass_tex <= GL_TEXTURE7_ARB) { + pass_tex -= GL_TEXTURE0_ARB; + COPY_4V(machine->Registers[idx], + span->array->attribs[FRAG_ATTRIB_TEX0 + pass_tex][column]); + } + else if (pass_tex >= GL_REG_0_ATI && pass_tex <= GL_REG_5_ATI) { + pass_tex -= GL_REG_0_ATI; + COPY_4V(machine->Registers[idx], machine->PrevPassRegisters[pass_tex]); + } + apply_swizzle(machine->Registers[idx], swizzle); + +} + +static void +handle_sample_op(struct gl_context * ctx, struct atifs_machine *machine, + struct atifs_setupinst *texinst, const SWspan *span, + GLuint column, GLuint idx) +{ +/* sample from unit idx using texinst->src as coords */ + GLuint swizzle = texinst->swizzle; + GLuint coord_source = texinst->src; + GLfloat tex_coords[4] = { 0 }; + + if (coord_source >= GL_TEXTURE0_ARB && coord_source <= GL_TEXTURE7_ARB) { + coord_source -= GL_TEXTURE0_ARB; + COPY_4V(tex_coords, + span->array->attribs[FRAG_ATTRIB_TEX0 + coord_source][column]); + } + else if (coord_source >= GL_REG_0_ATI && coord_source <= GL_REG_5_ATI) { + coord_source -= GL_REG_0_ATI; + COPY_4V(tex_coords, machine->PrevPassRegisters[coord_source]); + } + apply_swizzle(tex_coords, swizzle); + fetch_texel(ctx, tex_coords, 0.0F, idx, machine->Registers[idx]); +} + +#define SETUP_SRC_REG(optype, i, x) \ +do { \ + COPY_4V(src[optype][i], x); \ +} while (0) + + + +/** + * Execute the given fragment shader. + * NOTE: we do everything in single-precision floating point + * \param ctx - rendering context + * \param shader - the shader to execute + * \param machine - virtual machine state + * \param span - the SWspan we're operating on + * \param column - which pixel [i] we're operating on in the span + */ +static void +execute_shader(struct gl_context *ctx, const struct ati_fragment_shader *shader, + struct atifs_machine *machine, const SWspan *span, + GLuint column) +{ + GLuint pc; + struct atifs_instruction *inst; + struct atifs_setupinst *texinst; + GLint optype; + GLuint i; + GLint j, pass; + GLint dstreg; + GLfloat src[2][3][4]; + GLfloat zeros[4] = { 0.0, 0.0, 0.0, 0.0 }; + GLfloat ones[4] = { 1.0, 1.0, 1.0, 1.0 }; + GLfloat dst[2][4], *dstp; + + for (pass = 0; pass < shader->NumPasses; pass++) { + if (pass > 0) + finish_pass(machine); + for (j = 0; j < MAX_NUM_FRAGMENT_REGISTERS_ATI; j++) { + texinst = &shader->SetupInst[pass][j]; + if (texinst->Opcode == ATI_FRAGMENT_SHADER_PASS_OP) + handle_pass_op(machine, texinst, span, column, j); + else if (texinst->Opcode == ATI_FRAGMENT_SHADER_SAMPLE_OP) + handle_sample_op(ctx, machine, texinst, span, column, j); + } + + for (pc = 0; pc < shader->numArithInstr[pass]; pc++) { + inst = &shader->Instructions[pass][pc]; + + /* setup the source registers for color and alpha ops */ + for (optype = 0; optype < 2; optype++) { + for (i = 0; i < inst->ArgCount[optype]; i++) { + GLint index = inst->SrcReg[optype][i].Index; + + if (index >= GL_REG_0_ATI && index <= GL_REG_5_ATI) + SETUP_SRC_REG(optype, i, + machine->Registers[index - GL_REG_0_ATI]); + else if (index >= GL_CON_0_ATI && index <= GL_CON_7_ATI) { + if (shader->LocalConstDef & (1 << (index - GL_CON_0_ATI))) { + SETUP_SRC_REG(optype, i, + shader->Constants[index - GL_CON_0_ATI]); + } else { + SETUP_SRC_REG(optype, i, + ctx->ATIFragmentShader.GlobalConstants[index - GL_CON_0_ATI]); + } + } + else if (index == GL_ONE) + SETUP_SRC_REG(optype, i, ones); + else if (index == GL_ZERO) + SETUP_SRC_REG(optype, i, zeros); + else if (index == GL_PRIMARY_COLOR_EXT) + SETUP_SRC_REG(optype, i, + machine->Inputs[ATI_FS_INPUT_PRIMARY]); + else if (index == GL_SECONDARY_INTERPOLATOR_ATI) + SETUP_SRC_REG(optype, i, + machine->Inputs[ATI_FS_INPUT_SECONDARY]); + + apply_src_rep(optype, inst->SrcReg[optype][i].argRep, + src[optype][i]); + apply_src_mod(optype, inst->SrcReg[optype][i].argMod, + src[optype][i]); + } + } + + /* Execute the operations - color then alpha */ + for (optype = 0; optype < 2; optype++) { + if (inst->Opcode[optype]) { + switch (inst->Opcode[optype]) { + case GL_ADD_ATI: + if (!optype) + for (i = 0; i < 3; i++) { + dst[optype][i] = + src[optype][0][i] + src[optype][1][i]; + } + else + dst[optype][3] = src[optype][0][3] + src[optype][1][3]; + break; + case GL_SUB_ATI: + if (!optype) + for (i = 0; i < 3; i++) { + dst[optype][i] = + src[optype][0][i] - src[optype][1][i]; + } + else + dst[optype][3] = src[optype][0][3] - src[optype][1][3]; + break; + case GL_MUL_ATI: + if (!optype) + for (i = 0; i < 3; i++) { + dst[optype][i] = + src[optype][0][i] * src[optype][1][i]; + } + else + dst[optype][3] = src[optype][0][3] * src[optype][1][3]; + break; + case GL_MAD_ATI: + if (!optype) + for (i = 0; i < 3; i++) { + dst[optype][i] = + src[optype][0][i] * src[optype][1][i] + + src[optype][2][i]; + } + else + dst[optype][3] = + src[optype][0][3] * src[optype][1][3] + + src[optype][2][3]; + break; + case GL_LERP_ATI: + if (!optype) + for (i = 0; i < 3; i++) { + dst[optype][i] = + src[optype][0][i] * src[optype][1][i] + (1 - + src + [optype] + [0][i]) * + src[optype][2][i]; + } + else + dst[optype][3] = + src[optype][0][3] * src[optype][1][3] + (1 - + src[optype] + [0][3]) * + src[optype][2][3]; + break; + + case GL_MOV_ATI: + if (!optype) + for (i = 0; i < 3; i++) { + dst[optype][i] = src[optype][0][i]; + } + else + dst[optype][3] = src[optype][0][3]; + break; + case GL_CND_ATI: + if (!optype) { + for (i = 0; i < 3; i++) { + dst[optype][i] = + (src[optype][2][i] > + 0.5) ? src[optype][0][i] : src[optype][1][i]; + } + } + else { + dst[optype][3] = + (src[optype][2][3] > + 0.5) ? src[optype][0][3] : src[optype][1][3]; + } + break; + + case GL_CND0_ATI: + if (!optype) + for (i = 0; i < 3; i++) { + dst[optype][i] = + (src[optype][2][i] >= + 0) ? src[optype][0][i] : src[optype][1][i]; + } + else { + dst[optype][3] = + (src[optype][2][3] >= + 0) ? src[optype][0][3] : src[optype][1][3]; + } + break; + case GL_DOT2_ADD_ATI: + { + GLfloat result; + + /* DOT 2 always uses the source from the color op */ + /* could save recalculation of dot products for alpha inst */ + result = src[0][0][0] * src[0][1][0] + + src[0][0][1] * src[0][1][1] + src[0][2][2]; + if (!optype) { + for (i = 0; i < 3; i++) { + dst[optype][i] = result; + } + } + else + dst[optype][3] = result; + } + break; + case GL_DOT3_ATI: + { + GLfloat result; + + /* DOT 3 always uses the source from the color op */ + result = src[0][0][0] * src[0][1][0] + + src[0][0][1] * src[0][1][1] + + src[0][0][2] * src[0][1][2]; + + if (!optype) { + for (i = 0; i < 3; i++) { + dst[optype][i] = result; + } + } + else + dst[optype][3] = result; + } + break; + case GL_DOT4_ATI: + { + GLfloat result; + + /* DOT 4 always uses the source from the color op */ + result = src[0][0][0] * src[0][1][0] + + src[0][0][1] * src[0][1][1] + + src[0][0][2] * src[0][1][2] + + src[0][0][3] * src[0][1][3]; + if (!optype) { + for (i = 0; i < 3; i++) { + dst[optype][i] = result; + } + } + else + dst[optype][3] = result; + } + break; + + } + } + } + + /* write out the destination registers */ + for (optype = 0; optype < 2; optype++) { + if (inst->Opcode[optype]) { + dstreg = inst->DstReg[optype].Index; + dstp = machine->Registers[dstreg - GL_REG_0_ATI]; + + if ((optype == 0) || ((inst->Opcode[1] != GL_DOT2_ADD_ATI) && + (inst->Opcode[1] != GL_DOT3_ATI) && (inst->Opcode[1] != GL_DOT4_ATI))) + write_dst_addr(optype, inst->DstReg[optype].dstMod, + inst->DstReg[optype].dstMask, dst[optype], + dstp); + else + write_dst_addr(1, inst->DstReg[0].dstMod, 0, dst[1], dstp); + } + } + } + } +} + + +/** + * Init fragment shader virtual machine state. + */ +static void +init_machine(struct gl_context * ctx, struct atifs_machine *machine, + const struct ati_fragment_shader *shader, + const SWspan *span, GLuint col) +{ + GLfloat (*inputs)[4] = machine->Inputs; + GLint i, j; + + for (i = 0; i < 6; i++) { + for (j = 0; j < 4; j++) + machine->Registers[i][j] = 0.0; + } + + COPY_4V(inputs[ATI_FS_INPUT_PRIMARY], span->array->attribs[FRAG_ATTRIB_COL0][col]); + COPY_4V(inputs[ATI_FS_INPUT_SECONDARY], span->array->attribs[FRAG_ATTRIB_COL1][col]); +} + + + +/** + * Execute the current ATI shader program, operating on the given span. + */ +void +_swrast_exec_fragment_shader(struct gl_context * ctx, SWspan *span) +{ + const struct ati_fragment_shader *shader = ctx->ATIFragmentShader.Current; + struct atifs_machine machine; + GLuint i; + + /* incoming colors should be floats */ + ASSERT(span->array->ChanType == GL_FLOAT); + + for (i = 0; i < span->end; i++) { + if (span->array->mask[i]) { + init_machine(ctx, &machine, shader, span, i); + + execute_shader(ctx, shader, &machine, span, i); + + /* store result color */ + { + const GLfloat *colOut = machine.Registers[0]; + /*fprintf(stderr,"outputs %f %f %f %f\n", + colOut[0], colOut[1], colOut[2], colOut[3]); */ + COPY_4V(span->array->attribs[FRAG_ATTRIB_COL0][i], colOut); + } + } + } +} diff --git a/mesalib/src/mesa/swrast/s_atifragshader.h b/mesalib/src/mesa/swrast/s_atifragshader.h index 16ed86fd7..10aaaa592 100644 --- a/mesalib/src/mesa/swrast/s_atifragshader.h +++ b/mesalib/src/mesa/swrast/s_atifragshader.h @@ -1,38 +1,38 @@ -/* - * Mesa 3-D graphics library - * Version: 6.1 - * - * Copyright (C) 1999-2003 David Airlie All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * DAVID AIRLIE BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_ATIFRAGSHADER_H -#define S_ATIFRAGSHADER_H - - -#include "s_span.h" - -struct gl_context; - -extern void -_swrast_exec_fragment_shader( struct gl_context *ctx, SWspan *span ); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.1 + * + * Copyright (C) 1999-2003 David Airlie All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * DAVID AIRLIE BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_ATIFRAGSHADER_H +#define S_ATIFRAGSHADER_H + + +#include "s_span.h" + +struct gl_context; + +extern void +_swrast_exec_fragment_shader( struct gl_context *ctx, SWspan *span ); + + +#endif diff --git a/mesalib/src/mesa/swrast/s_blend.h b/mesalib/src/mesa/swrast/s_blend.h index 8ecff52cf..69cd89e7a 100644 --- a/mesalib/src/mesa/swrast/s_blend.h +++ b/mesalib/src/mesa/swrast/s_blend.h @@ -1,45 +1,45 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.2 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_BLEND_H -#define S_BLEND_H - - -#include "main/glheader.h" -#include "s_span.h" - -struct gl_context; -struct gl_renderbuffer; - - -extern void -_swrast_blend_span(struct gl_context *ctx, struct gl_renderbuffer *rb, SWspan *span); - - -extern void -_swrast_choose_blend_func(struct gl_context *ctx, GLenum chanType); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.5.2 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_BLEND_H +#define S_BLEND_H + + +#include "main/glheader.h" +#include "s_span.h" + +struct gl_context; +struct gl_renderbuffer; + + +extern void +_swrast_blend_span(struct gl_context *ctx, struct gl_renderbuffer *rb, SWspan *span); + + +extern void +_swrast_choose_blend_func(struct gl_context *ctx, GLenum chanType); + + +#endif diff --git a/mesalib/src/mesa/swrast/s_feedback.c b/mesalib/src/mesa/swrast/s_feedback.c index e548bf9c3..00f92d463 100644 --- a/mesalib/src/mesa/swrast/s_feedback.c +++ b/mesalib/src/mesa/swrast/s_feedback.c @@ -1,138 +1,138 @@ -/* - * Mesa 3-D graphics library - * Version: 7.0 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/feedback.h" -#include "main/macros.h" - -#include "s_context.h" -#include "s_feedback.h" -#include "s_triangle.h" - - - -static void -feedback_vertex(struct gl_context * ctx, const SWvertex * v, const SWvertex * pv) -{ - GLfloat win[4]; - const GLfloat *vtc = v->attrib[FRAG_ATTRIB_TEX0]; - const GLfloat *color = v->attrib[FRAG_ATTRIB_COL0]; - - win[0] = v->attrib[FRAG_ATTRIB_WPOS][0]; - win[1] = v->attrib[FRAG_ATTRIB_WPOS][1]; - win[2] = v->attrib[FRAG_ATTRIB_WPOS][2] / ctx->DrawBuffer->_DepthMaxF; - win[3] = 1.0F / v->attrib[FRAG_ATTRIB_WPOS][3]; - - _mesa_feedback_vertex(ctx, win, color, vtc); -} - - -/* - * Put triangle in feedback buffer. - */ -void -_swrast_feedback_triangle(struct gl_context *ctx, const SWvertex *v0, - const SWvertex *v1, const SWvertex *v2) -{ - if (!_swrast_culltriangle(ctx, v0, v1, v2)) { - _mesa_feedback_token(ctx, (GLfloat) (GLint) GL_POLYGON_TOKEN); - _mesa_feedback_token(ctx, (GLfloat) 3); /* three vertices */ - - if (ctx->Light.ShadeModel == GL_SMOOTH) { - feedback_vertex(ctx, v0, v0); - feedback_vertex(ctx, v1, v1); - feedback_vertex(ctx, v2, v2); - } - else { - feedback_vertex(ctx, v0, v2); - feedback_vertex(ctx, v1, v2); - feedback_vertex(ctx, v2, v2); - } - } -} - - -void -_swrast_feedback_line(struct gl_context *ctx, const SWvertex *v0, - const SWvertex *v1) -{ - GLenum token = GL_LINE_TOKEN; - SWcontext *swrast = SWRAST_CONTEXT(ctx); - - if (swrast->StippleCounter == 0) - token = GL_LINE_RESET_TOKEN; - - _mesa_feedback_token(ctx, (GLfloat) (GLint) token); - - if (ctx->Light.ShadeModel == GL_SMOOTH) { - feedback_vertex(ctx, v0, v0); - feedback_vertex(ctx, v1, v1); - } - else { - feedback_vertex(ctx, v0, v1); - feedback_vertex(ctx, v1, v1); - } - - swrast->StippleCounter++; -} - - -void -_swrast_feedback_point(struct gl_context *ctx, const SWvertex *v) -{ - _mesa_feedback_token(ctx, (GLfloat) (GLint) GL_POINT_TOKEN); - feedback_vertex(ctx, v, v); -} - - -void -_swrast_select_triangle(struct gl_context *ctx, const SWvertex *v0, - const SWvertex *v1, const SWvertex *v2) -{ - if (!_swrast_culltriangle(ctx, v0, v1, v2)) { - const GLfloat zs = 1.0F / ctx->DrawBuffer->_DepthMaxF; - - _mesa_update_hitflag( ctx, v0->attrib[FRAG_ATTRIB_WPOS][2] * zs ); - _mesa_update_hitflag( ctx, v1->attrib[FRAG_ATTRIB_WPOS][2] * zs ); - _mesa_update_hitflag( ctx, v2->attrib[FRAG_ATTRIB_WPOS][2] * zs ); - } -} - - -void -_swrast_select_line(struct gl_context *ctx, const SWvertex *v0, const SWvertex *v1) -{ - const GLfloat zs = 1.0F / ctx->DrawBuffer->_DepthMaxF; - _mesa_update_hitflag( ctx, v0->attrib[FRAG_ATTRIB_WPOS][2] * zs ); - _mesa_update_hitflag( ctx, v1->attrib[FRAG_ATTRIB_WPOS][2] * zs ); -} - - -void -_swrast_select_point(struct gl_context *ctx, const SWvertex *v) -{ - const GLfloat zs = 1.0F / ctx->DrawBuffer->_DepthMaxF; - _mesa_update_hitflag( ctx, v->attrib[FRAG_ATTRIB_WPOS][2] * zs ); -} +/* + * Mesa 3-D graphics library + * Version: 7.0 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/feedback.h" +#include "main/macros.h" + +#include "s_context.h" +#include "s_feedback.h" +#include "s_triangle.h" + + + +static void +feedback_vertex(struct gl_context * ctx, const SWvertex * v, const SWvertex * pv) +{ + GLfloat win[4]; + const GLfloat *vtc = v->attrib[FRAG_ATTRIB_TEX0]; + const GLfloat *color = v->attrib[FRAG_ATTRIB_COL0]; + + win[0] = v->attrib[FRAG_ATTRIB_WPOS][0]; + win[1] = v->attrib[FRAG_ATTRIB_WPOS][1]; + win[2] = v->attrib[FRAG_ATTRIB_WPOS][2] / ctx->DrawBuffer->_DepthMaxF; + win[3] = 1.0F / v->attrib[FRAG_ATTRIB_WPOS][3]; + + _mesa_feedback_vertex(ctx, win, color, vtc); +} + + +/* + * Put triangle in feedback buffer. + */ +void +_swrast_feedback_triangle(struct gl_context *ctx, const SWvertex *v0, + const SWvertex *v1, const SWvertex *v2) +{ + if (!_swrast_culltriangle(ctx, v0, v1, v2)) { + _mesa_feedback_token(ctx, (GLfloat) (GLint) GL_POLYGON_TOKEN); + _mesa_feedback_token(ctx, (GLfloat) 3); /* three vertices */ + + if (ctx->Light.ShadeModel == GL_SMOOTH) { + feedback_vertex(ctx, v0, v0); + feedback_vertex(ctx, v1, v1); + feedback_vertex(ctx, v2, v2); + } + else { + feedback_vertex(ctx, v0, v2); + feedback_vertex(ctx, v1, v2); + feedback_vertex(ctx, v2, v2); + } + } +} + + +void +_swrast_feedback_line(struct gl_context *ctx, const SWvertex *v0, + const SWvertex *v1) +{ + GLenum token = GL_LINE_TOKEN; + SWcontext *swrast = SWRAST_CONTEXT(ctx); + + if (swrast->StippleCounter == 0) + token = GL_LINE_RESET_TOKEN; + + _mesa_feedback_token(ctx, (GLfloat) (GLint) token); + + if (ctx->Light.ShadeModel == GL_SMOOTH) { + feedback_vertex(ctx, v0, v0); + feedback_vertex(ctx, v1, v1); + } + else { + feedback_vertex(ctx, v0, v1); + feedback_vertex(ctx, v1, v1); + } + + swrast->StippleCounter++; +} + + +void +_swrast_feedback_point(struct gl_context *ctx, const SWvertex *v) +{ + _mesa_feedback_token(ctx, (GLfloat) (GLint) GL_POINT_TOKEN); + feedback_vertex(ctx, v, v); +} + + +void +_swrast_select_triangle(struct gl_context *ctx, const SWvertex *v0, + const SWvertex *v1, const SWvertex *v2) +{ + if (!_swrast_culltriangle(ctx, v0, v1, v2)) { + const GLfloat zs = 1.0F / ctx->DrawBuffer->_DepthMaxF; + + _mesa_update_hitflag( ctx, v0->attrib[FRAG_ATTRIB_WPOS][2] * zs ); + _mesa_update_hitflag( ctx, v1->attrib[FRAG_ATTRIB_WPOS][2] * zs ); + _mesa_update_hitflag( ctx, v2->attrib[FRAG_ATTRIB_WPOS][2] * zs ); + } +} + + +void +_swrast_select_line(struct gl_context *ctx, const SWvertex *v0, const SWvertex *v1) +{ + const GLfloat zs = 1.0F / ctx->DrawBuffer->_DepthMaxF; + _mesa_update_hitflag( ctx, v0->attrib[FRAG_ATTRIB_WPOS][2] * zs ); + _mesa_update_hitflag( ctx, v1->attrib[FRAG_ATTRIB_WPOS][2] * zs ); +} + + +void +_swrast_select_point(struct gl_context *ctx, const SWvertex *v) +{ + const GLfloat zs = 1.0F / ctx->DrawBuffer->_DepthMaxF; + _mesa_update_hitflag( ctx, v->attrib[FRAG_ATTRIB_WPOS][2] * zs ); +} diff --git a/mesalib/src/mesa/swrast/s_feedback.h b/mesalib/src/mesa/swrast/s_feedback.h index f0ca2c0e6..6bfd49735 100644 --- a/mesalib/src/mesa/swrast/s_feedback.h +++ b/mesalib/src/mesa/swrast/s_feedback.h @@ -1,50 +1,50 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_FEEDBACK_H -#define S_FEEDBACK_H - - -#include "swrast.h" - - -extern void _swrast_feedback_point( struct gl_context *ctx, const SWvertex *v ); - -extern void _swrast_feedback_line( struct gl_context *ctx, - const SWvertex *v1, const SWvertex *v2 ); - -extern void _swrast_feedback_triangle( struct gl_context *ctx, const SWvertex *v0, - const SWvertex *v1, const SWvertex *v2 ); - -extern void _swrast_select_point( struct gl_context *ctx, const SWvertex *v ); - -extern void _swrast_select_line( struct gl_context *ctx, - const SWvertex *v1, const SWvertex *v2 ); - -extern void _swrast_select_triangle( struct gl_context *ctx, const SWvertex *v0, - const SWvertex *v1, const SWvertex *v2 ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_FEEDBACK_H +#define S_FEEDBACK_H + + +#include "swrast.h" + + +extern void _swrast_feedback_point( struct gl_context *ctx, const SWvertex *v ); + +extern void _swrast_feedback_line( struct gl_context *ctx, + const SWvertex *v1, const SWvertex *v2 ); + +extern void _swrast_feedback_triangle( struct gl_context *ctx, const SWvertex *v0, + const SWvertex *v1, const SWvertex *v2 ); + +extern void _swrast_select_point( struct gl_context *ctx, const SWvertex *v ); + +extern void _swrast_select_line( struct gl_context *ctx, + const SWvertex *v1, const SWvertex *v2 ); + +extern void _swrast_select_triangle( struct gl_context *ctx, const SWvertex *v0, + const SWvertex *v1, const SWvertex *v2 ); + +#endif diff --git a/mesalib/src/mesa/swrast/s_fog.h b/mesalib/src/mesa/swrast/s_fog.h index eed9b8bdd..9f93b7050 100644 --- a/mesalib/src/mesa/swrast/s_fog.h +++ b/mesalib/src/mesa/swrast/s_fog.h @@ -1,42 +1,42 @@ - -/* - * Mesa 3-D graphics library - * Version: 4.1 - * - * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_FOG_H -#define S_FOG_H - - -#include "main/glheader.h" -#include "s_span.h" - -struct gl_context; - -extern GLfloat -_swrast_z_to_fogfactor(struct gl_context *ctx, GLfloat z); - -extern void -_swrast_fog_rgba_span( const struct gl_context *ctx, SWspan *span ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 4.1 + * + * Copyright (C) 1999-2002 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_FOG_H +#define S_FOG_H + + +#include "main/glheader.h" +#include "s_span.h" + +struct gl_context; + +extern GLfloat +_swrast_z_to_fogfactor(struct gl_context *ctx, GLfloat z); + +extern void +_swrast_fog_rgba_span( const struct gl_context *ctx, SWspan *span ); + +#endif diff --git a/mesalib/src/mesa/swrast/s_lines.h b/mesalib/src/mesa/swrast/s_lines.h index 0f95eecab..a4c98a855 100644 --- a/mesalib/src/mesa/swrast/s_lines.h +++ b/mesalib/src/mesa/swrast/s_lines.h @@ -1,41 +1,41 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_LINES_H -#define S_LINES_H - -#include "swrast.h" - -void -_swrast_choose_line( struct gl_context *ctx ); - -void -_swrast_add_spec_terms_line( struct gl_context *ctx, - const SWvertex *v0, - const SWvertex *v1 ); - - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_LINES_H +#define S_LINES_H + +#include "swrast.h" + +void +_swrast_choose_line( struct gl_context *ctx ); + +void +_swrast_add_spec_terms_line( struct gl_context *ctx, + const SWvertex *v0, + const SWvertex *v1 ); + + +#endif diff --git a/mesalib/src/mesa/swrast/s_logic.h b/mesalib/src/mesa/swrast/s_logic.h index 39713b990..0a3adfca5 100644 --- a/mesalib/src/mesa/swrast/s_logic.h +++ b/mesalib/src/mesa/swrast/s_logic.h @@ -1,40 +1,40 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.2 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_LOGIC_H -#define S_LOGIC_H - - -#include "s_span.h" - -struct gl_context; -struct gl_renderbuffer; - -extern void -_swrast_logicop_rgba_span(struct gl_context *ctx, struct gl_renderbuffer *rb, - SWspan *span); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.5.2 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_LOGIC_H +#define S_LOGIC_H + + +#include "s_span.h" + +struct gl_context; +struct gl_renderbuffer; + +extern void +_swrast_logicop_rgba_span(struct gl_context *ctx, struct gl_renderbuffer *rb, + SWspan *span); + + +#endif diff --git a/mesalib/src/mesa/swrast/s_masking.h b/mesalib/src/mesa/swrast/s_masking.h index 7ab02cc7a..5124509a0 100644 --- a/mesalib/src/mesa/swrast/s_masking.h +++ b/mesalib/src/mesa/swrast/s_masking.h @@ -1,41 +1,41 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.2 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_MASKING_H -#define S_MASKING_H - - -#include "main/glheader.h" -#include "s_span.h" - -struct gl_context; -struct gl_renderbuffer; - - -extern void -_swrast_mask_rgba_span(struct gl_context *ctx, struct gl_renderbuffer *rb, - SWspan *span, GLuint buf); - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.5.2 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_MASKING_H +#define S_MASKING_H + + +#include "main/glheader.h" +#include "s_span.h" + +struct gl_context; +struct gl_renderbuffer; + + +extern void +_swrast_mask_rgba_span(struct gl_context *ctx, struct gl_renderbuffer *rb, + SWspan *span, GLuint buf); + +#endif diff --git a/mesalib/src/mesa/swrast/s_points.h b/mesalib/src/mesa/swrast/s_points.h index d99470bfe..0b6550e80 100644 --- a/mesalib/src/mesa/swrast/s_points.h +++ b/mesalib/src/mesa/swrast/s_points.h @@ -1,39 +1,39 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_POINTS_H -#define S_POINTS_H - -#include "swrast.h" - -extern void -_swrast_choose_point( struct gl_context *ctx ); - -extern void -_swrast_add_spec_terms_point( struct gl_context *ctx, - const SWvertex *v0 ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_POINTS_H +#define S_POINTS_H + +#include "swrast.h" + +extern void +_swrast_choose_point( struct gl_context *ctx ); + +extern void +_swrast_add_spec_terms_point( struct gl_context *ctx, + const SWvertex *v0 ); + +#endif diff --git a/mesalib/src/mesa/swrast/s_texcombine.h b/mesalib/src/mesa/swrast/s_texcombine.h index 3897acd69..11049d86b 100644 --- a/mesalib/src/mesa/swrast/s_texcombine.h +++ b/mesalib/src/mesa/swrast/s_texcombine.h @@ -1,37 +1,37 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_TEXCOMBINE_H -#define S_TEXCOMBINE_H - - -#include "s_span.h" - -struct gl_context; - -extern void -_swrast_texture_span( struct gl_context *ctx, SWspan *span ); - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_TEXCOMBINE_H +#define S_TEXCOMBINE_H + + +#include "s_span.h" + +struct gl_context; + +extern void +_swrast_texture_span( struct gl_context *ctx, SWspan *span ); + +#endif diff --git a/mesalib/src/mesa/swrast/s_triangle.h b/mesalib/src/mesa/swrast/s_triangle.h index bde5dd58b..46e23d420 100644 --- a/mesalib/src/mesa/swrast/s_triangle.h +++ b/mesalib/src/mesa/swrast/s_triangle.h @@ -1,50 +1,50 @@ - -/* - * Mesa 3-D graphics library - * Version: 5.1 - * - * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef S_TRIANGLES_H -#define S_TRIANGLES_H - - -#include "swrast.h" - - -extern GLboolean -_swrast_culltriangle( struct gl_context *ctx, - const SWvertex *v0, - const SWvertex *v1, - const SWvertex *v2); - -extern void -_swrast_choose_triangle( struct gl_context *ctx ); - -extern void -_swrast_add_spec_terms_triangle( struct gl_context *ctx, - const SWvertex *v0, - const SWvertex *v1, - const SWvertex *v2 ); - - -#endif + +/* + * Mesa 3-D graphics library + * Version: 5.1 + * + * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef S_TRIANGLES_H +#define S_TRIANGLES_H + + +#include "swrast.h" + + +extern GLboolean +_swrast_culltriangle( struct gl_context *ctx, + const SWvertex *v0, + const SWvertex *v1, + const SWvertex *v2); + +extern void +_swrast_choose_triangle( struct gl_context *ctx ); + +extern void +_swrast_add_spec_terms_triangle( struct gl_context *ctx, + const SWvertex *v0, + const SWvertex *v1, + const SWvertex *v2 ); + + +#endif diff --git a/mesalib/src/mesa/swrast_setup/NOTES b/mesalib/src/mesa/swrast_setup/NOTES index b5d2c503e..bdf57c39e 100644 --- a/mesalib/src/mesa/swrast_setup/NOTES +++ b/mesalib/src/mesa/swrast_setup/NOTES @@ -1,65 +1,65 @@ -INTRODUCTION - -A helper module which provides glue to bind the software rasterizer to -the software t&l module. The main task of this module is to build -swrast vertices from the t&l vertex_buffer structs, and to use them to -perform triangle setup functions not implemented in the software -rasterizer. - -The module implements a full set of functions to plug into the -t_vb_render.c driver interface (tnl->Driver.Render.*). - -There are strong advantages to decoupling the software rasterizer from -the t&l module, primarily allowing hardware drivers better control -over fallbacks, the removal of implicit knowledge about the software -rasterizer in the t&l module, allowing the two modules to evolve -independently and allowing either to be substituted with equivalent -functionality from another codebase. - -This module implements triangle/quad setup for offset, unfilled and -twoside-lit triangles. The software rasterizer doesn't handle these -primitives directly. - -Hardware rasterization drivers typically use this module in situations -where no hardware rasterization is possible, ie during total -fallbacks. - -STATE - -To create and destroy the module: - - GLboolean _swsetup_CreateContext( struct gl_context *ctx ); - void _swsetup_DestroyContext( struct gl_context *ctx ); - -The module is not active by default, and must be installed by calling -_swrast_Wakeup(). This function installs internal swrast_setup -functions into all the tnl->Driver.Render driver hooks, thus taking -over the task of rasterization entirely: - - void _swrast_Wakeup( struct gl_context *ctx ); - - -This module tracks state changes internally and maintains derived -values based on the current state. For this to work, the driver -ensure the following funciton is called whenever the state changes and -the swsetup module is 'awake': - - void _swsetup_InvalidateState( struct gl_context *ctx, GLuint new_state ); - -There is no explicit call to put the swsetup module to sleep. Simply -install other function pointers into all the tnl->Driver.Render.* -hooks, and (optionally) cease calling _swsetup_InvalidateState(). - -DRIVER INTERFACE - -The module offers a minimal driver interface: - - void (*Start)( struct gl_context *ctx ); - void (*Finish)( struct gl_context *ctx ); - -These are called before and after the completion of all swrast drawing -activity. As swrast doesn't call callbacks during triangle, line or -point rasterization, these are necessary to provide locking hooks for -some drivers. They may otherwise be left null. - - +INTRODUCTION + +A helper module which provides glue to bind the software rasterizer to +the software t&l module. The main task of this module is to build +swrast vertices from the t&l vertex_buffer structs, and to use them to +perform triangle setup functions not implemented in the software +rasterizer. + +The module implements a full set of functions to plug into the +t_vb_render.c driver interface (tnl->Driver.Render.*). + +There are strong advantages to decoupling the software rasterizer from +the t&l module, primarily allowing hardware drivers better control +over fallbacks, the removal of implicit knowledge about the software +rasterizer in the t&l module, allowing the two modules to evolve +independently and allowing either to be substituted with equivalent +functionality from another codebase. + +This module implements triangle/quad setup for offset, unfilled and +twoside-lit triangles. The software rasterizer doesn't handle these +primitives directly. + +Hardware rasterization drivers typically use this module in situations +where no hardware rasterization is possible, ie during total +fallbacks. + +STATE + +To create and destroy the module: + + GLboolean _swsetup_CreateContext( struct gl_context *ctx ); + void _swsetup_DestroyContext( struct gl_context *ctx ); + +The module is not active by default, and must be installed by calling +_swrast_Wakeup(). This function installs internal swrast_setup +functions into all the tnl->Driver.Render driver hooks, thus taking +over the task of rasterization entirely: + + void _swrast_Wakeup( struct gl_context *ctx ); + + +This module tracks state changes internally and maintains derived +values based on the current state. For this to work, the driver +ensure the following funciton is called whenever the state changes and +the swsetup module is 'awake': + + void _swsetup_InvalidateState( struct gl_context *ctx, GLuint new_state ); + +There is no explicit call to put the swsetup module to sleep. Simply +install other function pointers into all the tnl->Driver.Render.* +hooks, and (optionally) cease calling _swsetup_InvalidateState(). + +DRIVER INTERFACE + +The module offers a minimal driver interface: + + void (*Start)( struct gl_context *ctx ); + void (*Finish)( struct gl_context *ctx ); + +These are called before and after the completion of all swrast drawing +activity. As swrast doesn't call callbacks during triangle, line or +point rasterization, these are necessary to provide locking hooks for +some drivers. They may otherwise be left null. + + diff --git a/mesalib/src/mesa/swrast_setup/ss_triangle.c b/mesalib/src/mesa/swrast_setup/ss_triangle.c index 4fc9a0756..5d1c70e94 100644 --- a/mesalib/src/mesa/swrast_setup/ss_triangle.c +++ b/mesalib/src/mesa/swrast_setup/ss_triangle.c @@ -1,265 +1,265 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/macros.h" -#include "main/mtypes.h" - -#include "tnl/t_context.h" - -#include "ss_triangle.h" -#include "ss_context.h" - -#define SS_OFFSET_BIT 0x1 -#define SS_TWOSIDE_BIT 0x2 -#define SS_UNFILLED_BIT 0x4 -#define SS_MAX_TRIFUNC 0x8 - -static tnl_triangle_func tri_tab[SS_MAX_TRIFUNC]; -static tnl_quad_func quad_tab[SS_MAX_TRIFUNC]; - - -/* - * Render a triangle respecting edge flags. - */ -typedef void (* swsetup_edge_render_prim_tri)(struct gl_context *ctx, - const GLubyte *ef, - GLuint e0, - GLuint e1, - GLuint e2, - const SWvertex *v0, - const SWvertex *v1, - const SWvertex *v2); - -/* - * Render a triangle using lines and respecting edge flags. - */ -static void -_swsetup_edge_render_line_tri(struct gl_context *ctx, - const GLubyte *ef, - GLuint e0, - GLuint e1, - GLuint e2, - const SWvertex *v0, - const SWvertex *v1, - const SWvertex *v2) -{ - SScontext *swsetup = SWSETUP_CONTEXT(ctx); - - if (swsetup->render_prim == GL_POLYGON) { - if (ef[e2]) _swrast_Line( ctx, v2, v0 ); - if (ef[e0]) _swrast_Line( ctx, v0, v1 ); - if (ef[e1]) _swrast_Line( ctx, v1, v2 ); - } else { - if (ef[e0]) _swrast_Line( ctx, v0, v1 ); - if (ef[e1]) _swrast_Line( ctx, v1, v2 ); - if (ef[e2]) _swrast_Line( ctx, v2, v0 ); - } -} - -/* - * Render a triangle using points and respecting edge flags. - */ -static void -_swsetup_edge_render_point_tri(struct gl_context *ctx, - const GLubyte *ef, - GLuint e0, - GLuint e1, - GLuint e2, - const SWvertex *v0, - const SWvertex *v1, - const SWvertex *v2) -{ - if (ef[e0]) _swrast_Point( ctx, v0 ); - if (ef[e1]) _swrast_Point( ctx, v1 ); - if (ef[e2]) _swrast_Point( ctx, v2 ); - - _swrast_flush(ctx); -} - -/* - * Render a triangle respecting cull and shade model. - */ -static void _swsetup_render_tri(struct gl_context *ctx, - GLuint e0, - GLuint e1, - GLuint e2, - GLuint facing, - swsetup_edge_render_prim_tri render) -{ - SScontext *swsetup = SWSETUP_CONTEXT(ctx); - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - GLubyte *ef = VB->EdgeFlag; - SWvertex *verts = swsetup->verts; - SWvertex *v0 = &verts[e0]; - SWvertex *v1 = &verts[e1]; - SWvertex *v2 = &verts[e2]; - - /* cull testing */ - if (ctx->Polygon.CullFlag) { - if (facing == 1 && ctx->Polygon.CullFaceMode != GL_FRONT) - return; - if (facing == 0 && ctx->Polygon.CullFaceMode != GL_BACK) - return; - } - - _swrast_SetFacing(ctx, facing); - - if (ctx->Light.ShadeModel == GL_FLAT) { - GLchan c[2][4]; - GLfloat s[2][4]; - - /* save colors/indexes for v0, v1 vertices */ - COPY_CHAN4(c[0], v0->color); - COPY_CHAN4(c[1], v1->color); - COPY_4V(s[0], v0->attrib[FRAG_ATTRIB_COL1]); - COPY_4V(s[1], v1->attrib[FRAG_ATTRIB_COL1]); - - /* copy v2 color/indexes to v0, v1 indexes */ - COPY_CHAN4(v0->color, v2->color); - COPY_CHAN4(v1->color, v2->color); - COPY_4V(v0->attrib[FRAG_ATTRIB_COL1], v2->attrib[FRAG_ATTRIB_COL1]); - COPY_4V(v1->attrib[FRAG_ATTRIB_COL1], v2->attrib[FRAG_ATTRIB_COL1]); - - render(ctx, ef, e0, e1, e2, v0, v1, v2); - - COPY_CHAN4(v0->color, c[0]); - COPY_CHAN4(v1->color, c[1]); - COPY_4V(v0->attrib[FRAG_ATTRIB_COL1], s[0]); - COPY_4V(v1->attrib[FRAG_ATTRIB_COL1], s[1]); - } - else { - render(ctx, ef, e0, e1, e2, v0, v1, v2); - } -} - -#define SS_COLOR(a,b) UNCLAMPED_FLOAT_TO_RGBA_CHAN(a,b) -#define SS_SPEC(a,b) COPY_4V(a,b) -#define SS_IND(a,b) (a = b) - -#define IND (0) -#define TAG(x) x##_rgba -#include "ss_tritmp.h" - -#define IND (SS_OFFSET_BIT) -#define TAG(x) x##_offset_rgba -#include "ss_tritmp.h" - -#define IND (SS_TWOSIDE_BIT) -#define TAG(x) x##_twoside_rgba -#include "ss_tritmp.h" - -#define IND (SS_OFFSET_BIT|SS_TWOSIDE_BIT) -#define TAG(x) x##_offset_twoside_rgba -#include "ss_tritmp.h" - -#define IND (SS_UNFILLED_BIT) -#define TAG(x) x##_unfilled_rgba -#include "ss_tritmp.h" - -#define IND (SS_OFFSET_BIT|SS_UNFILLED_BIT) -#define TAG(x) x##_offset_unfilled_rgba -#include "ss_tritmp.h" - -#define IND (SS_TWOSIDE_BIT|SS_UNFILLED_BIT) -#define TAG(x) x##_twoside_unfilled_rgba -#include "ss_tritmp.h" - -#define IND (SS_OFFSET_BIT|SS_TWOSIDE_BIT|SS_UNFILLED_BIT) -#define TAG(x) x##_offset_twoside_unfilled_rgba -#include "ss_tritmp.h" - - -void _swsetup_trifuncs_init( struct gl_context *ctx ) -{ - (void) ctx; - - init_rgba(); - init_offset_rgba(); - init_twoside_rgba(); - init_offset_twoside_rgba(); - init_unfilled_rgba(); - init_offset_unfilled_rgba(); - init_twoside_unfilled_rgba(); - init_offset_twoside_unfilled_rgba(); -} - - -static void swsetup_points( struct gl_context *ctx, GLuint first, GLuint last ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - SWvertex *verts = SWSETUP_CONTEXT(ctx)->verts; - GLuint i; - - if (VB->Elts) { - for (i = first; i < last; i++) - if (VB->ClipMask[VB->Elts[i]] == 0) - _swrast_Point( ctx, &verts[VB->Elts[i]] ); - } - else { - for (i = first; i < last; i++) - if (VB->ClipMask[i] == 0) - _swrast_Point( ctx, &verts[i] ); - } -} - -static void swsetup_line( struct gl_context *ctx, GLuint v0, GLuint v1 ) -{ - SWvertex *verts = SWSETUP_CONTEXT(ctx)->verts; - _swrast_Line( ctx, &verts[v0], &verts[v1] ); -} - - - -void _swsetup_choose_trifuncs( struct gl_context *ctx ) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - GLuint ind = 0; - - if (ctx->Polygon.OffsetPoint || - ctx->Polygon.OffsetLine || - ctx->Polygon.OffsetFill) - ind |= SS_OFFSET_BIT; - - if ((ctx->Light.Enabled && ctx->Light.Model.TwoSide) || - (ctx->VertexProgram._Current && ctx->VertexProgram.TwoSideEnabled)) - ind |= SS_TWOSIDE_BIT; - - /* We piggyback the two-sided stencil front/back determination on the - * unfilled triangle path. - */ - if (ctx->Polygon.FrontMode != GL_FILL || - ctx->Polygon.BackMode != GL_FILL || - (ctx->Stencil.Enabled && ctx->Stencil._TestTwoSide)) - ind |= SS_UNFILLED_BIT; - - tnl->Driver.Render.Triangle = tri_tab[ind]; - tnl->Driver.Render.Quad = quad_tab[ind]; - tnl->Driver.Render.Line = swsetup_line; - tnl->Driver.Render.Points = swsetup_points; -} +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/macros.h" +#include "main/mtypes.h" + +#include "tnl/t_context.h" + +#include "ss_triangle.h" +#include "ss_context.h" + +#define SS_OFFSET_BIT 0x1 +#define SS_TWOSIDE_BIT 0x2 +#define SS_UNFILLED_BIT 0x4 +#define SS_MAX_TRIFUNC 0x8 + +static tnl_triangle_func tri_tab[SS_MAX_TRIFUNC]; +static tnl_quad_func quad_tab[SS_MAX_TRIFUNC]; + + +/* + * Render a triangle respecting edge flags. + */ +typedef void (* swsetup_edge_render_prim_tri)(struct gl_context *ctx, + const GLubyte *ef, + GLuint e0, + GLuint e1, + GLuint e2, + const SWvertex *v0, + const SWvertex *v1, + const SWvertex *v2); + +/* + * Render a triangle using lines and respecting edge flags. + */ +static void +_swsetup_edge_render_line_tri(struct gl_context *ctx, + const GLubyte *ef, + GLuint e0, + GLuint e1, + GLuint e2, + const SWvertex *v0, + const SWvertex *v1, + const SWvertex *v2) +{ + SScontext *swsetup = SWSETUP_CONTEXT(ctx); + + if (swsetup->render_prim == GL_POLYGON) { + if (ef[e2]) _swrast_Line( ctx, v2, v0 ); + if (ef[e0]) _swrast_Line( ctx, v0, v1 ); + if (ef[e1]) _swrast_Line( ctx, v1, v2 ); + } else { + if (ef[e0]) _swrast_Line( ctx, v0, v1 ); + if (ef[e1]) _swrast_Line( ctx, v1, v2 ); + if (ef[e2]) _swrast_Line( ctx, v2, v0 ); + } +} + +/* + * Render a triangle using points and respecting edge flags. + */ +static void +_swsetup_edge_render_point_tri(struct gl_context *ctx, + const GLubyte *ef, + GLuint e0, + GLuint e1, + GLuint e2, + const SWvertex *v0, + const SWvertex *v1, + const SWvertex *v2) +{ + if (ef[e0]) _swrast_Point( ctx, v0 ); + if (ef[e1]) _swrast_Point( ctx, v1 ); + if (ef[e2]) _swrast_Point( ctx, v2 ); + + _swrast_flush(ctx); +} + +/* + * Render a triangle respecting cull and shade model. + */ +static void _swsetup_render_tri(struct gl_context *ctx, + GLuint e0, + GLuint e1, + GLuint e2, + GLuint facing, + swsetup_edge_render_prim_tri render) +{ + SScontext *swsetup = SWSETUP_CONTEXT(ctx); + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + GLubyte *ef = VB->EdgeFlag; + SWvertex *verts = swsetup->verts; + SWvertex *v0 = &verts[e0]; + SWvertex *v1 = &verts[e1]; + SWvertex *v2 = &verts[e2]; + + /* cull testing */ + if (ctx->Polygon.CullFlag) { + if (facing == 1 && ctx->Polygon.CullFaceMode != GL_FRONT) + return; + if (facing == 0 && ctx->Polygon.CullFaceMode != GL_BACK) + return; + } + + _swrast_SetFacing(ctx, facing); + + if (ctx->Light.ShadeModel == GL_FLAT) { + GLchan c[2][4]; + GLfloat s[2][4]; + + /* save colors/indexes for v0, v1 vertices */ + COPY_CHAN4(c[0], v0->color); + COPY_CHAN4(c[1], v1->color); + COPY_4V(s[0], v0->attrib[FRAG_ATTRIB_COL1]); + COPY_4V(s[1], v1->attrib[FRAG_ATTRIB_COL1]); + + /* copy v2 color/indexes to v0, v1 indexes */ + COPY_CHAN4(v0->color, v2->color); + COPY_CHAN4(v1->color, v2->color); + COPY_4V(v0->attrib[FRAG_ATTRIB_COL1], v2->attrib[FRAG_ATTRIB_COL1]); + COPY_4V(v1->attrib[FRAG_ATTRIB_COL1], v2->attrib[FRAG_ATTRIB_COL1]); + + render(ctx, ef, e0, e1, e2, v0, v1, v2); + + COPY_CHAN4(v0->color, c[0]); + COPY_CHAN4(v1->color, c[1]); + COPY_4V(v0->attrib[FRAG_ATTRIB_COL1], s[0]); + COPY_4V(v1->attrib[FRAG_ATTRIB_COL1], s[1]); + } + else { + render(ctx, ef, e0, e1, e2, v0, v1, v2); + } +} + +#define SS_COLOR(a,b) UNCLAMPED_FLOAT_TO_RGBA_CHAN(a,b) +#define SS_SPEC(a,b) COPY_4V(a,b) +#define SS_IND(a,b) (a = b) + +#define IND (0) +#define TAG(x) x##_rgba +#include "ss_tritmp.h" + +#define IND (SS_OFFSET_BIT) +#define TAG(x) x##_offset_rgba +#include "ss_tritmp.h" + +#define IND (SS_TWOSIDE_BIT) +#define TAG(x) x##_twoside_rgba +#include "ss_tritmp.h" + +#define IND (SS_OFFSET_BIT|SS_TWOSIDE_BIT) +#define TAG(x) x##_offset_twoside_rgba +#include "ss_tritmp.h" + +#define IND (SS_UNFILLED_BIT) +#define TAG(x) x##_unfilled_rgba +#include "ss_tritmp.h" + +#define IND (SS_OFFSET_BIT|SS_UNFILLED_BIT) +#define TAG(x) x##_offset_unfilled_rgba +#include "ss_tritmp.h" + +#define IND (SS_TWOSIDE_BIT|SS_UNFILLED_BIT) +#define TAG(x) x##_twoside_unfilled_rgba +#include "ss_tritmp.h" + +#define IND (SS_OFFSET_BIT|SS_TWOSIDE_BIT|SS_UNFILLED_BIT) +#define TAG(x) x##_offset_twoside_unfilled_rgba +#include "ss_tritmp.h" + + +void _swsetup_trifuncs_init( struct gl_context *ctx ) +{ + (void) ctx; + + init_rgba(); + init_offset_rgba(); + init_twoside_rgba(); + init_offset_twoside_rgba(); + init_unfilled_rgba(); + init_offset_unfilled_rgba(); + init_twoside_unfilled_rgba(); + init_offset_twoside_unfilled_rgba(); +} + + +static void swsetup_points( struct gl_context *ctx, GLuint first, GLuint last ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + SWvertex *verts = SWSETUP_CONTEXT(ctx)->verts; + GLuint i; + + if (VB->Elts) { + for (i = first; i < last; i++) + if (VB->ClipMask[VB->Elts[i]] == 0) + _swrast_Point( ctx, &verts[VB->Elts[i]] ); + } + else { + for (i = first; i < last; i++) + if (VB->ClipMask[i] == 0) + _swrast_Point( ctx, &verts[i] ); + } +} + +static void swsetup_line( struct gl_context *ctx, GLuint v0, GLuint v1 ) +{ + SWvertex *verts = SWSETUP_CONTEXT(ctx)->verts; + _swrast_Line( ctx, &verts[v0], &verts[v1] ); +} + + + +void _swsetup_choose_trifuncs( struct gl_context *ctx ) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + GLuint ind = 0; + + if (ctx->Polygon.OffsetPoint || + ctx->Polygon.OffsetLine || + ctx->Polygon.OffsetFill) + ind |= SS_OFFSET_BIT; + + if ((ctx->Light.Enabled && ctx->Light.Model.TwoSide) || + (ctx->VertexProgram._Current && ctx->VertexProgram.TwoSideEnabled)) + ind |= SS_TWOSIDE_BIT; + + /* We piggyback the two-sided stencil front/back determination on the + * unfilled triangle path. + */ + if (ctx->Polygon.FrontMode != GL_FILL || + ctx->Polygon.BackMode != GL_FILL || + (ctx->Stencil.Enabled && ctx->Stencil._TestTwoSide)) + ind |= SS_UNFILLED_BIT; + + tnl->Driver.Render.Triangle = tri_tab[ind]; + tnl->Driver.Render.Quad = quad_tab[ind]; + tnl->Driver.Render.Line = swsetup_line; + tnl->Driver.Render.Points = swsetup_points; +} diff --git a/mesalib/src/mesa/swrast_setup/ss_triangle.h b/mesalib/src/mesa/swrast_setup/ss_triangle.h index e21a9749a..a027f4826 100644 --- a/mesalib/src/mesa/swrast_setup/ss_triangle.h +++ b/mesalib/src/mesa/swrast_setup/ss_triangle.h @@ -1,38 +1,38 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - -#ifndef SS_TRIANGLE_H -#define SS_TRIANGLE_H - -struct gl_context; - - -void _swsetup_trifuncs_init( struct gl_context *ctx ); -void _swsetup_choose_trifuncs( struct gl_context *ctx ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + +#ifndef SS_TRIANGLE_H +#define SS_TRIANGLE_H + +struct gl_context; + + +void _swsetup_trifuncs_init( struct gl_context *ctx ); +void _swsetup_choose_trifuncs( struct gl_context *ctx ); + +#endif diff --git a/mesalib/src/mesa/swrast_setup/ss_tritmp.h b/mesalib/src/mesa/swrast_setup/ss_tritmp.h index 9b2b7cf6a..5844ad594 100644 --- a/mesalib/src/mesa/swrast_setup/ss_tritmp.h +++ b/mesalib/src/mesa/swrast_setup/ss_tritmp.h @@ -1,248 +1,248 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - -/** - * This is where we handle assigning vertex colors based on front/back - * facing, compute polygon offset and handle glPolygonMode(). - */ -static void TAG(triangle)(struct gl_context *ctx, GLuint e0, GLuint e1, GLuint e2 ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - SScontext *swsetup = SWSETUP_CONTEXT(ctx); - SWvertex *verts = SWSETUP_CONTEXT(ctx)->verts; - SWvertex *v[3]; - GLfloat z[3]; - GLfloat offset, oz0, oz1, oz2; - GLenum mode = GL_FILL; - GLuint facing = 0; - GLchan saved_color[3][4] = { { 0 } }; - GLfloat saved_col0[3][4] = { { 0 } }; - GLfloat saved_spec[3][4] = { { 0 } }; - - v[0] = &verts[e0]; - v[1] = &verts[e1]; - v[2] = &verts[e2]; - - if (IND & (SS_TWOSIDE_BIT | SS_OFFSET_BIT | SS_UNFILLED_BIT)) - { - GLfloat ex = v[0]->attrib[FRAG_ATTRIB_WPOS][0] - v[2]->attrib[FRAG_ATTRIB_WPOS][0]; - GLfloat ey = v[0]->attrib[FRAG_ATTRIB_WPOS][1] - v[2]->attrib[FRAG_ATTRIB_WPOS][1]; - GLfloat fx = v[1]->attrib[FRAG_ATTRIB_WPOS][0] - v[2]->attrib[FRAG_ATTRIB_WPOS][0]; - GLfloat fy = v[1]->attrib[FRAG_ATTRIB_WPOS][1] - v[2]->attrib[FRAG_ATTRIB_WPOS][1]; - GLfloat cc = ex*fy - ey*fx; - - if (IND & (SS_TWOSIDE_BIT | SS_UNFILLED_BIT)) - { - facing = (cc < 0.0) ^ ctx->Polygon._FrontBit; - - if (IND & SS_UNFILLED_BIT) - mode = facing ? ctx->Polygon.BackMode : ctx->Polygon.FrontMode; - - if (facing == 1) { - if (IND & SS_TWOSIDE_BIT) { - if (VB->BackfaceColorPtr) { - GLfloat (*vbcolor)[4] = VB->BackfaceColorPtr->data; - - if (swsetup->intColors) { - COPY_CHAN4(saved_color[0], v[0]->color); - COPY_CHAN4(saved_color[1], v[1]->color); - COPY_CHAN4(saved_color[2], v[2]->color); - } - else { - COPY_4V(saved_col0[0], v[0]->attrib[FRAG_ATTRIB_COL0]); - COPY_4V(saved_col0[1], v[1]->attrib[FRAG_ATTRIB_COL0]); - COPY_4V(saved_col0[2], v[2]->attrib[FRAG_ATTRIB_COL0]); - } - - if (VB->BackfaceColorPtr->stride) { - if (swsetup->intColors) { - SS_COLOR(v[0]->color, vbcolor[e0]); - SS_COLOR(v[1]->color, vbcolor[e1]); - SS_COLOR(v[2]->color, vbcolor[e2]); - } - else { - COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL0], vbcolor[e0]); - COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL0], vbcolor[e1]); - COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL0], vbcolor[e2]); - } - } - else { - /* flat shade */ - if (swsetup->intColors) { - SS_COLOR(v[0]->color, vbcolor[0]); - SS_COLOR(v[1]->color, vbcolor[0]); - SS_COLOR(v[2]->color, vbcolor[0]); - } - else { - COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL0], vbcolor[0]); - COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL0], vbcolor[0]); - COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL0], vbcolor[0]); - } - } - } - - if (VB->BackfaceSecondaryColorPtr) { - GLfloat (*vbspec)[4] = VB->BackfaceSecondaryColorPtr->data; - - COPY_4V(saved_spec[0], v[0]->attrib[FRAG_ATTRIB_COL1]); - COPY_4V(saved_spec[1], v[1]->attrib[FRAG_ATTRIB_COL1]); - COPY_4V(saved_spec[2], v[2]->attrib[FRAG_ATTRIB_COL1]); - - if (VB->BackfaceSecondaryColorPtr->stride) { - SS_SPEC(v[0]->attrib[FRAG_ATTRIB_COL1], vbspec[e0]); - SS_SPEC(v[1]->attrib[FRAG_ATTRIB_COL1], vbspec[e1]); - SS_SPEC(v[2]->attrib[FRAG_ATTRIB_COL1], vbspec[e2]); - } - else { - SS_SPEC(v[0]->attrib[FRAG_ATTRIB_COL1], vbspec[0]); - SS_SPEC(v[1]->attrib[FRAG_ATTRIB_COL1], vbspec[0]); - SS_SPEC(v[2]->attrib[FRAG_ATTRIB_COL1], vbspec[0]); - } - } - } - } - } - - if (IND & SS_OFFSET_BIT) { - const GLfloat max = ctx->DrawBuffer->_DepthMaxF; - /* save original Z values (restored later) */ - z[0] = v[0]->attrib[FRAG_ATTRIB_WPOS][2]; - z[1] = v[1]->attrib[FRAG_ATTRIB_WPOS][2]; - z[2] = v[2]->attrib[FRAG_ATTRIB_WPOS][2]; - /* Note that Z values are already scaled to [0,65535] (for example) - * so no MRD value is used here. - */ - offset = ctx->Polygon.OffsetUnits; - if (cc * cc > 1e-16) { - const GLfloat ez = z[0] - z[2]; - const GLfloat fz = z[1] - z[2]; - const GLfloat oneOverArea = 1.0F / cc; - const GLfloat dzdx = FABSF((ey * fz - ez * fy) * oneOverArea); - const GLfloat dzdy = FABSF((ez * fx - ex * fz) * oneOverArea); - offset += MAX2(dzdx, dzdy) * ctx->Polygon.OffsetFactor; - } - /* new Z values */ - oz0 = CLAMP(v[0]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0F, max); - oz1 = CLAMP(v[1]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0F, max); - oz2 = CLAMP(v[2]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0F, max); - } - } - - if (mode == GL_POINT) { - if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetPoint) { - v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0; - v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1; - v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2; - } - _swsetup_render_tri(ctx, e0, e1, e2, facing, _swsetup_edge_render_point_tri); - } else if (mode == GL_LINE) { - if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetLine) { - v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0; - v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1; - v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2; - } - _swsetup_render_tri(ctx, e0, e1, e2, facing, _swsetup_edge_render_line_tri); - } else { - if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetFill) { - v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0; - v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1; - v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2; - } - _swrast_Triangle( ctx, v[0], v[1], v[2] ); - } - - /* - * Restore original vertex colors, etc. - */ - if (IND & SS_OFFSET_BIT) { - v[0]->attrib[FRAG_ATTRIB_WPOS][2] = z[0]; - v[1]->attrib[FRAG_ATTRIB_WPOS][2] = z[1]; - v[2]->attrib[FRAG_ATTRIB_WPOS][2] = z[2]; - } - - if (IND & SS_TWOSIDE_BIT) { - if (facing == 1) { - if (VB->BackfaceColorPtr) { - if (swsetup->intColors) { - COPY_CHAN4(v[0]->color, saved_color[0]); - COPY_CHAN4(v[1]->color, saved_color[1]); - COPY_CHAN4(v[2]->color, saved_color[2]); - } - else { - COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL0], saved_col0[0]); - COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL0], saved_col0[1]); - COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL0], saved_col0[2]); - } - } - - if (VB->BackfaceSecondaryColorPtr) { - COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL1], saved_spec[0]); - COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL1], saved_spec[1]); - COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL1], saved_spec[2]); - } - } - } -} - - - -/* Need to fixup edgeflags when decomposing to triangles: - */ -static void TAG(quadfunc)( struct gl_context *ctx, GLuint v0, - GLuint v1, GLuint v2, GLuint v3 ) -{ - if (IND & SS_UNFILLED_BIT) { - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - if (VB->EdgeFlag) { /* XXX this test shouldn't be needed (bug 12614) */ - GLubyte ef1 = VB->EdgeFlag[v1]; - GLubyte ef3 = VB->EdgeFlag[v3]; - VB->EdgeFlag[v1] = 0; - TAG(triangle)( ctx, v0, v1, v3 ); - VB->EdgeFlag[v1] = ef1; - VB->EdgeFlag[v3] = 0; - TAG(triangle)( ctx, v1, v2, v3 ); - VB->EdgeFlag[v3] = ef3; - } - } else { - TAG(triangle)( ctx, v0, v1, v3 ); - TAG(triangle)( ctx, v1, v2, v3 ); - } -} - - - - -static void TAG(init)( void ) -{ - tri_tab[IND] = TAG(triangle); - quad_tab[IND] = TAG(quadfunc); -} - - -#undef IND -#undef TAG +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + +/** + * This is where we handle assigning vertex colors based on front/back + * facing, compute polygon offset and handle glPolygonMode(). + */ +static void TAG(triangle)(struct gl_context *ctx, GLuint e0, GLuint e1, GLuint e2 ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + SScontext *swsetup = SWSETUP_CONTEXT(ctx); + SWvertex *verts = SWSETUP_CONTEXT(ctx)->verts; + SWvertex *v[3]; + GLfloat z[3]; + GLfloat offset, oz0, oz1, oz2; + GLenum mode = GL_FILL; + GLuint facing = 0; + GLchan saved_color[3][4] = { { 0 } }; + GLfloat saved_col0[3][4] = { { 0 } }; + GLfloat saved_spec[3][4] = { { 0 } }; + + v[0] = &verts[e0]; + v[1] = &verts[e1]; + v[2] = &verts[e2]; + + if (IND & (SS_TWOSIDE_BIT | SS_OFFSET_BIT | SS_UNFILLED_BIT)) + { + GLfloat ex = v[0]->attrib[FRAG_ATTRIB_WPOS][0] - v[2]->attrib[FRAG_ATTRIB_WPOS][0]; + GLfloat ey = v[0]->attrib[FRAG_ATTRIB_WPOS][1] - v[2]->attrib[FRAG_ATTRIB_WPOS][1]; + GLfloat fx = v[1]->attrib[FRAG_ATTRIB_WPOS][0] - v[2]->attrib[FRAG_ATTRIB_WPOS][0]; + GLfloat fy = v[1]->attrib[FRAG_ATTRIB_WPOS][1] - v[2]->attrib[FRAG_ATTRIB_WPOS][1]; + GLfloat cc = ex*fy - ey*fx; + + if (IND & (SS_TWOSIDE_BIT | SS_UNFILLED_BIT)) + { + facing = (cc < 0.0) ^ ctx->Polygon._FrontBit; + + if (IND & SS_UNFILLED_BIT) + mode = facing ? ctx->Polygon.BackMode : ctx->Polygon.FrontMode; + + if (facing == 1) { + if (IND & SS_TWOSIDE_BIT) { + if (VB->BackfaceColorPtr) { + GLfloat (*vbcolor)[4] = VB->BackfaceColorPtr->data; + + if (swsetup->intColors) { + COPY_CHAN4(saved_color[0], v[0]->color); + COPY_CHAN4(saved_color[1], v[1]->color); + COPY_CHAN4(saved_color[2], v[2]->color); + } + else { + COPY_4V(saved_col0[0], v[0]->attrib[FRAG_ATTRIB_COL0]); + COPY_4V(saved_col0[1], v[1]->attrib[FRAG_ATTRIB_COL0]); + COPY_4V(saved_col0[2], v[2]->attrib[FRAG_ATTRIB_COL0]); + } + + if (VB->BackfaceColorPtr->stride) { + if (swsetup->intColors) { + SS_COLOR(v[0]->color, vbcolor[e0]); + SS_COLOR(v[1]->color, vbcolor[e1]); + SS_COLOR(v[2]->color, vbcolor[e2]); + } + else { + COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL0], vbcolor[e0]); + COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL0], vbcolor[e1]); + COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL0], vbcolor[e2]); + } + } + else { + /* flat shade */ + if (swsetup->intColors) { + SS_COLOR(v[0]->color, vbcolor[0]); + SS_COLOR(v[1]->color, vbcolor[0]); + SS_COLOR(v[2]->color, vbcolor[0]); + } + else { + COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL0], vbcolor[0]); + COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL0], vbcolor[0]); + COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL0], vbcolor[0]); + } + } + } + + if (VB->BackfaceSecondaryColorPtr) { + GLfloat (*vbspec)[4] = VB->BackfaceSecondaryColorPtr->data; + + COPY_4V(saved_spec[0], v[0]->attrib[FRAG_ATTRIB_COL1]); + COPY_4V(saved_spec[1], v[1]->attrib[FRAG_ATTRIB_COL1]); + COPY_4V(saved_spec[2], v[2]->attrib[FRAG_ATTRIB_COL1]); + + if (VB->BackfaceSecondaryColorPtr->stride) { + SS_SPEC(v[0]->attrib[FRAG_ATTRIB_COL1], vbspec[e0]); + SS_SPEC(v[1]->attrib[FRAG_ATTRIB_COL1], vbspec[e1]); + SS_SPEC(v[2]->attrib[FRAG_ATTRIB_COL1], vbspec[e2]); + } + else { + SS_SPEC(v[0]->attrib[FRAG_ATTRIB_COL1], vbspec[0]); + SS_SPEC(v[1]->attrib[FRAG_ATTRIB_COL1], vbspec[0]); + SS_SPEC(v[2]->attrib[FRAG_ATTRIB_COL1], vbspec[0]); + } + } + } + } + } + + if (IND & SS_OFFSET_BIT) { + const GLfloat max = ctx->DrawBuffer->_DepthMaxF; + /* save original Z values (restored later) */ + z[0] = v[0]->attrib[FRAG_ATTRIB_WPOS][2]; + z[1] = v[1]->attrib[FRAG_ATTRIB_WPOS][2]; + z[2] = v[2]->attrib[FRAG_ATTRIB_WPOS][2]; + /* Note that Z values are already scaled to [0,65535] (for example) + * so no MRD value is used here. + */ + offset = ctx->Polygon.OffsetUnits; + if (cc * cc > 1e-16) { + const GLfloat ez = z[0] - z[2]; + const GLfloat fz = z[1] - z[2]; + const GLfloat oneOverArea = 1.0F / cc; + const GLfloat dzdx = FABSF((ey * fz - ez * fy) * oneOverArea); + const GLfloat dzdy = FABSF((ez * fx - ex * fz) * oneOverArea); + offset += MAX2(dzdx, dzdy) * ctx->Polygon.OffsetFactor; + } + /* new Z values */ + oz0 = CLAMP(v[0]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0F, max); + oz1 = CLAMP(v[1]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0F, max); + oz2 = CLAMP(v[2]->attrib[FRAG_ATTRIB_WPOS][2] + offset, 0.0F, max); + } + } + + if (mode == GL_POINT) { + if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetPoint) { + v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0; + v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1; + v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2; + } + _swsetup_render_tri(ctx, e0, e1, e2, facing, _swsetup_edge_render_point_tri); + } else if (mode == GL_LINE) { + if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetLine) { + v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0; + v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1; + v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2; + } + _swsetup_render_tri(ctx, e0, e1, e2, facing, _swsetup_edge_render_line_tri); + } else { + if ((IND & SS_OFFSET_BIT) && ctx->Polygon.OffsetFill) { + v[0]->attrib[FRAG_ATTRIB_WPOS][2] = oz0; + v[1]->attrib[FRAG_ATTRIB_WPOS][2] = oz1; + v[2]->attrib[FRAG_ATTRIB_WPOS][2] = oz2; + } + _swrast_Triangle( ctx, v[0], v[1], v[2] ); + } + + /* + * Restore original vertex colors, etc. + */ + if (IND & SS_OFFSET_BIT) { + v[0]->attrib[FRAG_ATTRIB_WPOS][2] = z[0]; + v[1]->attrib[FRAG_ATTRIB_WPOS][2] = z[1]; + v[2]->attrib[FRAG_ATTRIB_WPOS][2] = z[2]; + } + + if (IND & SS_TWOSIDE_BIT) { + if (facing == 1) { + if (VB->BackfaceColorPtr) { + if (swsetup->intColors) { + COPY_CHAN4(v[0]->color, saved_color[0]); + COPY_CHAN4(v[1]->color, saved_color[1]); + COPY_CHAN4(v[2]->color, saved_color[2]); + } + else { + COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL0], saved_col0[0]); + COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL0], saved_col0[1]); + COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL0], saved_col0[2]); + } + } + + if (VB->BackfaceSecondaryColorPtr) { + COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL1], saved_spec[0]); + COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL1], saved_spec[1]); + COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL1], saved_spec[2]); + } + } + } +} + + + +/* Need to fixup edgeflags when decomposing to triangles: + */ +static void TAG(quadfunc)( struct gl_context *ctx, GLuint v0, + GLuint v1, GLuint v2, GLuint v3 ) +{ + if (IND & SS_UNFILLED_BIT) { + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + if (VB->EdgeFlag) { /* XXX this test shouldn't be needed (bug 12614) */ + GLubyte ef1 = VB->EdgeFlag[v1]; + GLubyte ef3 = VB->EdgeFlag[v3]; + VB->EdgeFlag[v1] = 0; + TAG(triangle)( ctx, v0, v1, v3 ); + VB->EdgeFlag[v1] = ef1; + VB->EdgeFlag[v3] = 0; + TAG(triangle)( ctx, v1, v2, v3 ); + VB->EdgeFlag[v3] = ef3; + } + } else { + TAG(triangle)( ctx, v0, v1, v3 ); + TAG(triangle)( ctx, v1, v2, v3 ); + } +} + + + + +static void TAG(init)( void ) +{ + tri_tab[IND] = TAG(triangle); + quad_tab[IND] = TAG(quadfunc); +} + + +#undef IND +#undef TAG diff --git a/mesalib/src/mesa/swrast_setup/ss_vb.h b/mesalib/src/mesa/swrast_setup/ss_vb.h index 35d5da6f7..05e665b5c 100644 --- a/mesalib/src/mesa/swrast_setup/ss_vb.h +++ b/mesalib/src/mesa/swrast_setup/ss_vb.h @@ -1,37 +1,37 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - -#ifndef SS_VB_H -#define SS_VB_H - -struct gl_context; - -void _swsetup_vb_init( struct gl_context *ctx ); -void _swsetup_choose_rastersetup_func( struct gl_context *ctx ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + +#ifndef SS_VB_H +#define SS_VB_H + +struct gl_context; + +void _swsetup_vb_init( struct gl_context *ctx ); +void _swsetup_choose_rastersetup_func( struct gl_context *ctx ); + +#endif diff --git a/mesalib/src/mesa/swrast_setup/swrast_setup.h b/mesalib/src/mesa/swrast_setup/swrast_setup.h index 6d82ef58f..1d87ec108 100644 --- a/mesalib/src/mesa/swrast_setup/swrast_setup.h +++ b/mesalib/src/mesa/swrast_setup/swrast_setup.h @@ -1,61 +1,61 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - -/* Public interface to the swrast_setup module. This module provides - * an implementation of the driver interface to t_vb_render.c, and uses - * the software rasterizer (swrast) to perform actual rasterization. - * - * The internals of the implementation are private, but can be hooked - * into tnl at any time (except between RenderStart/RenderEnd) by - * calling _swsetup_Wakeup(). - */ - -#ifndef SWRAST_SETUP_H -#define SWRAST_SETUP_H - -#include "swrast/swrast.h" - -extern GLboolean -_swsetup_CreateContext( struct gl_context *ctx ); - -extern void -_swsetup_DestroyContext( struct gl_context *ctx ); - -extern void -_swsetup_InvalidateState( struct gl_context *ctx, GLuint new_state ); - -extern void -_swsetup_Wakeup( struct gl_context *ctx ); - -/* Helper function to translate a hardware vertex (as understood by - * the tnl/t_vertex.c code) to a swrast vertex. - */ -extern void -_swsetup_Translate( struct gl_context *ctx, const void *vertex, SWvertex *dest ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + +/* Public interface to the swrast_setup module. This module provides + * an implementation of the driver interface to t_vb_render.c, and uses + * the software rasterizer (swrast) to perform actual rasterization. + * + * The internals of the implementation are private, but can be hooked + * into tnl at any time (except between RenderStart/RenderEnd) by + * calling _swsetup_Wakeup(). + */ + +#ifndef SWRAST_SETUP_H +#define SWRAST_SETUP_H + +#include "swrast/swrast.h" + +extern GLboolean +_swsetup_CreateContext( struct gl_context *ctx ); + +extern void +_swsetup_DestroyContext( struct gl_context *ctx ); + +extern void +_swsetup_InvalidateState( struct gl_context *ctx, GLuint new_state ); + +extern void +_swsetup_Wakeup( struct gl_context *ctx ); + +/* Helper function to translate a hardware vertex (as understood by + * the tnl/t_vertex.c code) to a swrast vertex. + */ +extern void +_swsetup_Translate( struct gl_context *ctx, const void *vertex, SWvertex *dest ); + +#endif diff --git a/mesalib/src/mesa/tnl/t_pipeline.h b/mesalib/src/mesa/tnl/t_pipeline.h index 579207889..0eb03395c 100644 --- a/mesalib/src/mesa/tnl/t_pipeline.h +++ b/mesalib/src/mesa/tnl/t_pipeline.h @@ -1,73 +1,73 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.3 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - - -#ifndef _T_PIPELINE_H_ -#define _T_PIPELINE_H_ - -#include "main/mtypes.h" -#include "t_context.h" - -extern void _tnl_run_pipeline( struct gl_context *ctx ); - -extern void _tnl_destroy_pipeline( struct gl_context *ctx ); - -extern void _tnl_install_pipeline( struct gl_context *ctx, - const struct tnl_pipeline_stage **stages ); - - -/* These are implemented in the t_vb_*.c files: - */ -extern const struct tnl_pipeline_stage _tnl_vertex_transform_stage; -extern const struct tnl_pipeline_stage _tnl_normal_transform_stage; -extern const struct tnl_pipeline_stage _tnl_lighting_stage; -extern const struct tnl_pipeline_stage _tnl_fog_coordinate_stage; -extern const struct tnl_pipeline_stage _tnl_texgen_stage; -extern const struct tnl_pipeline_stage _tnl_texture_transform_stage; -extern const struct tnl_pipeline_stage _tnl_point_attenuation_stage; -extern const struct tnl_pipeline_stage _tnl_vertex_program_stage; -extern const struct tnl_pipeline_stage _tnl_render_stage; - -/* Shorthand to plug in the default pipeline: - */ -extern const struct tnl_pipeline_stage *_tnl_default_pipeline[]; -extern const struct tnl_pipeline_stage *_tnl_vp_pipeline[]; - - -/* Convenience routines provided by t_vb_render.c: - */ -extern tnl_render_func _tnl_render_tab_elts[]; -extern tnl_render_func _tnl_render_tab_verts[]; - -extern void _tnl_RenderClippedPolygon( struct gl_context *ctx, - const GLuint *elts, GLuint n ); - -extern void _tnl_RenderClippedLine( struct gl_context *ctx, GLuint ii, GLuint jj ); - - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.5.3 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + + +#ifndef _T_PIPELINE_H_ +#define _T_PIPELINE_H_ + +#include "main/mtypes.h" +#include "t_context.h" + +extern void _tnl_run_pipeline( struct gl_context *ctx ); + +extern void _tnl_destroy_pipeline( struct gl_context *ctx ); + +extern void _tnl_install_pipeline( struct gl_context *ctx, + const struct tnl_pipeline_stage **stages ); + + +/* These are implemented in the t_vb_*.c files: + */ +extern const struct tnl_pipeline_stage _tnl_vertex_transform_stage; +extern const struct tnl_pipeline_stage _tnl_normal_transform_stage; +extern const struct tnl_pipeline_stage _tnl_lighting_stage; +extern const struct tnl_pipeline_stage _tnl_fog_coordinate_stage; +extern const struct tnl_pipeline_stage _tnl_texgen_stage; +extern const struct tnl_pipeline_stage _tnl_texture_transform_stage; +extern const struct tnl_pipeline_stage _tnl_point_attenuation_stage; +extern const struct tnl_pipeline_stage _tnl_vertex_program_stage; +extern const struct tnl_pipeline_stage _tnl_render_stage; + +/* Shorthand to plug in the default pipeline: + */ +extern const struct tnl_pipeline_stage *_tnl_default_pipeline[]; +extern const struct tnl_pipeline_stage *_tnl_vp_pipeline[]; + + +/* Convenience routines provided by t_vb_render.c: + */ +extern tnl_render_func _tnl_render_tab_elts[]; +extern tnl_render_func _tnl_render_tab_verts[]; + +extern void _tnl_RenderClippedPolygon( struct gl_context *ctx, + const GLuint *elts, GLuint n ); + +extern void _tnl_RenderClippedLine( struct gl_context *ctx, GLuint ii, GLuint jj ); + + +#endif diff --git a/mesalib/src/mesa/tnl/t_vb_fog.c b/mesalib/src/mesa/tnl/t_vb_fog.c index 3e5c99735..cbd8dfc96 100644 --- a/mesalib/src/mesa/tnl/t_vb_fog.c +++ b/mesalib/src/mesa/tnl/t_vb_fog.c @@ -1,276 +1,276 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/macros.h" -#include "main/imports.h" -#include "main/mtypes.h" - -#include "math/m_xform.h" - -#include "t_context.h" -#include "t_pipeline.h" - - -struct fog_stage_data { - GLvector4f fogcoord; /* has actual storage allocated */ -}; - -#define FOG_STAGE_DATA(stage) ((struct fog_stage_data *)stage->privatePtr) - -#define FOG_EXP_TABLE_SIZE 256 -#define FOG_MAX (10.0) -#define EXP_FOG_MAX .0006595 -#define FOG_INCR (FOG_MAX/FOG_EXP_TABLE_SIZE) -static GLfloat exp_table[FOG_EXP_TABLE_SIZE]; -static GLfloat inited = 0; - -#if 1 -#define NEG_EXP( result, narg ) \ -do { \ - GLfloat f = (GLfloat) (narg * (1.0/FOG_INCR)); \ - GLint k = (GLint) f; \ - if (k > FOG_EXP_TABLE_SIZE-2) \ - result = (GLfloat) EXP_FOG_MAX; \ - else \ - result = exp_table[k] + (f-k)*(exp_table[k+1]-exp_table[k]); \ -} while (0) -#else -#define NEG_EXP( result, narg ) \ -do { \ - result = exp(-narg); \ -} while (0) -#endif - - -/** - * Initialize the exp_table[] lookup table for approximating exp(). - */ -static void -init_static_data( void ) -{ - GLfloat f = 0.0F; - GLint i = 0; - for ( ; i < FOG_EXP_TABLE_SIZE ; i++, f += FOG_INCR) { - exp_table[i] = EXPF(-f); - } - inited = 1; -} - - -/** - * Compute per-vertex fog blend factors from fog coordinates by - * evaluating the GL_LINEAR, GL_EXP or GL_EXP2 fog function. - * Fog coordinates are distances from the eye (typically between the - * near and far clip plane distances). - * Note that fogcoords may be negative, if eye z is source absolute - * value must be taken earlier. - * Fog blend factors are in the range [0,1]. - */ -static void -compute_fog_blend_factors(struct gl_context *ctx, GLvector4f *out, const GLvector4f *in) -{ - GLfloat end = ctx->Fog.End; - GLfloat *v = in->start; - GLuint stride = in->stride; - GLuint n = in->count; - GLfloat (*data)[4] = out->data; - GLfloat d; - GLuint i; - - out->count = in->count; - - switch (ctx->Fog.Mode) { - case GL_LINEAR: - if (ctx->Fog.Start == ctx->Fog.End) - d = 1.0F; - else - d = 1.0F / (ctx->Fog.End - ctx->Fog.Start); - for ( i = 0 ; i < n ; i++, STRIDE_F(v, stride)) { - const GLfloat z = *v; - GLfloat f = (end - z) * d; - data[i][0] = CLAMP(f, 0.0F, 1.0F); - } - break; - case GL_EXP: - d = ctx->Fog.Density; - for ( i = 0 ; i < n ; i++, STRIDE_F(v,stride)) { - const GLfloat z = *v; - NEG_EXP( data[i][0], d * z ); - } - break; - case GL_EXP2: - d = ctx->Fog.Density*ctx->Fog.Density; - for ( i = 0 ; i < n ; i++, STRIDE_F(v, stride)) { - const GLfloat z = *v; - NEG_EXP( data[i][0], d * z * z ); - } - break; - default: - _mesa_problem(ctx, "Bad fog mode in make_fog_coord"); - return; - } -} - - -static GLboolean -run_fog_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - struct vertex_buffer *VB = &tnl->vb; - struct fog_stage_data *store = FOG_STAGE_DATA(stage); - GLvector4f *input; - - - if (!ctx->Fog.Enabled) - return GL_TRUE; - - if (ctx->Fog.FogCoordinateSource == GL_FRAGMENT_DEPTH_EXT && !ctx->VertexProgram._Current) { - GLuint i; - GLfloat *coord; - /* Fog is computed from vertex or fragment Z values */ - /* source = VB->AttribPtr[_TNL_ATTRIB_POS] or VB->EyePtr coords */ - /* dest = VB->AttribPtr[_TNL_ATTRIB_FOG] = fog stage private storage */ - VB->AttribPtr[_TNL_ATTRIB_FOG] = &store->fogcoord; - - if (!ctx->_NeedEyeCoords) { - /* compute fog coords from object coords */ - const GLfloat *m = ctx->ModelviewMatrixStack.Top->m; - GLfloat plane[4]; - - /* Use this to store calculated eye z values: - */ - input = &store->fogcoord; - - plane[0] = m[2]; - plane[1] = m[6]; - plane[2] = m[10]; - plane[3] = m[14]; - /* Full eye coords weren't required, just calculate the - * eye Z values. - */ - _mesa_dotprod_tab[VB->AttribPtr[_TNL_ATTRIB_POS]->size] - ( (GLfloat *) input->data, - 4 * sizeof(GLfloat), - VB->AttribPtr[_TNL_ATTRIB_POS], plane ); - - input->count = VB->AttribPtr[_TNL_ATTRIB_POS]->count; - - /* make sure coords are really positive - NOTE should avoid going through array twice */ - coord = input->start; - for (i = 0; i < input->count; i++) { - *coord = FABSF(*coord); - STRIDE_F(coord, input->stride); - } - } - else { - /* fog coordinates = eye Z coordinates - need to copy for ABS */ - input = &store->fogcoord; - - if (VB->EyePtr->size < 2) - _mesa_vector4f_clean_elem( VB->EyePtr, VB->Count, 2 ); - - input->stride = 4 * sizeof(GLfloat); - input->count = VB->EyePtr->count; - coord = VB->EyePtr->start; - for (i = 0 ; i < VB->EyePtr->count; i++) { - input->data[i][0] = FABSF(coord[2]); - STRIDE_F(coord, VB->EyePtr->stride); - } - } - } - else { - /* use glFogCoord() coordinates */ - input = VB->AttribPtr[_TNL_ATTRIB_FOG]; /* source data */ - - /* input->count may be one if glFogCoord was only called once - * before glBegin. But we need to compute fog for all vertices. - */ - input->count = VB->AttribPtr[_TNL_ATTRIB_POS]->count; - - VB->AttribPtr[_TNL_ATTRIB_FOG] = &store->fogcoord; /* dest data */ - } - - if (tnl->_DoVertexFog) { - /* compute blend factors from fog coordinates */ - compute_fog_blend_factors( ctx, VB->AttribPtr[_TNL_ATTRIB_FOG], input ); - } - else { - /* results = incoming fog coords (compute fog per-fragment later) */ - VB->AttribPtr[_TNL_ATTRIB_FOG] = input; - } - - return GL_TRUE; -} - - - -/* Called the first time stage->run() is invoked. - */ -static GLboolean -alloc_fog_data(struct gl_context *ctx, struct tnl_pipeline_stage *stage) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - struct fog_stage_data *store; - stage->privatePtr = MALLOC(sizeof(*store)); - store = FOG_STAGE_DATA(stage); - if (!store) - return GL_FALSE; - - _mesa_vector4f_alloc( &store->fogcoord, 0, tnl->vb.Size, 32 ); - - if (!inited) - init_static_data(); - - return GL_TRUE; -} - - -static void -free_fog_data(struct tnl_pipeline_stage *stage) -{ - struct fog_stage_data *store = FOG_STAGE_DATA(stage); - if (store) { - _mesa_vector4f_free( &store->fogcoord ); - FREE( store ); - stage->privatePtr = NULL; - } -} - - -const struct tnl_pipeline_stage _tnl_fog_coordinate_stage = -{ - "build fog coordinates", /* name */ - NULL, /* private_data */ - alloc_fog_data, /* dtr */ - free_fog_data, /* dtr */ - NULL, /* check */ - run_fog_stage /* run -- initially set to init. */ -}; +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/macros.h" +#include "main/imports.h" +#include "main/mtypes.h" + +#include "math/m_xform.h" + +#include "t_context.h" +#include "t_pipeline.h" + + +struct fog_stage_data { + GLvector4f fogcoord; /* has actual storage allocated */ +}; + +#define FOG_STAGE_DATA(stage) ((struct fog_stage_data *)stage->privatePtr) + +#define FOG_EXP_TABLE_SIZE 256 +#define FOG_MAX (10.0) +#define EXP_FOG_MAX .0006595 +#define FOG_INCR (FOG_MAX/FOG_EXP_TABLE_SIZE) +static GLfloat exp_table[FOG_EXP_TABLE_SIZE]; +static GLfloat inited = 0; + +#if 1 +#define NEG_EXP( result, narg ) \ +do { \ + GLfloat f = (GLfloat) (narg * (1.0/FOG_INCR)); \ + GLint k = (GLint) f; \ + if (k > FOG_EXP_TABLE_SIZE-2) \ + result = (GLfloat) EXP_FOG_MAX; \ + else \ + result = exp_table[k] + (f-k)*(exp_table[k+1]-exp_table[k]); \ +} while (0) +#else +#define NEG_EXP( result, narg ) \ +do { \ + result = exp(-narg); \ +} while (0) +#endif + + +/** + * Initialize the exp_table[] lookup table for approximating exp(). + */ +static void +init_static_data( void ) +{ + GLfloat f = 0.0F; + GLint i = 0; + for ( ; i < FOG_EXP_TABLE_SIZE ; i++, f += FOG_INCR) { + exp_table[i] = EXPF(-f); + } + inited = 1; +} + + +/** + * Compute per-vertex fog blend factors from fog coordinates by + * evaluating the GL_LINEAR, GL_EXP or GL_EXP2 fog function. + * Fog coordinates are distances from the eye (typically between the + * near and far clip plane distances). + * Note that fogcoords may be negative, if eye z is source absolute + * value must be taken earlier. + * Fog blend factors are in the range [0,1]. + */ +static void +compute_fog_blend_factors(struct gl_context *ctx, GLvector4f *out, const GLvector4f *in) +{ + GLfloat end = ctx->Fog.End; + GLfloat *v = in->start; + GLuint stride = in->stride; + GLuint n = in->count; + GLfloat (*data)[4] = out->data; + GLfloat d; + GLuint i; + + out->count = in->count; + + switch (ctx->Fog.Mode) { + case GL_LINEAR: + if (ctx->Fog.Start == ctx->Fog.End) + d = 1.0F; + else + d = 1.0F / (ctx->Fog.End - ctx->Fog.Start); + for ( i = 0 ; i < n ; i++, STRIDE_F(v, stride)) { + const GLfloat z = *v; + GLfloat f = (end - z) * d; + data[i][0] = CLAMP(f, 0.0F, 1.0F); + } + break; + case GL_EXP: + d = ctx->Fog.Density; + for ( i = 0 ; i < n ; i++, STRIDE_F(v,stride)) { + const GLfloat z = *v; + NEG_EXP( data[i][0], d * z ); + } + break; + case GL_EXP2: + d = ctx->Fog.Density*ctx->Fog.Density; + for ( i = 0 ; i < n ; i++, STRIDE_F(v, stride)) { + const GLfloat z = *v; + NEG_EXP( data[i][0], d * z * z ); + } + break; + default: + _mesa_problem(ctx, "Bad fog mode in make_fog_coord"); + return; + } +} + + +static GLboolean +run_fog_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + struct vertex_buffer *VB = &tnl->vb; + struct fog_stage_data *store = FOG_STAGE_DATA(stage); + GLvector4f *input; + + + if (!ctx->Fog.Enabled) + return GL_TRUE; + + if (ctx->Fog.FogCoordinateSource == GL_FRAGMENT_DEPTH_EXT && !ctx->VertexProgram._Current) { + GLuint i; + GLfloat *coord; + /* Fog is computed from vertex or fragment Z values */ + /* source = VB->AttribPtr[_TNL_ATTRIB_POS] or VB->EyePtr coords */ + /* dest = VB->AttribPtr[_TNL_ATTRIB_FOG] = fog stage private storage */ + VB->AttribPtr[_TNL_ATTRIB_FOG] = &store->fogcoord; + + if (!ctx->_NeedEyeCoords) { + /* compute fog coords from object coords */ + const GLfloat *m = ctx->ModelviewMatrixStack.Top->m; + GLfloat plane[4]; + + /* Use this to store calculated eye z values: + */ + input = &store->fogcoord; + + plane[0] = m[2]; + plane[1] = m[6]; + plane[2] = m[10]; + plane[3] = m[14]; + /* Full eye coords weren't required, just calculate the + * eye Z values. + */ + _mesa_dotprod_tab[VB->AttribPtr[_TNL_ATTRIB_POS]->size] + ( (GLfloat *) input->data, + 4 * sizeof(GLfloat), + VB->AttribPtr[_TNL_ATTRIB_POS], plane ); + + input->count = VB->AttribPtr[_TNL_ATTRIB_POS]->count; + + /* make sure coords are really positive + NOTE should avoid going through array twice */ + coord = input->start; + for (i = 0; i < input->count; i++) { + *coord = FABSF(*coord); + STRIDE_F(coord, input->stride); + } + } + else { + /* fog coordinates = eye Z coordinates - need to copy for ABS */ + input = &store->fogcoord; + + if (VB->EyePtr->size < 2) + _mesa_vector4f_clean_elem( VB->EyePtr, VB->Count, 2 ); + + input->stride = 4 * sizeof(GLfloat); + input->count = VB->EyePtr->count; + coord = VB->EyePtr->start; + for (i = 0 ; i < VB->EyePtr->count; i++) { + input->data[i][0] = FABSF(coord[2]); + STRIDE_F(coord, VB->EyePtr->stride); + } + } + } + else { + /* use glFogCoord() coordinates */ + input = VB->AttribPtr[_TNL_ATTRIB_FOG]; /* source data */ + + /* input->count may be one if glFogCoord was only called once + * before glBegin. But we need to compute fog for all vertices. + */ + input->count = VB->AttribPtr[_TNL_ATTRIB_POS]->count; + + VB->AttribPtr[_TNL_ATTRIB_FOG] = &store->fogcoord; /* dest data */ + } + + if (tnl->_DoVertexFog) { + /* compute blend factors from fog coordinates */ + compute_fog_blend_factors( ctx, VB->AttribPtr[_TNL_ATTRIB_FOG], input ); + } + else { + /* results = incoming fog coords (compute fog per-fragment later) */ + VB->AttribPtr[_TNL_ATTRIB_FOG] = input; + } + + return GL_TRUE; +} + + + +/* Called the first time stage->run() is invoked. + */ +static GLboolean +alloc_fog_data(struct gl_context *ctx, struct tnl_pipeline_stage *stage) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + struct fog_stage_data *store; + stage->privatePtr = MALLOC(sizeof(*store)); + store = FOG_STAGE_DATA(stage); + if (!store) + return GL_FALSE; + + _mesa_vector4f_alloc( &store->fogcoord, 0, tnl->vb.Size, 32 ); + + if (!inited) + init_static_data(); + + return GL_TRUE; +} + + +static void +free_fog_data(struct tnl_pipeline_stage *stage) +{ + struct fog_stage_data *store = FOG_STAGE_DATA(stage); + if (store) { + _mesa_vector4f_free( &store->fogcoord ); + FREE( store ); + stage->privatePtr = NULL; + } +} + + +const struct tnl_pipeline_stage _tnl_fog_coordinate_stage = +{ + "build fog coordinates", /* name */ + NULL, /* private_data */ + alloc_fog_data, /* dtr */ + free_fog_data, /* dtr */ + NULL, /* check */ + run_fog_stage /* run -- initially set to init. */ +}; diff --git a/mesalib/src/mesa/tnl/t_vb_normals.c b/mesalib/src/mesa/tnl/t_vb_normals.c index 94e8d0ab5..c19b48e51 100644 --- a/mesalib/src/mesa/tnl/t_vb_normals.c +++ b/mesalib/src/mesa/tnl/t_vb_normals.c @@ -1,187 +1,187 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/macros.h" -#include "main/imports.h" -#include "main/mtypes.h" - -#include "math/m_xform.h" - -#include "t_context.h" -#include "t_pipeline.h" - - -struct normal_stage_data { - normal_func NormalTransform; - GLvector4f normal; -}; - -#define NORMAL_STAGE_DATA(stage) ((struct normal_stage_data *)stage->privatePtr) - - -static GLboolean -run_normal_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) -{ - struct normal_stage_data *store = NORMAL_STAGE_DATA(stage); - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - const GLfloat *lengths; - - if (!store->NormalTransform) - return GL_TRUE; - - /* We can only use the display list's saved normal lengths if we've - * got a transformation matrix with uniform scaling. - */ - if (_math_matrix_is_general_scale(ctx->ModelviewMatrixStack.Top)) - lengths = NULL; - else - lengths = VB->NormalLengthPtr; - - store->NormalTransform( ctx->ModelviewMatrixStack.Top, - ctx->_ModelViewInvScale, - VB->AttribPtr[_TNL_ATTRIB_NORMAL], /* input normals */ - lengths, - &store->normal ); /* resulting normals */ - - if (VB->AttribPtr[_TNL_ATTRIB_NORMAL]->count > 1) { - store->normal.stride = 4 * sizeof(GLfloat); - } - else { - store->normal.stride = 0; - } - - VB->AttribPtr[_TNL_ATTRIB_NORMAL] = &store->normal; - - VB->NormalLengthPtr = NULL; /* no longer valid */ - return GL_TRUE; -} - - -/** - * Examine current GL state and set the store->NormalTransform pointer - * to point to the appropriate normal transformation routine. - */ -static void -validate_normal_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) -{ - struct normal_stage_data *store = NORMAL_STAGE_DATA(stage); - - if (ctx->VertexProgram._Current || - (!ctx->Light.Enabled && - !(ctx->Texture._GenFlags & TEXGEN_NEED_NORMALS))) { - store->NormalTransform = NULL; - return; - } - - if (ctx->_NeedEyeCoords) { - /* Eye coordinates are needed, for whatever reasons. - * Do lighting in eye coordinates, as the GL spec says. - */ - GLuint transform = NORM_TRANSFORM_NO_ROT; - - if (_math_matrix_has_rotation(ctx->ModelviewMatrixStack.Top)) { - /* need to do full (3x3) matrix transform */ - transform = NORM_TRANSFORM; - } - - if (ctx->Transform.Normalize) { - store->NormalTransform = _mesa_normal_tab[transform | NORM_NORMALIZE]; - } - else if (ctx->Transform.RescaleNormals && - ctx->_ModelViewInvScale != 1.0) { - store->NormalTransform = _mesa_normal_tab[transform | NORM_RESCALE]; - } - else { - store->NormalTransform = _mesa_normal_tab[transform]; - } - } - else { - /* We don't need eye coordinates. - * Do lighting in object coordinates. Thus, we don't need to fully - * transform normal vectors (just leave them in object coordinates) - * but we still need to do normalization/rescaling if enabled. - */ - if (ctx->Transform.Normalize) { - store->NormalTransform = _mesa_normal_tab[NORM_NORMALIZE]; - } - else if (!ctx->Transform.RescaleNormals && - ctx->_ModelViewInvScale != 1.0) { - store->NormalTransform = _mesa_normal_tab[NORM_RESCALE]; - } - else { - store->NormalTransform = NULL; - } - } -} - - -/** - * Allocate stage's private data (storage for transformed normals). - */ -static GLboolean -alloc_normal_data(struct gl_context *ctx, struct tnl_pipeline_stage *stage) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - struct normal_stage_data *store; - - stage->privatePtr = malloc(sizeof(*store)); - store = NORMAL_STAGE_DATA(stage); - if (!store) - return GL_FALSE; - - _mesa_vector4f_alloc( &store->normal, 0, tnl->vb.Size, 32 ); - return GL_TRUE; -} - - -/** - * Free stage's private data. - */ -static void -free_normal_data(struct tnl_pipeline_stage *stage) -{ - struct normal_stage_data *store = NORMAL_STAGE_DATA(stage); - if (store) { - _mesa_vector4f_free( &store->normal ); - free( store ); - stage->privatePtr = NULL; - } -} - - -const struct tnl_pipeline_stage _tnl_normal_transform_stage = -{ - "normal transform", /* name */ - NULL, /* privatePtr */ - alloc_normal_data, /* create */ - free_normal_data, /* destroy */ - validate_normal_stage, /* validate */ - run_normal_stage /* run */ -}; +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/macros.h" +#include "main/imports.h" +#include "main/mtypes.h" + +#include "math/m_xform.h" + +#include "t_context.h" +#include "t_pipeline.h" + + +struct normal_stage_data { + normal_func NormalTransform; + GLvector4f normal; +}; + +#define NORMAL_STAGE_DATA(stage) ((struct normal_stage_data *)stage->privatePtr) + + +static GLboolean +run_normal_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) +{ + struct normal_stage_data *store = NORMAL_STAGE_DATA(stage); + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + const GLfloat *lengths; + + if (!store->NormalTransform) + return GL_TRUE; + + /* We can only use the display list's saved normal lengths if we've + * got a transformation matrix with uniform scaling. + */ + if (_math_matrix_is_general_scale(ctx->ModelviewMatrixStack.Top)) + lengths = NULL; + else + lengths = VB->NormalLengthPtr; + + store->NormalTransform( ctx->ModelviewMatrixStack.Top, + ctx->_ModelViewInvScale, + VB->AttribPtr[_TNL_ATTRIB_NORMAL], /* input normals */ + lengths, + &store->normal ); /* resulting normals */ + + if (VB->AttribPtr[_TNL_ATTRIB_NORMAL]->count > 1) { + store->normal.stride = 4 * sizeof(GLfloat); + } + else { + store->normal.stride = 0; + } + + VB->AttribPtr[_TNL_ATTRIB_NORMAL] = &store->normal; + + VB->NormalLengthPtr = NULL; /* no longer valid */ + return GL_TRUE; +} + + +/** + * Examine current GL state and set the store->NormalTransform pointer + * to point to the appropriate normal transformation routine. + */ +static void +validate_normal_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) +{ + struct normal_stage_data *store = NORMAL_STAGE_DATA(stage); + + if (ctx->VertexProgram._Current || + (!ctx->Light.Enabled && + !(ctx->Texture._GenFlags & TEXGEN_NEED_NORMALS))) { + store->NormalTransform = NULL; + return; + } + + if (ctx->_NeedEyeCoords) { + /* Eye coordinates are needed, for whatever reasons. + * Do lighting in eye coordinates, as the GL spec says. + */ + GLuint transform = NORM_TRANSFORM_NO_ROT; + + if (_math_matrix_has_rotation(ctx->ModelviewMatrixStack.Top)) { + /* need to do full (3x3) matrix transform */ + transform = NORM_TRANSFORM; + } + + if (ctx->Transform.Normalize) { + store->NormalTransform = _mesa_normal_tab[transform | NORM_NORMALIZE]; + } + else if (ctx->Transform.RescaleNormals && + ctx->_ModelViewInvScale != 1.0) { + store->NormalTransform = _mesa_normal_tab[transform | NORM_RESCALE]; + } + else { + store->NormalTransform = _mesa_normal_tab[transform]; + } + } + else { + /* We don't need eye coordinates. + * Do lighting in object coordinates. Thus, we don't need to fully + * transform normal vectors (just leave them in object coordinates) + * but we still need to do normalization/rescaling if enabled. + */ + if (ctx->Transform.Normalize) { + store->NormalTransform = _mesa_normal_tab[NORM_NORMALIZE]; + } + else if (!ctx->Transform.RescaleNormals && + ctx->_ModelViewInvScale != 1.0) { + store->NormalTransform = _mesa_normal_tab[NORM_RESCALE]; + } + else { + store->NormalTransform = NULL; + } + } +} + + +/** + * Allocate stage's private data (storage for transformed normals). + */ +static GLboolean +alloc_normal_data(struct gl_context *ctx, struct tnl_pipeline_stage *stage) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + struct normal_stage_data *store; + + stage->privatePtr = malloc(sizeof(*store)); + store = NORMAL_STAGE_DATA(stage); + if (!store) + return GL_FALSE; + + _mesa_vector4f_alloc( &store->normal, 0, tnl->vb.Size, 32 ); + return GL_TRUE; +} + + +/** + * Free stage's private data. + */ +static void +free_normal_data(struct tnl_pipeline_stage *stage) +{ + struct normal_stage_data *store = NORMAL_STAGE_DATA(stage); + if (store) { + _mesa_vector4f_free( &store->normal ); + free( store ); + stage->privatePtr = NULL; + } +} + + +const struct tnl_pipeline_stage _tnl_normal_transform_stage = +{ + "normal transform", /* name */ + NULL, /* privatePtr */ + alloc_normal_data, /* create */ + free_normal_data, /* destroy */ + validate_normal_stage, /* validate */ + run_normal_stage /* run */ +}; diff --git a/mesalib/src/mesa/tnl/t_vb_points.c b/mesalib/src/mesa/tnl/t_vb_points.c index 1bf961a6d..9edbbc708 100644 --- a/mesalib/src/mesa/tnl/t_vb_points.c +++ b/mesalib/src/mesa/tnl/t_vb_points.c @@ -1,114 +1,114 @@ -/* - * Mesa 3-D graphics library - * Version: 7.0 - * - * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Brian Paul - */ - -#include "main/glheader.h" -#include "main/mtypes.h" -#include "main/dd.h" -#include "main/imports.h" -#include "t_context.h" -#include "t_pipeline.h" - - -struct point_stage_data { - GLvector4f PointSize; -}; - -#define POINT_STAGE_DATA(stage) ((struct point_stage_data *)stage->privatePtr) - - -/** - * Compute point size for each vertex from the vertex eye-space Z - * coordinate and the point size attenuation factors. - * Only done when point size attenuation is enabled and vertex program is - * disabled. - */ -static GLboolean -run_point_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) -{ - if (ctx->Point._Attenuated && !ctx->VertexProgram._Current) { - struct point_stage_data *store = POINT_STAGE_DATA(stage); - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - const GLfloat *eyeCoord = (GLfloat *) VB->EyePtr->data + 2; - const GLint eyeCoordStride = VB->EyePtr->stride / sizeof(GLfloat); - const GLfloat p0 = ctx->Point.Params[0]; - const GLfloat p1 = ctx->Point.Params[1]; - const GLfloat p2 = ctx->Point.Params[2]; - const GLfloat pointSize = ctx->Point.Size; - GLfloat (*size)[4] = store->PointSize.data; - GLuint i; - - for (i = 0; i < VB->Count; i++) { - const GLfloat dist = FABSF(*eyeCoord); - const GLfloat q = p0 + dist * (p1 + dist * p2); - const GLfloat atten = (q != 0.0F) ? SQRTF(1.0F / q) : 1.0F; - size[i][0] = pointSize * atten; /* clamping done in rasterization */ - eyeCoord += eyeCoordStride; - } - - VB->AttribPtr[_TNL_ATTRIB_POINTSIZE] = &store->PointSize; - } - - return GL_TRUE; -} - - -static GLboolean -alloc_point_data(struct gl_context *ctx, struct tnl_pipeline_stage *stage) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - struct point_stage_data *store; - stage->privatePtr = malloc(sizeof(*store)); - store = POINT_STAGE_DATA(stage); - if (!store) - return GL_FALSE; - - _mesa_vector4f_alloc( &store->PointSize, 0, VB->Size, 32 ); - return GL_TRUE; -} - - -static void -free_point_data(struct tnl_pipeline_stage *stage) -{ - struct point_stage_data *store = POINT_STAGE_DATA(stage); - if (store) { - _mesa_vector4f_free( &store->PointSize ); - free( store ); - stage->privatePtr = NULL; - } -} - - -const struct tnl_pipeline_stage _tnl_point_attenuation_stage = -{ - "point size attenuation", /* name */ - NULL, /* stage private data */ - alloc_point_data, /* alloc data */ - free_point_data, /* destructor */ - NULL, - run_point_stage /* run */ -}; +/* + * Mesa 3-D graphics library + * Version: 7.0 + * + * Copyright (C) 1999-2007 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Brian Paul + */ + +#include "main/glheader.h" +#include "main/mtypes.h" +#include "main/dd.h" +#include "main/imports.h" +#include "t_context.h" +#include "t_pipeline.h" + + +struct point_stage_data { + GLvector4f PointSize; +}; + +#define POINT_STAGE_DATA(stage) ((struct point_stage_data *)stage->privatePtr) + + +/** + * Compute point size for each vertex from the vertex eye-space Z + * coordinate and the point size attenuation factors. + * Only done when point size attenuation is enabled and vertex program is + * disabled. + */ +static GLboolean +run_point_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage) +{ + if (ctx->Point._Attenuated && !ctx->VertexProgram._Current) { + struct point_stage_data *store = POINT_STAGE_DATA(stage); + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + const GLfloat *eyeCoord = (GLfloat *) VB->EyePtr->data + 2; + const GLint eyeCoordStride = VB->EyePtr->stride / sizeof(GLfloat); + const GLfloat p0 = ctx->Point.Params[0]; + const GLfloat p1 = ctx->Point.Params[1]; + const GLfloat p2 = ctx->Point.Params[2]; + const GLfloat pointSize = ctx->Point.Size; + GLfloat (*size)[4] = store->PointSize.data; + GLuint i; + + for (i = 0; i < VB->Count; i++) { + const GLfloat dist = FABSF(*eyeCoord); + const GLfloat q = p0 + dist * (p1 + dist * p2); + const GLfloat atten = (q != 0.0F) ? SQRTF(1.0F / q) : 1.0F; + size[i][0] = pointSize * atten; /* clamping done in rasterization */ + eyeCoord += eyeCoordStride; + } + + VB->AttribPtr[_TNL_ATTRIB_POINTSIZE] = &store->PointSize; + } + + return GL_TRUE; +} + + +static GLboolean +alloc_point_data(struct gl_context *ctx, struct tnl_pipeline_stage *stage) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + struct point_stage_data *store; + stage->privatePtr = malloc(sizeof(*store)); + store = POINT_STAGE_DATA(stage); + if (!store) + return GL_FALSE; + + _mesa_vector4f_alloc( &store->PointSize, 0, VB->Size, 32 ); + return GL_TRUE; +} + + +static void +free_point_data(struct tnl_pipeline_stage *stage) +{ + struct point_stage_data *store = POINT_STAGE_DATA(stage); + if (store) { + _mesa_vector4f_free( &store->PointSize ); + free( store ); + stage->privatePtr = NULL; + } +} + + +const struct tnl_pipeline_stage _tnl_point_attenuation_stage = +{ + "point size attenuation", /* name */ + NULL, /* stage private data */ + alloc_point_data, /* alloc data */ + free_point_data, /* destructor */ + NULL, + run_point_stage /* run */ +}; diff --git a/mesalib/src/mesa/tnl/t_vb_render.c b/mesalib/src/mesa/tnl/t_vb_render.c index 64dabca98..cb3192132 100644 --- a/mesalib/src/mesa/tnl/t_vb_render.c +++ b/mesalib/src/mesa/tnl/t_vb_render.c @@ -1,348 +1,348 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - -/* - * Render whole vertex buffers, including projection of vertices from - * clip space and clipping of primitives. - * - * This file makes calls to project vertices and to the point, line - * and triangle rasterizers via the function pointers: - * - * context->Driver.Render.* - * - */ - - -#include "main/glheader.h" -#include "main/context.h" -#include "main/enums.h" -#include "main/macros.h" -#include "main/imports.h" -#include "main/mtypes.h" -#include "math/m_xform.h" - -#include "t_pipeline.h" - - - -/**********************************************************************/ -/* Clip single primitives */ -/**********************************************************************/ - - -#define W(i) coord[i][3] -#define Z(i) coord[i][2] -#define Y(i) coord[i][1] -#define X(i) coord[i][0] -#define SIZE 4 -#define TAG(x) x##_4 -#include "t_vb_cliptmp.h" - - - -/**********************************************************************/ -/* Clip and render whole begin/end objects */ -/**********************************************************************/ - -#define NEED_EDGEFLAG_SETUP (ctx->Polygon.FrontMode != GL_FILL || ctx->Polygon.BackMode != GL_FILL) -#define EDGEFLAG_GET(idx) VB->EdgeFlag[idx] -#define EDGEFLAG_SET(idx, val) VB->EdgeFlag[idx] = val - - -/* This does NOT include the CLIP_USER_BIT! */ -#define CLIPMASK (CLIP_FRUSTUM_BITS | CLIP_CULL_BIT) - - -/* Vertices, with the possibility of clipping. - */ -#define RENDER_POINTS( start, count ) \ - tnl->Driver.Render.Points( ctx, start, count ) - -#define RENDER_LINE( v1, v2 ) \ -do { \ - GLubyte c1 = mask[v1], c2 = mask[v2]; \ - GLubyte ormask = c1|c2; \ - if (!ormask) \ - LineFunc( ctx, v1, v2 ); \ - else if (!(c1 & c2 & CLIPMASK)) \ - clip_line_4( ctx, v1, v2, ormask ); \ -} while (0) - -#define RENDER_TRI( v1, v2, v3 ) \ -do { \ - GLubyte c1 = mask[v1], c2 = mask[v2], c3 = mask[v3]; \ - GLubyte ormask = c1|c2|c3; \ - if (!ormask) \ - TriangleFunc( ctx, v1, v2, v3 ); \ - else if (!(c1 & c2 & c3 & CLIPMASK)) \ - clip_tri_4( ctx, v1, v2, v3, ormask ); \ -} while (0) - -#define RENDER_QUAD( v1, v2, v3, v4 ) \ -do { \ - GLubyte c1 = mask[v1], c2 = mask[v2]; \ - GLubyte c3 = mask[v3], c4 = mask[v4]; \ - GLubyte ormask = c1|c2|c3|c4; \ - if (!ormask) \ - QuadFunc( ctx, v1, v2, v3, v4 ); \ - else if (!(c1 & c2 & c3 & c4 & CLIPMASK)) \ - clip_quad_4( ctx, v1, v2, v3, v4, ormask ); \ -} while (0) - - -#define LOCAL_VARS \ - TNLcontext *tnl = TNL_CONTEXT(ctx); \ - struct vertex_buffer *VB = &tnl->vb; \ - const GLuint * const elt = VB->Elts; \ - const GLubyte *mask = VB->ClipMask; \ - const GLuint sz = VB->ClipPtr->size; \ - const tnl_line_func LineFunc = tnl->Driver.Render.Line; \ - const tnl_triangle_func TriangleFunc = tnl->Driver.Render.Triangle; \ - const tnl_quad_func QuadFunc = tnl->Driver.Render.Quad; \ - const GLboolean stipple = ctx->Line.StippleFlag; \ - (void) (LineFunc && TriangleFunc && QuadFunc); \ - (void) elt; (void) mask; (void) sz; (void) stipple; - -#define TAG(x) clip_##x##_verts -#define INIT(x) tnl->Driver.Render.PrimitiveNotify( ctx, x ) -#define RESET_STIPPLE if (stipple) tnl->Driver.Render.ResetLineStipple( ctx ) -#define PRESERVE_VB_DEFS -#include "t_vb_rendertmp.h" - - - -/* Elts, with the possibility of clipping. - */ -#undef ELT -#undef TAG -#define ELT(x) elt[x] -#define TAG(x) clip_##x##_elts -#include "t_vb_rendertmp.h" - -/* TODO: do this for all primitives, verts and elts: - */ -static void clip_elt_triangles( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - tnl_render_func render_tris = tnl->Driver.Render.PrimTabElts[GL_TRIANGLES]; - struct vertex_buffer *VB = &tnl->vb; - const GLuint * const elt = VB->Elts; - GLubyte *mask = VB->ClipMask; - GLuint last = count-2; - GLuint j; - (void) flags; - - tnl->Driver.Render.PrimitiveNotify( ctx, GL_TRIANGLES ); - - for (j=start; j < last; j+=3 ) { - GLubyte c1 = mask[elt[j]]; - GLubyte c2 = mask[elt[j+1]]; - GLubyte c3 = mask[elt[j+2]]; - GLubyte ormask = c1|c2|c3; - if (ormask) { - if (start < j) - render_tris( ctx, start, j, 0 ); - if (!(c1&c2&c3&CLIPMASK)) - clip_tri_4( ctx, elt[j], elt[j+1], elt[j+2], ormask ); - start = j+3; - } - } - - if (start < j) - render_tris( ctx, start, j, 0 ); -} - -/**********************************************************************/ -/* Render whole begin/end objects */ -/**********************************************************************/ - -#define NEED_EDGEFLAG_SETUP (ctx->Polygon.FrontMode != GL_FILL || ctx->Polygon.BackMode != GL_FILL) -#define EDGEFLAG_GET(idx) VB->EdgeFlag[idx] -#define EDGEFLAG_SET(idx, val) VB->EdgeFlag[idx] = val - - -/* Vertices, no clipping. - */ -#define RENDER_POINTS( start, count ) \ - tnl->Driver.Render.Points( ctx, start, count ) - -#define RENDER_LINE( v1, v2 ) \ - LineFunc( ctx, v1, v2 ) - -#define RENDER_TRI( v1, v2, v3 ) \ - TriangleFunc( ctx, v1, v2, v3 ) - -#define RENDER_QUAD( v1, v2, v3, v4 ) \ - QuadFunc( ctx, v1, v2, v3, v4 ) - -#define TAG(x) _tnl_##x##_verts - -#define LOCAL_VARS \ - TNLcontext *tnl = TNL_CONTEXT(ctx); \ - struct vertex_buffer *VB = &tnl->vb; \ - const GLuint * const elt = VB->Elts; \ - const tnl_line_func LineFunc = tnl->Driver.Render.Line; \ - const tnl_triangle_func TriangleFunc = tnl->Driver.Render.Triangle; \ - const tnl_quad_func QuadFunc = tnl->Driver.Render.Quad; \ - const GLboolean stipple = ctx->Line.StippleFlag; \ - (void) (LineFunc && TriangleFunc && QuadFunc); \ - (void) elt; (void) stipple - -#define RESET_STIPPLE if (stipple) tnl->Driver.Render.ResetLineStipple( ctx ) -#define INIT(x) tnl->Driver.Render.PrimitiveNotify( ctx, x ) -#define RENDER_TAB_QUALIFIER -#define PRESERVE_VB_DEFS -#include "t_vb_rendertmp.h" - - -/* Elts, no clipping. - */ -#undef ELT -#define TAG(x) _tnl_##x##_elts -#define ELT(x) elt[x] -#include "t_vb_rendertmp.h" - - -/**********************************************************************/ -/* Helper functions for drivers */ -/**********************************************************************/ - -void _tnl_RenderClippedPolygon( struct gl_context *ctx, const GLuint *elts, GLuint n ) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - struct vertex_buffer *VB = &tnl->vb; - GLuint *tmp = VB->Elts; - - VB->Elts = (GLuint *)elts; - tnl->Driver.Render.PrimTabElts[GL_POLYGON]( ctx, 0, n, PRIM_BEGIN|PRIM_END ); - VB->Elts = tmp; -} - -void _tnl_RenderClippedLine( struct gl_context *ctx, GLuint ii, GLuint jj ) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - tnl->Driver.Render.Line( ctx, ii, jj ); -} - - - -/**********************************************************************/ -/* Clip and render whole vertex buffers */ -/**********************************************************************/ - - -static GLboolean run_render( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - struct vertex_buffer *VB = &tnl->vb; - tnl_render_func *tab; - GLint pass = 0; - - /* Allow the drivers to lock before projected verts are built so - * that window coordinates are guarenteed not to change before - * rendering. - */ - ASSERT(tnl->Driver.Render.Start); - - tnl->Driver.Render.Start( ctx ); - - ASSERT(tnl->Driver.Render.BuildVertices); - ASSERT(tnl->Driver.Render.PrimitiveNotify); - ASSERT(tnl->Driver.Render.Points); - ASSERT(tnl->Driver.Render.Line); - ASSERT(tnl->Driver.Render.Triangle); - ASSERT(tnl->Driver.Render.Quad); - ASSERT(tnl->Driver.Render.ResetLineStipple); - ASSERT(tnl->Driver.Render.Interp); - ASSERT(tnl->Driver.Render.CopyPV); - ASSERT(tnl->Driver.Render.ClippedLine); - ASSERT(tnl->Driver.Render.ClippedPolygon); - ASSERT(tnl->Driver.Render.Finish); - - tnl->Driver.Render.BuildVertices( ctx, 0, VB->Count, ~0 ); - - if (VB->ClipOrMask) { - tab = VB->Elts ? clip_render_tab_elts : clip_render_tab_verts; - clip_render_tab_elts[GL_TRIANGLES] = clip_elt_triangles; - } - else { - tab = (VB->Elts ? - tnl->Driver.Render.PrimTabElts : - tnl->Driver.Render.PrimTabVerts); - } - - do - { - GLuint i; - - for (i = 0 ; i < VB->PrimitiveCount ; i++) - { - GLuint prim = _tnl_translate_prim(&VB->Primitive[i]); - GLuint start = VB->Primitive[i].start; - GLuint length = VB->Primitive[i].count; - - assert((prim & PRIM_MODE_MASK) <= GL_POLYGON); - - if (MESA_VERBOSE & VERBOSE_PRIMS) - _mesa_debug(NULL, "MESA prim %s %d..%d\n", - _mesa_lookup_enum_by_nr(prim & PRIM_MODE_MASK), - start, start+length); - - if (length) - tab[prim & PRIM_MODE_MASK]( ctx, start, start + length, prim ); - } - } while (tnl->Driver.Render.Multipass && - tnl->Driver.Render.Multipass( ctx, ++pass )); - - tnl->Driver.Render.Finish( ctx ); - - return GL_FALSE; /* finished the pipe */ -} - - -/**********************************************************************/ -/* Render pipeline stage */ -/**********************************************************************/ - - - - - -const struct tnl_pipeline_stage _tnl_render_stage = -{ - "render", /* name */ - NULL, /* private data */ - NULL, /* creator */ - NULL, /* destructor */ - NULL, /* validate */ - run_render /* run */ -}; +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + +/* + * Render whole vertex buffers, including projection of vertices from + * clip space and clipping of primitives. + * + * This file makes calls to project vertices and to the point, line + * and triangle rasterizers via the function pointers: + * + * context->Driver.Render.* + * + */ + + +#include "main/glheader.h" +#include "main/context.h" +#include "main/enums.h" +#include "main/macros.h" +#include "main/imports.h" +#include "main/mtypes.h" +#include "math/m_xform.h" + +#include "t_pipeline.h" + + + +/**********************************************************************/ +/* Clip single primitives */ +/**********************************************************************/ + + +#define W(i) coord[i][3] +#define Z(i) coord[i][2] +#define Y(i) coord[i][1] +#define X(i) coord[i][0] +#define SIZE 4 +#define TAG(x) x##_4 +#include "t_vb_cliptmp.h" + + + +/**********************************************************************/ +/* Clip and render whole begin/end objects */ +/**********************************************************************/ + +#define NEED_EDGEFLAG_SETUP (ctx->Polygon.FrontMode != GL_FILL || ctx->Polygon.BackMode != GL_FILL) +#define EDGEFLAG_GET(idx) VB->EdgeFlag[idx] +#define EDGEFLAG_SET(idx, val) VB->EdgeFlag[idx] = val + + +/* This does NOT include the CLIP_USER_BIT! */ +#define CLIPMASK (CLIP_FRUSTUM_BITS | CLIP_CULL_BIT) + + +/* Vertices, with the possibility of clipping. + */ +#define RENDER_POINTS( start, count ) \ + tnl->Driver.Render.Points( ctx, start, count ) + +#define RENDER_LINE( v1, v2 ) \ +do { \ + GLubyte c1 = mask[v1], c2 = mask[v2]; \ + GLubyte ormask = c1|c2; \ + if (!ormask) \ + LineFunc( ctx, v1, v2 ); \ + else if (!(c1 & c2 & CLIPMASK)) \ + clip_line_4( ctx, v1, v2, ormask ); \ +} while (0) + +#define RENDER_TRI( v1, v2, v3 ) \ +do { \ + GLubyte c1 = mask[v1], c2 = mask[v2], c3 = mask[v3]; \ + GLubyte ormask = c1|c2|c3; \ + if (!ormask) \ + TriangleFunc( ctx, v1, v2, v3 ); \ + else if (!(c1 & c2 & c3 & CLIPMASK)) \ + clip_tri_4( ctx, v1, v2, v3, ormask ); \ +} while (0) + +#define RENDER_QUAD( v1, v2, v3, v4 ) \ +do { \ + GLubyte c1 = mask[v1], c2 = mask[v2]; \ + GLubyte c3 = mask[v3], c4 = mask[v4]; \ + GLubyte ormask = c1|c2|c3|c4; \ + if (!ormask) \ + QuadFunc( ctx, v1, v2, v3, v4 ); \ + else if (!(c1 & c2 & c3 & c4 & CLIPMASK)) \ + clip_quad_4( ctx, v1, v2, v3, v4, ormask ); \ +} while (0) + + +#define LOCAL_VARS \ + TNLcontext *tnl = TNL_CONTEXT(ctx); \ + struct vertex_buffer *VB = &tnl->vb; \ + const GLuint * const elt = VB->Elts; \ + const GLubyte *mask = VB->ClipMask; \ + const GLuint sz = VB->ClipPtr->size; \ + const tnl_line_func LineFunc = tnl->Driver.Render.Line; \ + const tnl_triangle_func TriangleFunc = tnl->Driver.Render.Triangle; \ + const tnl_quad_func QuadFunc = tnl->Driver.Render.Quad; \ + const GLboolean stipple = ctx->Line.StippleFlag; \ + (void) (LineFunc && TriangleFunc && QuadFunc); \ + (void) elt; (void) mask; (void) sz; (void) stipple; + +#define TAG(x) clip_##x##_verts +#define INIT(x) tnl->Driver.Render.PrimitiveNotify( ctx, x ) +#define RESET_STIPPLE if (stipple) tnl->Driver.Render.ResetLineStipple( ctx ) +#define PRESERVE_VB_DEFS +#include "t_vb_rendertmp.h" + + + +/* Elts, with the possibility of clipping. + */ +#undef ELT +#undef TAG +#define ELT(x) elt[x] +#define TAG(x) clip_##x##_elts +#include "t_vb_rendertmp.h" + +/* TODO: do this for all primitives, verts and elts: + */ +static void clip_elt_triangles( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + tnl_render_func render_tris = tnl->Driver.Render.PrimTabElts[GL_TRIANGLES]; + struct vertex_buffer *VB = &tnl->vb; + const GLuint * const elt = VB->Elts; + GLubyte *mask = VB->ClipMask; + GLuint last = count-2; + GLuint j; + (void) flags; + + tnl->Driver.Render.PrimitiveNotify( ctx, GL_TRIANGLES ); + + for (j=start; j < last; j+=3 ) { + GLubyte c1 = mask[elt[j]]; + GLubyte c2 = mask[elt[j+1]]; + GLubyte c3 = mask[elt[j+2]]; + GLubyte ormask = c1|c2|c3; + if (ormask) { + if (start < j) + render_tris( ctx, start, j, 0 ); + if (!(c1&c2&c3&CLIPMASK)) + clip_tri_4( ctx, elt[j], elt[j+1], elt[j+2], ormask ); + start = j+3; + } + } + + if (start < j) + render_tris( ctx, start, j, 0 ); +} + +/**********************************************************************/ +/* Render whole begin/end objects */ +/**********************************************************************/ + +#define NEED_EDGEFLAG_SETUP (ctx->Polygon.FrontMode != GL_FILL || ctx->Polygon.BackMode != GL_FILL) +#define EDGEFLAG_GET(idx) VB->EdgeFlag[idx] +#define EDGEFLAG_SET(idx, val) VB->EdgeFlag[idx] = val + + +/* Vertices, no clipping. + */ +#define RENDER_POINTS( start, count ) \ + tnl->Driver.Render.Points( ctx, start, count ) + +#define RENDER_LINE( v1, v2 ) \ + LineFunc( ctx, v1, v2 ) + +#define RENDER_TRI( v1, v2, v3 ) \ + TriangleFunc( ctx, v1, v2, v3 ) + +#define RENDER_QUAD( v1, v2, v3, v4 ) \ + QuadFunc( ctx, v1, v2, v3, v4 ) + +#define TAG(x) _tnl_##x##_verts + +#define LOCAL_VARS \ + TNLcontext *tnl = TNL_CONTEXT(ctx); \ + struct vertex_buffer *VB = &tnl->vb; \ + const GLuint * const elt = VB->Elts; \ + const tnl_line_func LineFunc = tnl->Driver.Render.Line; \ + const tnl_triangle_func TriangleFunc = tnl->Driver.Render.Triangle; \ + const tnl_quad_func QuadFunc = tnl->Driver.Render.Quad; \ + const GLboolean stipple = ctx->Line.StippleFlag; \ + (void) (LineFunc && TriangleFunc && QuadFunc); \ + (void) elt; (void) stipple + +#define RESET_STIPPLE if (stipple) tnl->Driver.Render.ResetLineStipple( ctx ) +#define INIT(x) tnl->Driver.Render.PrimitiveNotify( ctx, x ) +#define RENDER_TAB_QUALIFIER +#define PRESERVE_VB_DEFS +#include "t_vb_rendertmp.h" + + +/* Elts, no clipping. + */ +#undef ELT +#define TAG(x) _tnl_##x##_elts +#define ELT(x) elt[x] +#include "t_vb_rendertmp.h" + + +/**********************************************************************/ +/* Helper functions for drivers */ +/**********************************************************************/ + +void _tnl_RenderClippedPolygon( struct gl_context *ctx, const GLuint *elts, GLuint n ) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + struct vertex_buffer *VB = &tnl->vb; + GLuint *tmp = VB->Elts; + + VB->Elts = (GLuint *)elts; + tnl->Driver.Render.PrimTabElts[GL_POLYGON]( ctx, 0, n, PRIM_BEGIN|PRIM_END ); + VB->Elts = tmp; +} + +void _tnl_RenderClippedLine( struct gl_context *ctx, GLuint ii, GLuint jj ) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + tnl->Driver.Render.Line( ctx, ii, jj ); +} + + + +/**********************************************************************/ +/* Clip and render whole vertex buffers */ +/**********************************************************************/ + + +static GLboolean run_render( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + struct vertex_buffer *VB = &tnl->vb; + tnl_render_func *tab; + GLint pass = 0; + + /* Allow the drivers to lock before projected verts are built so + * that window coordinates are guarenteed not to change before + * rendering. + */ + ASSERT(tnl->Driver.Render.Start); + + tnl->Driver.Render.Start( ctx ); + + ASSERT(tnl->Driver.Render.BuildVertices); + ASSERT(tnl->Driver.Render.PrimitiveNotify); + ASSERT(tnl->Driver.Render.Points); + ASSERT(tnl->Driver.Render.Line); + ASSERT(tnl->Driver.Render.Triangle); + ASSERT(tnl->Driver.Render.Quad); + ASSERT(tnl->Driver.Render.ResetLineStipple); + ASSERT(tnl->Driver.Render.Interp); + ASSERT(tnl->Driver.Render.CopyPV); + ASSERT(tnl->Driver.Render.ClippedLine); + ASSERT(tnl->Driver.Render.ClippedPolygon); + ASSERT(tnl->Driver.Render.Finish); + + tnl->Driver.Render.BuildVertices( ctx, 0, VB->Count, ~0 ); + + if (VB->ClipOrMask) { + tab = VB->Elts ? clip_render_tab_elts : clip_render_tab_verts; + clip_render_tab_elts[GL_TRIANGLES] = clip_elt_triangles; + } + else { + tab = (VB->Elts ? + tnl->Driver.Render.PrimTabElts : + tnl->Driver.Render.PrimTabVerts); + } + + do + { + GLuint i; + + for (i = 0 ; i < VB->PrimitiveCount ; i++) + { + GLuint prim = _tnl_translate_prim(&VB->Primitive[i]); + GLuint start = VB->Primitive[i].start; + GLuint length = VB->Primitive[i].count; + + assert((prim & PRIM_MODE_MASK) <= GL_POLYGON); + + if (MESA_VERBOSE & VERBOSE_PRIMS) + _mesa_debug(NULL, "MESA prim %s %d..%d\n", + _mesa_lookup_enum_by_nr(prim & PRIM_MODE_MASK), + start, start+length); + + if (length) + tab[prim & PRIM_MODE_MASK]( ctx, start, start + length, prim ); + } + } while (tnl->Driver.Render.Multipass && + tnl->Driver.Render.Multipass( ctx, ++pass )); + + tnl->Driver.Render.Finish( ctx ); + + return GL_FALSE; /* finished the pipe */ +} + + +/**********************************************************************/ +/* Render pipeline stage */ +/**********************************************************************/ + + + + + +const struct tnl_pipeline_stage _tnl_render_stage = +{ + "render", /* name */ + NULL, /* private data */ + NULL, /* creator */ + NULL, /* destructor */ + NULL, /* validate */ + run_render /* run */ +}; diff --git a/mesalib/src/mesa/tnl/t_vb_rendertmp.h b/mesalib/src/mesa/tnl/t_vb_rendertmp.h index 92dbf219d..4ed485a7b 100644 --- a/mesalib/src/mesa/tnl/t_vb_rendertmp.h +++ b/mesalib/src/mesa/tnl/t_vb_rendertmp.h @@ -1,486 +1,486 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - -#ifndef POSTFIX -#define POSTFIX -#endif - -#ifndef INIT -#define INIT(x) -#endif - -#ifndef NEED_EDGEFLAG_SETUP -#define NEED_EDGEFLAG_SETUP 0 -#define EDGEFLAG_GET(a) 0 -#define EDGEFLAG_SET(a,b) (void)b -#endif - -#ifndef RESET_STIPPLE -#define RESET_STIPPLE -#endif - -#ifndef TEST_PRIM_END -#define TEST_PRIM_END(prim) (flags & PRIM_END) -#define TEST_PRIM_BEGIN(prim) (flags & PRIM_BEGIN) -#endif - -#ifndef ELT -#define ELT(x) x -#endif - -#ifndef RENDER_TAB_QUALIFIER -#define RENDER_TAB_QUALIFIER static -#endif - -static void TAG(render_points)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - LOCAL_VARS; - (void) flags; - - INIT(GL_POINTS); - RENDER_POINTS( start, count ); - POSTFIX; -} - -static void TAG(render_lines)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint j; - LOCAL_VARS; - (void) flags; - - INIT(GL_LINES); - for (j=start+1; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_LINE( ELT(j-1), ELT(j) ); - else - RENDER_LINE( ELT(j), ELT(j-1) ); - } - POSTFIX; -} - - -static void TAG(render_line_strip)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint j; - LOCAL_VARS; - (void) flags; - - INIT(GL_LINE_STRIP); - - if (TEST_PRIM_BEGIN(flags)) { - RESET_STIPPLE; - } - - for (j=start+1; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_LINE( ELT(j-1), ELT(j) ); - else - RENDER_LINE( ELT(j), ELT(j-1) ); - } - POSTFIX; -} - - -static void TAG(render_line_loop)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint i; - LOCAL_VARS; - - (void) flags; - - INIT(GL_LINE_LOOP); - - if (start+1 < count) { - if (TEST_PRIM_BEGIN(flags)) { - RESET_STIPPLE; - if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_LINE( ELT(start), ELT(start+1) ); - else - RENDER_LINE( ELT(start+1), ELT(start) ); - } - - for ( i = start+2 ; i < count ; i++) { - if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_LINE( ELT(i-1), ELT(i) ); - else - RENDER_LINE( ELT(i), ELT(i-1) ); - } - - if ( TEST_PRIM_END(flags)) { - if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_LINE( ELT(count-1), ELT(start) ); - else - RENDER_LINE( ELT(start), ELT(count-1) ); - } - } - - POSTFIX; -} - - -static void TAG(render_triangles)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint j; - LOCAL_VARS; - (void) flags; - - INIT(GL_TRIANGLES); - if (NEED_EDGEFLAG_SETUP) { - for (j=start+2; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_TRI( ELT(j-2), ELT(j-1), ELT(j) ); - else - RENDER_TRI( ELT(j-1), ELT(j), ELT(j-2) ); - } - } else { - for (j=start+2; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_TRI( ELT(j-2), ELT(j-1), ELT(j) ); - else - RENDER_TRI( ELT(j-1), ELT(j), ELT(j-2) ); - } - } - POSTFIX; -} - - - -static void TAG(render_tri_strip)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint j; - GLuint parity = 0; - LOCAL_VARS; - - INIT(GL_TRIANGLE_STRIP); - if (NEED_EDGEFLAG_SETUP) { - for (j=start+2;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) { - ej2 = ELT(j-2+parity); - ej1 = ELT(j-1-parity); - ej = ELT(j); - } - else { - ej2 = ELT(j-1+parity); - ej1 = ELT(j-parity); - ej = ELT(j-2); - } - ef2 = EDGEFLAG_GET( ej2 ); - ef1 = EDGEFLAG_GET( ej1 ); - ef = EDGEFLAG_GET( ej ); - if (TEST_PRIM_BEGIN(flags)) { - RESET_STIPPLE; - } - EDGEFLAG_SET( ej2, GL_TRUE ); - EDGEFLAG_SET( ej1, GL_TRUE ); - EDGEFLAG_SET( ej, GL_TRUE ); - RENDER_TRI( ej2, ej1, ej ); - EDGEFLAG_SET( ej2, ef2 ); - EDGEFLAG_SET( ej1, ef1 ); - EDGEFLAG_SET( ej, ef ); - } - } else { - for (j=start+2; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_TRI( ELT(j-2+parity), ELT(j-1-parity), ELT(j) ); - else - RENDER_TRI( ELT(j-1+parity), ELT(j-parity), ELT(j-2) ); - } - } - POSTFIX; -} - - -static void TAG(render_tri_fan)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint j; - LOCAL_VARS; - (void) flags; - - INIT(GL_TRIANGLE_FAN); - if (NEED_EDGEFLAG_SETUP) { - for (j=start+2;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_TRI( ejs, ej1, ej); - else - RENDER_TRI( ej, ejs, ej1); - EDGEFLAG_SET( ejs, efs ); - EDGEFLAG_SET( ej1, ef1 ); - EDGEFLAG_SET( ej, ef ); - } - } else { - for (j=start+2;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) - RENDER_TRI( ELT(start), ELT(j-1), ELT(j) ); - else - RENDER_TRI( ELT(j), ELT(start), ELT(j-1) ); - } - } - - POSTFIX; -} - - -static void TAG(render_poly)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint j = start+2; - LOCAL_VARS; - (void) flags; - - INIT(GL_POLYGON); - if (NEED_EDGEFLAG_SETUP) { - GLboolean efstart = EDGEFLAG_GET( ELT(start) ); - GLboolean efcount = EDGEFLAG_GET( ELT(count-1) ); - - /* If the primitive does not begin here, the first edge - * is non-boundary. - */ - if (!TEST_PRIM_BEGIN(flags)) - EDGEFLAG_SET( ELT(start), GL_FALSE ); - else { - RESET_STIPPLE; - } - - /* If the primitive does not end here, the final edge is - * non-boundary. - */ - if (!TEST_PRIM_END(flags)) - EDGEFLAG_SET( ELT(count-1), GL_FALSE ); - - /* Draw the first triangles (possibly zero) - */ - if (j+1Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || - !ctx->Const.QuadsFollowProvokingVertexConvention) - RENDER_QUAD( ELT(j-3), ELT(j-2), ELT(j-1), ELT(j) ); - else - RENDER_QUAD( ELT(j-2), ELT(j-1), ELT(j), ELT(j-3) ); - } - } else { - for (j=start+3; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || - !ctx->Const.QuadsFollowProvokingVertexConvention) - RENDER_QUAD( ELT(j-3), ELT(j-2), ELT(j-1), ELT(j) ); - else - RENDER_QUAD( ELT(j-2), ELT(j-1), ELT(j), ELT(j-3) ); - } - } - POSTFIX; -} - -static void TAG(render_quad_strip)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - GLuint j; - LOCAL_VARS; - (void) flags; - - INIT(GL_QUAD_STRIP); - if (NEED_EDGEFLAG_SETUP) { - for (j=start+3;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || - !ctx->Const.QuadsFollowProvokingVertexConvention) - RENDER_QUAD( ELT(j-1), ELT(j-3), ELT(j-2), ELT(j) ); - else - RENDER_QUAD( ELT(j-2), ELT(j), ELT(j-1), ELT(j-3) ); - EDGEFLAG_SET( ELT(j-3), ef3 ); - EDGEFLAG_SET( ELT(j-2), ef2 ); - EDGEFLAG_SET( ELT(j-1), ef1 ); - EDGEFLAG_SET( ELT(j), ef ); - } - } else { - for (j=start+3;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || - !ctx->Const.QuadsFollowProvokingVertexConvention) - RENDER_QUAD( ELT(j-1), ELT(j-3), ELT(j-2), ELT(j) ); - else - RENDER_QUAD( ELT(j-2), ELT(j), ELT(j-1), ELT(j-3) ); - } - } - POSTFIX; -} - -static void TAG(render_noop)( struct gl_context *ctx, - GLuint start, - GLuint count, - GLuint flags ) -{ - (void)(ctx && start && count && flags); -} - -RENDER_TAB_QUALIFIER void (*TAG(render_tab)[GL_POLYGON+2])(struct gl_context *, - GLuint, - GLuint, - GLuint) = -{ - TAG(render_points), - TAG(render_lines), - TAG(render_line_loop), - TAG(render_line_strip), - TAG(render_triangles), - TAG(render_tri_strip), - TAG(render_tri_fan), - TAG(render_quads), - TAG(render_quad_strip), - TAG(render_poly), - TAG(render_noop), -}; - - - -#ifndef PRESERVE_VB_DEFS -#undef RENDER_TRI -#undef RENDER_QUAD -#undef RENDER_LINE -#undef RENDER_POINTS -#undef LOCAL_VARS -#undef INIT -#undef POSTFIX -#undef RESET_STIPPLE -#undef DBG -#undef ELT -#undef RENDER_TAB_QUALIFIER -#endif - -#ifndef PRESERVE_TAG -#undef TAG -#endif - -#undef PRESERVE_VB_DEFS -#undef PRESERVE_TAG +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2005 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + +#ifndef POSTFIX +#define POSTFIX +#endif + +#ifndef INIT +#define INIT(x) +#endif + +#ifndef NEED_EDGEFLAG_SETUP +#define NEED_EDGEFLAG_SETUP 0 +#define EDGEFLAG_GET(a) 0 +#define EDGEFLAG_SET(a,b) (void)b +#endif + +#ifndef RESET_STIPPLE +#define RESET_STIPPLE +#endif + +#ifndef TEST_PRIM_END +#define TEST_PRIM_END(prim) (flags & PRIM_END) +#define TEST_PRIM_BEGIN(prim) (flags & PRIM_BEGIN) +#endif + +#ifndef ELT +#define ELT(x) x +#endif + +#ifndef RENDER_TAB_QUALIFIER +#define RENDER_TAB_QUALIFIER static +#endif + +static void TAG(render_points)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + LOCAL_VARS; + (void) flags; + + INIT(GL_POINTS); + RENDER_POINTS( start, count ); + POSTFIX; +} + +static void TAG(render_lines)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint j; + LOCAL_VARS; + (void) flags; + + INIT(GL_LINES); + for (j=start+1; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_LINE( ELT(j-1), ELT(j) ); + else + RENDER_LINE( ELT(j), ELT(j-1) ); + } + POSTFIX; +} + + +static void TAG(render_line_strip)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint j; + LOCAL_VARS; + (void) flags; + + INIT(GL_LINE_STRIP); + + if (TEST_PRIM_BEGIN(flags)) { + RESET_STIPPLE; + } + + for (j=start+1; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_LINE( ELT(j-1), ELT(j) ); + else + RENDER_LINE( ELT(j), ELT(j-1) ); + } + POSTFIX; +} + + +static void TAG(render_line_loop)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint i; + LOCAL_VARS; + + (void) flags; + + INIT(GL_LINE_LOOP); + + if (start+1 < count) { + if (TEST_PRIM_BEGIN(flags)) { + RESET_STIPPLE; + if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_LINE( ELT(start), ELT(start+1) ); + else + RENDER_LINE( ELT(start+1), ELT(start) ); + } + + for ( i = start+2 ; i < count ; i++) { + if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_LINE( ELT(i-1), ELT(i) ); + else + RENDER_LINE( ELT(i), ELT(i-1) ); + } + + if ( TEST_PRIM_END(flags)) { + if (ctx->Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_LINE( ELT(count-1), ELT(start) ); + else + RENDER_LINE( ELT(start), ELT(count-1) ); + } + } + + POSTFIX; +} + + +static void TAG(render_triangles)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint j; + LOCAL_VARS; + (void) flags; + + INIT(GL_TRIANGLES); + if (NEED_EDGEFLAG_SETUP) { + for (j=start+2; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_TRI( ELT(j-2), ELT(j-1), ELT(j) ); + else + RENDER_TRI( ELT(j-1), ELT(j), ELT(j-2) ); + } + } else { + for (j=start+2; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_TRI( ELT(j-2), ELT(j-1), ELT(j) ); + else + RENDER_TRI( ELT(j-1), ELT(j), ELT(j-2) ); + } + } + POSTFIX; +} + + + +static void TAG(render_tri_strip)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint j; + GLuint parity = 0; + LOCAL_VARS; + + INIT(GL_TRIANGLE_STRIP); + if (NEED_EDGEFLAG_SETUP) { + for (j=start+2;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) { + ej2 = ELT(j-2+parity); + ej1 = ELT(j-1-parity); + ej = ELT(j); + } + else { + ej2 = ELT(j-1+parity); + ej1 = ELT(j-parity); + ej = ELT(j-2); + } + ef2 = EDGEFLAG_GET( ej2 ); + ef1 = EDGEFLAG_GET( ej1 ); + ef = EDGEFLAG_GET( ej ); + if (TEST_PRIM_BEGIN(flags)) { + RESET_STIPPLE; + } + EDGEFLAG_SET( ej2, GL_TRUE ); + EDGEFLAG_SET( ej1, GL_TRUE ); + EDGEFLAG_SET( ej, GL_TRUE ); + RENDER_TRI( ej2, ej1, ej ); + EDGEFLAG_SET( ej2, ef2 ); + EDGEFLAG_SET( ej1, ef1 ); + EDGEFLAG_SET( ej, ef ); + } + } else { + for (j=start+2; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_TRI( ELT(j-2+parity), ELT(j-1-parity), ELT(j) ); + else + RENDER_TRI( ELT(j-1+parity), ELT(j-parity), ELT(j-2) ); + } + } + POSTFIX; +} + + +static void TAG(render_tri_fan)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint j; + LOCAL_VARS; + (void) flags; + + INIT(GL_TRIANGLE_FAN); + if (NEED_EDGEFLAG_SETUP) { + for (j=start+2;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_TRI( ejs, ej1, ej); + else + RENDER_TRI( ej, ejs, ej1); + EDGEFLAG_SET( ejs, efs ); + EDGEFLAG_SET( ej1, ef1 ); + EDGEFLAG_SET( ej, ef ); + } + } else { + for (j=start+2;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT) + RENDER_TRI( ELT(start), ELT(j-1), ELT(j) ); + else + RENDER_TRI( ELT(j), ELT(start), ELT(j-1) ); + } + } + + POSTFIX; +} + + +static void TAG(render_poly)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint j = start+2; + LOCAL_VARS; + (void) flags; + + INIT(GL_POLYGON); + if (NEED_EDGEFLAG_SETUP) { + GLboolean efstart = EDGEFLAG_GET( ELT(start) ); + GLboolean efcount = EDGEFLAG_GET( ELT(count-1) ); + + /* If the primitive does not begin here, the first edge + * is non-boundary. + */ + if (!TEST_PRIM_BEGIN(flags)) + EDGEFLAG_SET( ELT(start), GL_FALSE ); + else { + RESET_STIPPLE; + } + + /* If the primitive does not end here, the final edge is + * non-boundary. + */ + if (!TEST_PRIM_END(flags)) + EDGEFLAG_SET( ELT(count-1), GL_FALSE ); + + /* Draw the first triangles (possibly zero) + */ + if (j+1Light.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || + !ctx->Const.QuadsFollowProvokingVertexConvention) + RENDER_QUAD( ELT(j-3), ELT(j-2), ELT(j-1), ELT(j) ); + else + RENDER_QUAD( ELT(j-2), ELT(j-1), ELT(j), ELT(j-3) ); + } + } else { + for (j=start+3; jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || + !ctx->Const.QuadsFollowProvokingVertexConvention) + RENDER_QUAD( ELT(j-3), ELT(j-2), ELT(j-1), ELT(j) ); + else + RENDER_QUAD( ELT(j-2), ELT(j-1), ELT(j), ELT(j-3) ); + } + } + POSTFIX; +} + +static void TAG(render_quad_strip)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + GLuint j; + LOCAL_VARS; + (void) flags; + + INIT(GL_QUAD_STRIP); + if (NEED_EDGEFLAG_SETUP) { + for (j=start+3;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || + !ctx->Const.QuadsFollowProvokingVertexConvention) + RENDER_QUAD( ELT(j-1), ELT(j-3), ELT(j-2), ELT(j) ); + else + RENDER_QUAD( ELT(j-2), ELT(j), ELT(j-1), ELT(j-3) ); + EDGEFLAG_SET( ELT(j-3), ef3 ); + EDGEFLAG_SET( ELT(j-2), ef2 ); + EDGEFLAG_SET( ELT(j-1), ef1 ); + EDGEFLAG_SET( ELT(j), ef ); + } + } else { + for (j=start+3;jLight.ProvokingVertex == GL_LAST_VERTEX_CONVENTION_EXT || + !ctx->Const.QuadsFollowProvokingVertexConvention) + RENDER_QUAD( ELT(j-1), ELT(j-3), ELT(j-2), ELT(j) ); + else + RENDER_QUAD( ELT(j-2), ELT(j), ELT(j-1), ELT(j-3) ); + } + } + POSTFIX; +} + +static void TAG(render_noop)( struct gl_context *ctx, + GLuint start, + GLuint count, + GLuint flags ) +{ + (void)(ctx && start && count && flags); +} + +RENDER_TAB_QUALIFIER void (*TAG(render_tab)[GL_POLYGON+2])(struct gl_context *, + GLuint, + GLuint, + GLuint) = +{ + TAG(render_points), + TAG(render_lines), + TAG(render_line_loop), + TAG(render_line_strip), + TAG(render_triangles), + TAG(render_tri_strip), + TAG(render_tri_fan), + TAG(render_quads), + TAG(render_quad_strip), + TAG(render_poly), + TAG(render_noop), +}; + + + +#ifndef PRESERVE_VB_DEFS +#undef RENDER_TRI +#undef RENDER_QUAD +#undef RENDER_LINE +#undef RENDER_POINTS +#undef LOCAL_VARS +#undef INIT +#undef POSTFIX +#undef RESET_STIPPLE +#undef DBG +#undef ELT +#undef RENDER_TAB_QUALIFIER +#endif + +#ifndef PRESERVE_TAG +#undef TAG +#endif + +#undef PRESERVE_VB_DEFS +#undef PRESERVE_TAG diff --git a/mesalib/src/mesa/tnl/t_vb_texgen.c b/mesalib/src/mesa/tnl/t_vb_texgen.c index 228808461..61430c396 100644 --- a/mesalib/src/mesa/tnl/t_vb_texgen.c +++ b/mesalib/src/mesa/tnl/t_vb_texgen.c @@ -1,609 +1,609 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Brian Paul - * Keith Whitwell - */ - -/* - * Regarding GL_NV_texgen_reflection: - * - * Portions of this software may use or implement intellectual - * property owned and licensed by NVIDIA Corporation. NVIDIA disclaims - * any and all warranties with respect to such intellectual property, - * including any use thereof or modifications thereto. - */ - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/macros.h" -#include "main/imports.h" -#include "main/mtypes.h" - -#include "math/m_xform.h" - -#include "t_context.h" -#include "t_pipeline.h" - - -/*********************************************************************** - * Automatic texture coordinate generation (texgen) code. - */ - - -struct texgen_stage_data; - -typedef void (*texgen_func)( struct gl_context *ctx, - struct texgen_stage_data *store, - GLuint unit); - - -struct texgen_stage_data { - - /* Per-texunit derived state. - */ - GLuint TexgenSize[MAX_TEXTURE_COORD_UNITS]; - texgen_func TexgenFunc[MAX_TEXTURE_COORD_UNITS]; - - /* Temporary values used in texgen. - */ - GLfloat (*tmp_f)[3]; - GLfloat *tmp_m; - - /* Buffered outputs of the stage. - */ - GLvector4f texcoord[MAX_TEXTURE_COORD_UNITS]; -}; - - -#define TEXGEN_STAGE_DATA(stage) ((struct texgen_stage_data *)stage->privatePtr) - - - -static GLuint all_bits[5] = { - 0, - VEC_SIZE_1, - VEC_SIZE_2, - VEC_SIZE_3, - VEC_SIZE_4, -}; - -#define VEC_SIZE_FLAGS (VEC_SIZE_1|VEC_SIZE_2|VEC_SIZE_3|VEC_SIZE_4) - -#define TEXGEN_NEED_M (TEXGEN_SPHERE_MAP) -#define TEXGEN_NEED_F (TEXGEN_SPHERE_MAP | \ - TEXGEN_REFLECTION_MAP_NV) - - - -static void build_m3( GLfloat f[][3], GLfloat m[], - const GLvector4f *normal, - const GLvector4f *eye ) -{ - GLuint stride = eye->stride; - GLfloat *coord = (GLfloat *)eye->start; - GLuint count = eye->count; - const GLfloat *norm = normal->start; - GLuint i; - - for (i=0;istride)) { - GLfloat u[3], two_nu, fx, fy, fz; - COPY_3V( u, coord ); - NORMALIZE_3FV( u ); - two_nu = 2.0F * DOT3(norm,u); - fx = f[i][0] = u[0] - norm[0] * two_nu; - fy = f[i][1] = u[1] - norm[1] * two_nu; - fz = f[i][2] = u[2] - norm[2] * two_nu; - m[i] = fx * fx + fy * fy + (fz + 1.0F) * (fz + 1.0F); - if (m[i] != 0.0F) { - m[i] = 0.5F * _mesa_inv_sqrtf(m[i]); - } - } -} - - - -static void build_m2( GLfloat f[][3], GLfloat m[], - const GLvector4f *normal, - const GLvector4f *eye ) -{ - GLuint stride = eye->stride; - GLfloat *coord = eye->start; - GLuint count = eye->count; - - GLfloat *norm = normal->start; - GLuint i; - - for (i=0;istride)) { - GLfloat u[3], two_nu, fx, fy, fz; - COPY_2V( u, coord ); - u[2] = 0; - NORMALIZE_3FV( u ); - two_nu = 2.0F * DOT3(norm,u); - fx = f[i][0] = u[0] - norm[0] * two_nu; - fy = f[i][1] = u[1] - norm[1] * two_nu; - fz = f[i][2] = u[2] - norm[2] * two_nu; - m[i] = fx * fx + fy * fy + (fz + 1.0F) * (fz + 1.0F); - if (m[i] != 0.0F) { - m[i] = 0.5F * _mesa_inv_sqrtf(m[i]); - } - } -} - - - -typedef void (*build_m_func)( GLfloat f[][3], - GLfloat m[], - const GLvector4f *normal, - const GLvector4f *eye ); - - -static build_m_func build_m_tab[5] = { - NULL, - NULL, - build_m2, - build_m3, - build_m3 -}; - - -/* This is unusual in that we respect the stride of the output vector - * (f). This allows us to pass in either a texcoord vector4f, or a - * temporary vector3f. - */ -static void build_f3( GLfloat *f, - GLuint fstride, - const GLvector4f *normal, - const GLvector4f *eye ) -{ - GLuint stride = eye->stride; - GLfloat *coord = eye->start; - GLuint count = eye->count; - - GLfloat *norm = normal->start; - GLuint i; - - for (i=0;istride); - } -} - - -static void build_f2( GLfloat *f, - GLuint fstride, - const GLvector4f *normal, - const GLvector4f *eye ) -{ - GLuint stride = eye->stride; - GLfloat *coord = eye->start; - GLuint count = eye->count; - GLfloat *norm = normal->start; - GLuint i; - - for (i=0;istride); - } -} - -typedef void (*build_f_func)( GLfloat *f, - GLuint fstride, - const GLvector4f *normal_vec, - const GLvector4f *eye ); - - - -/* Just treat 4-vectors as 3-vectors. - */ -static build_f_func build_f_tab[5] = { - NULL, - NULL, - build_f2, - build_f3, - build_f3 -}; - - - -/* Special case texgen functions. - */ -static void texgen_reflection_map_nv( struct gl_context *ctx, - struct texgen_stage_data *store, - GLuint unit ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; - GLvector4f *out = &store->texcoord[unit]; - - build_f_tab[VB->EyePtr->size]( out->start, - out->stride, - VB->AttribPtr[_TNL_ATTRIB_NORMAL], - VB->EyePtr ); - - out->flags |= (in->flags & VEC_SIZE_FLAGS) | VEC_SIZE_3; - out->count = VB->Count; - out->size = MAX2(in->size, 3); - if (in->size == 4) - _mesa_copy_tab[0x8]( out, in ); -} - - - -static void texgen_normal_map_nv( struct gl_context *ctx, - struct texgen_stage_data *store, - GLuint unit ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; - GLvector4f *out = &store->texcoord[unit]; - GLvector4f *normal = VB->AttribPtr[_TNL_ATTRIB_NORMAL]; - GLfloat (*texcoord)[4] = (GLfloat (*)[4])out->start; - GLuint count = VB->Count; - GLuint i; - const GLfloat *norm = normal->start; - - for (i=0;istride)) { - texcoord[i][0] = norm[0]; - texcoord[i][1] = norm[1]; - texcoord[i][2] = norm[2]; - } - - - out->flags |= (in->flags & VEC_SIZE_FLAGS) | VEC_SIZE_3; - out->count = count; - out->size = MAX2(in->size, 3); - if (in->size == 4) - _mesa_copy_tab[0x8]( out, in ); -} - - -static void texgen_sphere_map( struct gl_context *ctx, - struct texgen_stage_data *store, - GLuint unit ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; - GLvector4f *out = &store->texcoord[unit]; - GLfloat (*texcoord)[4] = (GLfloat (*)[4]) out->start; - GLuint count = VB->Count; - GLuint i; - GLfloat (*f)[3] = store->tmp_f; - GLfloat *m = store->tmp_m; - - (build_m_tab[VB->EyePtr->size])( store->tmp_f, - store->tmp_m, - VB->AttribPtr[_TNL_ATTRIB_NORMAL], - VB->EyePtr ); - - out->size = MAX2(in->size,2); - - for (i=0;icount = count; - out->flags |= (in->flags & VEC_SIZE_FLAGS) | VEC_SIZE_2; - if (in->size > 2) - _mesa_copy_tab[all_bits[in->size] & ~0x3]( out, in ); -} - - - -static void texgen( struct gl_context *ctx, - struct texgen_stage_data *store, - GLuint unit ) -{ - TNLcontext *tnl = TNL_CONTEXT(ctx); - struct vertex_buffer *VB = &tnl->vb; - GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; - GLvector4f *out = &store->texcoord[unit]; - const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit]; - const GLvector4f *obj = VB->AttribPtr[_TNL_ATTRIB_POS]; - const GLvector4f *eye = VB->EyePtr; - const GLvector4f *normal = VB->AttribPtr[_TNL_ATTRIB_NORMAL]; - const GLfloat *m = store->tmp_m; - const GLuint count = VB->Count; - GLfloat (*texcoord)[4] = (GLfloat (*)[4])out->data; - GLfloat (*f)[3] = store->tmp_f; - GLuint copy; - - if (texUnit->_GenFlags & TEXGEN_NEED_M) { - build_m_tab[eye->size]( store->tmp_f, store->tmp_m, normal, eye ); - } else if (texUnit->_GenFlags & TEXGEN_NEED_F) { - build_f_tab[eye->size]( (GLfloat *)store->tmp_f, 3, normal, eye ); - } - - - out->size = MAX2(in->size, store->TexgenSize[unit]); - out->flags |= (in->flags & VEC_SIZE_FLAGS) | texUnit->TexGenEnabled; - out->count = count; - - copy = (all_bits[in->size] & ~texUnit->TexGenEnabled); - if (copy) - _mesa_copy_tab[copy]( out, in ); - - if (texUnit->TexGenEnabled & S_BIT) { - GLuint i; - switch (texUnit->GenS.Mode) { - case GL_OBJECT_LINEAR: - _mesa_dotprod_tab[obj->size]( (GLfloat *)out->data, - sizeof(out->data[0]), obj, - texUnit->GenS.ObjectPlane ); - break; - case GL_EYE_LINEAR: - _mesa_dotprod_tab[eye->size]( (GLfloat *)out->data, - sizeof(out->data[0]), eye, - texUnit->GenS.EyePlane ); - break; - case GL_SPHERE_MAP: - for (i = 0; i < count; i++) - texcoord[i][0] = f[i][0] * m[i] + 0.5F; - break; - case GL_REFLECTION_MAP_NV: - for (i=0;istart; - for (i=0;istride)) { - texcoord[i][0] = norm[0]; - } - break; - } - default: - _mesa_problem(ctx, "Bad S texgen"); - } - } - - if (texUnit->TexGenEnabled & T_BIT) { - GLuint i; - switch (texUnit->GenT.Mode) { - case GL_OBJECT_LINEAR: - _mesa_dotprod_tab[obj->size]( &(out->data[0][1]), - sizeof(out->data[0]), obj, - texUnit->GenT.ObjectPlane ); - break; - case GL_EYE_LINEAR: - _mesa_dotprod_tab[eye->size]( &(out->data[0][1]), - sizeof(out->data[0]), eye, - texUnit->GenT.EyePlane ); - break; - case GL_SPHERE_MAP: - for (i = 0; i < count; i++) - texcoord[i][1] = f[i][1] * m[i] + 0.5F; - break; - case GL_REFLECTION_MAP_NV: - for (i=0;istart; - for (i=0;istride)) { - texcoord[i][1] = norm[1]; - } - break; - } - default: - _mesa_problem(ctx, "Bad T texgen"); - } - } - - if (texUnit->TexGenEnabled & R_BIT) { - GLuint i; - switch (texUnit->GenR.Mode) { - case GL_OBJECT_LINEAR: - _mesa_dotprod_tab[obj->size]( &(out->data[0][2]), - sizeof(out->data[0]), obj, - texUnit->GenR.ObjectPlane ); - break; - case GL_EYE_LINEAR: - _mesa_dotprod_tab[eye->size]( &(out->data[0][2]), - sizeof(out->data[0]), eye, - texUnit->GenR.EyePlane ); - break; - case GL_REFLECTION_MAP_NV: - for (i=0;istart; - for (i=0;istride)) { - texcoord[i][2] = norm[2]; - } - break; - } - default: - _mesa_problem(ctx, "Bad R texgen"); - } - } - - if (texUnit->TexGenEnabled & Q_BIT) { - switch (texUnit->GenQ.Mode) { - case GL_OBJECT_LINEAR: - _mesa_dotprod_tab[obj->size]( &(out->data[0][3]), - sizeof(out->data[0]), obj, - texUnit->GenQ.ObjectPlane ); - break; - case GL_EYE_LINEAR: - _mesa_dotprod_tab[eye->size]( &(out->data[0][3]), - sizeof(out->data[0]), eye, - texUnit->GenQ.EyePlane ); - break; - default: - _mesa_problem(ctx, "Bad Q texgen"); - } - } -} - - - - -static GLboolean run_texgen_stage( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - struct texgen_stage_data *store = TEXGEN_STAGE_DATA(stage); - GLuint i; - - if (!ctx->Texture._TexGenEnabled || ctx->VertexProgram._Current) - return GL_TRUE; - - for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) { - struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i]; - - if (texUnit->TexGenEnabled) { - - store->TexgenFunc[i]( ctx, store, i ); - - VB->AttribPtr[VERT_ATTRIB_TEX0 + i] = &store->texcoord[i]; - } - } - - return GL_TRUE; -} - - -static void validate_texgen_stage( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - struct texgen_stage_data *store = TEXGEN_STAGE_DATA(stage); - GLuint i; - - if (!ctx->Texture._TexGenEnabled || ctx->VertexProgram._Current) - return; - - for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) { - struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i]; - - if (texUnit->TexGenEnabled) { - GLuint sz; - - if (texUnit->TexGenEnabled & Q_BIT) - sz = 4; - else if (texUnit->TexGenEnabled & R_BIT) - sz = 3; - else if (texUnit->TexGenEnabled & T_BIT) - sz = 2; - else - sz = 1; - - store->TexgenSize[i] = sz; - store->TexgenFunc[i] = texgen; /* general solution */ - - /* look for special texgen cases */ - if (texUnit->TexGenEnabled == (S_BIT|T_BIT|R_BIT)) { - if (texUnit->_GenFlags == TEXGEN_REFLECTION_MAP_NV) { - store->TexgenFunc[i] = texgen_reflection_map_nv; - } - else if (texUnit->_GenFlags == TEXGEN_NORMAL_MAP_NV) { - store->TexgenFunc[i] = texgen_normal_map_nv; - } - } - else if (texUnit->TexGenEnabled == (S_BIT|T_BIT) && - texUnit->_GenFlags == TEXGEN_SPHERE_MAP) { - store->TexgenFunc[i] = texgen_sphere_map; - } - } - } -} - - - - - -/* Called the first time stage->run() is invoked. - */ -static GLboolean alloc_texgen_data( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - struct texgen_stage_data *store; - GLuint i; - - stage->privatePtr = CALLOC(sizeof(*store)); - store = TEXGEN_STAGE_DATA(stage); - if (!store) - return GL_FALSE; - - for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) - _mesa_vector4f_alloc( &store->texcoord[i], 0, VB->Size, 32 ); - - store->tmp_f = (GLfloat (*)[3]) MALLOC(VB->Size * sizeof(GLfloat) * 3); - store->tmp_m = (GLfloat *) MALLOC(VB->Size * sizeof(GLfloat)); - - return GL_TRUE; -} - - -static void free_texgen_data( struct tnl_pipeline_stage *stage ) - -{ - struct texgen_stage_data *store = TEXGEN_STAGE_DATA(stage); - GLuint i; - - if (store) { - for (i = 0 ; i < MAX_TEXTURE_COORD_UNITS ; i++) - if (store->texcoord[i].data) - _mesa_vector4f_free( &store->texcoord[i] ); - - - if (store->tmp_f) FREE( store->tmp_f ); - if (store->tmp_m) FREE( store->tmp_m ); - FREE( store ); - stage->privatePtr = NULL; - } -} - - - -const struct tnl_pipeline_stage _tnl_texgen_stage = -{ - "texgen", /* name */ - NULL, /* private data */ - alloc_texgen_data, /* destructor */ - free_texgen_data, /* destructor */ - validate_texgen_stage, /* check */ - run_texgen_stage /* run -- initially set to alloc data */ -}; +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Brian Paul + * Keith Whitwell + */ + +/* + * Regarding GL_NV_texgen_reflection: + * + * Portions of this software may use or implement intellectual + * property owned and licensed by NVIDIA Corporation. NVIDIA disclaims + * any and all warranties with respect to such intellectual property, + * including any use thereof or modifications thereto. + */ + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/macros.h" +#include "main/imports.h" +#include "main/mtypes.h" + +#include "math/m_xform.h" + +#include "t_context.h" +#include "t_pipeline.h" + + +/*********************************************************************** + * Automatic texture coordinate generation (texgen) code. + */ + + +struct texgen_stage_data; + +typedef void (*texgen_func)( struct gl_context *ctx, + struct texgen_stage_data *store, + GLuint unit); + + +struct texgen_stage_data { + + /* Per-texunit derived state. + */ + GLuint TexgenSize[MAX_TEXTURE_COORD_UNITS]; + texgen_func TexgenFunc[MAX_TEXTURE_COORD_UNITS]; + + /* Temporary values used in texgen. + */ + GLfloat (*tmp_f)[3]; + GLfloat *tmp_m; + + /* Buffered outputs of the stage. + */ + GLvector4f texcoord[MAX_TEXTURE_COORD_UNITS]; +}; + + +#define TEXGEN_STAGE_DATA(stage) ((struct texgen_stage_data *)stage->privatePtr) + + + +static GLuint all_bits[5] = { + 0, + VEC_SIZE_1, + VEC_SIZE_2, + VEC_SIZE_3, + VEC_SIZE_4, +}; + +#define VEC_SIZE_FLAGS (VEC_SIZE_1|VEC_SIZE_2|VEC_SIZE_3|VEC_SIZE_4) + +#define TEXGEN_NEED_M (TEXGEN_SPHERE_MAP) +#define TEXGEN_NEED_F (TEXGEN_SPHERE_MAP | \ + TEXGEN_REFLECTION_MAP_NV) + + + +static void build_m3( GLfloat f[][3], GLfloat m[], + const GLvector4f *normal, + const GLvector4f *eye ) +{ + GLuint stride = eye->stride; + GLfloat *coord = (GLfloat *)eye->start; + GLuint count = eye->count; + const GLfloat *norm = normal->start; + GLuint i; + + for (i=0;istride)) { + GLfloat u[3], two_nu, fx, fy, fz; + COPY_3V( u, coord ); + NORMALIZE_3FV( u ); + two_nu = 2.0F * DOT3(norm,u); + fx = f[i][0] = u[0] - norm[0] * two_nu; + fy = f[i][1] = u[1] - norm[1] * two_nu; + fz = f[i][2] = u[2] - norm[2] * two_nu; + m[i] = fx * fx + fy * fy + (fz + 1.0F) * (fz + 1.0F); + if (m[i] != 0.0F) { + m[i] = 0.5F * _mesa_inv_sqrtf(m[i]); + } + } +} + + + +static void build_m2( GLfloat f[][3], GLfloat m[], + const GLvector4f *normal, + const GLvector4f *eye ) +{ + GLuint stride = eye->stride; + GLfloat *coord = eye->start; + GLuint count = eye->count; + + GLfloat *norm = normal->start; + GLuint i; + + for (i=0;istride)) { + GLfloat u[3], two_nu, fx, fy, fz; + COPY_2V( u, coord ); + u[2] = 0; + NORMALIZE_3FV( u ); + two_nu = 2.0F * DOT3(norm,u); + fx = f[i][0] = u[0] - norm[0] * two_nu; + fy = f[i][1] = u[1] - norm[1] * two_nu; + fz = f[i][2] = u[2] - norm[2] * two_nu; + m[i] = fx * fx + fy * fy + (fz + 1.0F) * (fz + 1.0F); + if (m[i] != 0.0F) { + m[i] = 0.5F * _mesa_inv_sqrtf(m[i]); + } + } +} + + + +typedef void (*build_m_func)( GLfloat f[][3], + GLfloat m[], + const GLvector4f *normal, + const GLvector4f *eye ); + + +static build_m_func build_m_tab[5] = { + NULL, + NULL, + build_m2, + build_m3, + build_m3 +}; + + +/* This is unusual in that we respect the stride of the output vector + * (f). This allows us to pass in either a texcoord vector4f, or a + * temporary vector3f. + */ +static void build_f3( GLfloat *f, + GLuint fstride, + const GLvector4f *normal, + const GLvector4f *eye ) +{ + GLuint stride = eye->stride; + GLfloat *coord = eye->start; + GLuint count = eye->count; + + GLfloat *norm = normal->start; + GLuint i; + + for (i=0;istride); + } +} + + +static void build_f2( GLfloat *f, + GLuint fstride, + const GLvector4f *normal, + const GLvector4f *eye ) +{ + GLuint stride = eye->stride; + GLfloat *coord = eye->start; + GLuint count = eye->count; + GLfloat *norm = normal->start; + GLuint i; + + for (i=0;istride); + } +} + +typedef void (*build_f_func)( GLfloat *f, + GLuint fstride, + const GLvector4f *normal_vec, + const GLvector4f *eye ); + + + +/* Just treat 4-vectors as 3-vectors. + */ +static build_f_func build_f_tab[5] = { + NULL, + NULL, + build_f2, + build_f3, + build_f3 +}; + + + +/* Special case texgen functions. + */ +static void texgen_reflection_map_nv( struct gl_context *ctx, + struct texgen_stage_data *store, + GLuint unit ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; + GLvector4f *out = &store->texcoord[unit]; + + build_f_tab[VB->EyePtr->size]( out->start, + out->stride, + VB->AttribPtr[_TNL_ATTRIB_NORMAL], + VB->EyePtr ); + + out->flags |= (in->flags & VEC_SIZE_FLAGS) | VEC_SIZE_3; + out->count = VB->Count; + out->size = MAX2(in->size, 3); + if (in->size == 4) + _mesa_copy_tab[0x8]( out, in ); +} + + + +static void texgen_normal_map_nv( struct gl_context *ctx, + struct texgen_stage_data *store, + GLuint unit ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; + GLvector4f *out = &store->texcoord[unit]; + GLvector4f *normal = VB->AttribPtr[_TNL_ATTRIB_NORMAL]; + GLfloat (*texcoord)[4] = (GLfloat (*)[4])out->start; + GLuint count = VB->Count; + GLuint i; + const GLfloat *norm = normal->start; + + for (i=0;istride)) { + texcoord[i][0] = norm[0]; + texcoord[i][1] = norm[1]; + texcoord[i][2] = norm[2]; + } + + + out->flags |= (in->flags & VEC_SIZE_FLAGS) | VEC_SIZE_3; + out->count = count; + out->size = MAX2(in->size, 3); + if (in->size == 4) + _mesa_copy_tab[0x8]( out, in ); +} + + +static void texgen_sphere_map( struct gl_context *ctx, + struct texgen_stage_data *store, + GLuint unit ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; + GLvector4f *out = &store->texcoord[unit]; + GLfloat (*texcoord)[4] = (GLfloat (*)[4]) out->start; + GLuint count = VB->Count; + GLuint i; + GLfloat (*f)[3] = store->tmp_f; + GLfloat *m = store->tmp_m; + + (build_m_tab[VB->EyePtr->size])( store->tmp_f, + store->tmp_m, + VB->AttribPtr[_TNL_ATTRIB_NORMAL], + VB->EyePtr ); + + out->size = MAX2(in->size,2); + + for (i=0;icount = count; + out->flags |= (in->flags & VEC_SIZE_FLAGS) | VEC_SIZE_2; + if (in->size > 2) + _mesa_copy_tab[all_bits[in->size] & ~0x3]( out, in ); +} + + + +static void texgen( struct gl_context *ctx, + struct texgen_stage_data *store, + GLuint unit ) +{ + TNLcontext *tnl = TNL_CONTEXT(ctx); + struct vertex_buffer *VB = &tnl->vb; + GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit]; + GLvector4f *out = &store->texcoord[unit]; + const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit]; + const GLvector4f *obj = VB->AttribPtr[_TNL_ATTRIB_POS]; + const GLvector4f *eye = VB->EyePtr; + const GLvector4f *normal = VB->AttribPtr[_TNL_ATTRIB_NORMAL]; + const GLfloat *m = store->tmp_m; + const GLuint count = VB->Count; + GLfloat (*texcoord)[4] = (GLfloat (*)[4])out->data; + GLfloat (*f)[3] = store->tmp_f; + GLuint copy; + + if (texUnit->_GenFlags & TEXGEN_NEED_M) { + build_m_tab[eye->size]( store->tmp_f, store->tmp_m, normal, eye ); + } else if (texUnit->_GenFlags & TEXGEN_NEED_F) { + build_f_tab[eye->size]( (GLfloat *)store->tmp_f, 3, normal, eye ); + } + + + out->size = MAX2(in->size, store->TexgenSize[unit]); + out->flags |= (in->flags & VEC_SIZE_FLAGS) | texUnit->TexGenEnabled; + out->count = count; + + copy = (all_bits[in->size] & ~texUnit->TexGenEnabled); + if (copy) + _mesa_copy_tab[copy]( out, in ); + + if (texUnit->TexGenEnabled & S_BIT) { + GLuint i; + switch (texUnit->GenS.Mode) { + case GL_OBJECT_LINEAR: + _mesa_dotprod_tab[obj->size]( (GLfloat *)out->data, + sizeof(out->data[0]), obj, + texUnit->GenS.ObjectPlane ); + break; + case GL_EYE_LINEAR: + _mesa_dotprod_tab[eye->size]( (GLfloat *)out->data, + sizeof(out->data[0]), eye, + texUnit->GenS.EyePlane ); + break; + case GL_SPHERE_MAP: + for (i = 0; i < count; i++) + texcoord[i][0] = f[i][0] * m[i] + 0.5F; + break; + case GL_REFLECTION_MAP_NV: + for (i=0;istart; + for (i=0;istride)) { + texcoord[i][0] = norm[0]; + } + break; + } + default: + _mesa_problem(ctx, "Bad S texgen"); + } + } + + if (texUnit->TexGenEnabled & T_BIT) { + GLuint i; + switch (texUnit->GenT.Mode) { + case GL_OBJECT_LINEAR: + _mesa_dotprod_tab[obj->size]( &(out->data[0][1]), + sizeof(out->data[0]), obj, + texUnit->GenT.ObjectPlane ); + break; + case GL_EYE_LINEAR: + _mesa_dotprod_tab[eye->size]( &(out->data[0][1]), + sizeof(out->data[0]), eye, + texUnit->GenT.EyePlane ); + break; + case GL_SPHERE_MAP: + for (i = 0; i < count; i++) + texcoord[i][1] = f[i][1] * m[i] + 0.5F; + break; + case GL_REFLECTION_MAP_NV: + for (i=0;istart; + for (i=0;istride)) { + texcoord[i][1] = norm[1]; + } + break; + } + default: + _mesa_problem(ctx, "Bad T texgen"); + } + } + + if (texUnit->TexGenEnabled & R_BIT) { + GLuint i; + switch (texUnit->GenR.Mode) { + case GL_OBJECT_LINEAR: + _mesa_dotprod_tab[obj->size]( &(out->data[0][2]), + sizeof(out->data[0]), obj, + texUnit->GenR.ObjectPlane ); + break; + case GL_EYE_LINEAR: + _mesa_dotprod_tab[eye->size]( &(out->data[0][2]), + sizeof(out->data[0]), eye, + texUnit->GenR.EyePlane ); + break; + case GL_REFLECTION_MAP_NV: + for (i=0;istart; + for (i=0;istride)) { + texcoord[i][2] = norm[2]; + } + break; + } + default: + _mesa_problem(ctx, "Bad R texgen"); + } + } + + if (texUnit->TexGenEnabled & Q_BIT) { + switch (texUnit->GenQ.Mode) { + case GL_OBJECT_LINEAR: + _mesa_dotprod_tab[obj->size]( &(out->data[0][3]), + sizeof(out->data[0]), obj, + texUnit->GenQ.ObjectPlane ); + break; + case GL_EYE_LINEAR: + _mesa_dotprod_tab[eye->size]( &(out->data[0][3]), + sizeof(out->data[0]), eye, + texUnit->GenQ.EyePlane ); + break; + default: + _mesa_problem(ctx, "Bad Q texgen"); + } + } +} + + + + +static GLboolean run_texgen_stage( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + struct texgen_stage_data *store = TEXGEN_STAGE_DATA(stage); + GLuint i; + + if (!ctx->Texture._TexGenEnabled || ctx->VertexProgram._Current) + return GL_TRUE; + + for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) { + struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i]; + + if (texUnit->TexGenEnabled) { + + store->TexgenFunc[i]( ctx, store, i ); + + VB->AttribPtr[VERT_ATTRIB_TEX0 + i] = &store->texcoord[i]; + } + } + + return GL_TRUE; +} + + +static void validate_texgen_stage( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + struct texgen_stage_data *store = TEXGEN_STAGE_DATA(stage); + GLuint i; + + if (!ctx->Texture._TexGenEnabled || ctx->VertexProgram._Current) + return; + + for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) { + struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i]; + + if (texUnit->TexGenEnabled) { + GLuint sz; + + if (texUnit->TexGenEnabled & Q_BIT) + sz = 4; + else if (texUnit->TexGenEnabled & R_BIT) + sz = 3; + else if (texUnit->TexGenEnabled & T_BIT) + sz = 2; + else + sz = 1; + + store->TexgenSize[i] = sz; + store->TexgenFunc[i] = texgen; /* general solution */ + + /* look for special texgen cases */ + if (texUnit->TexGenEnabled == (S_BIT|T_BIT|R_BIT)) { + if (texUnit->_GenFlags == TEXGEN_REFLECTION_MAP_NV) { + store->TexgenFunc[i] = texgen_reflection_map_nv; + } + else if (texUnit->_GenFlags == TEXGEN_NORMAL_MAP_NV) { + store->TexgenFunc[i] = texgen_normal_map_nv; + } + } + else if (texUnit->TexGenEnabled == (S_BIT|T_BIT) && + texUnit->_GenFlags == TEXGEN_SPHERE_MAP) { + store->TexgenFunc[i] = texgen_sphere_map; + } + } + } +} + + + + + +/* Called the first time stage->run() is invoked. + */ +static GLboolean alloc_texgen_data( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + struct texgen_stage_data *store; + GLuint i; + + stage->privatePtr = CALLOC(sizeof(*store)); + store = TEXGEN_STAGE_DATA(stage); + if (!store) + return GL_FALSE; + + for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) + _mesa_vector4f_alloc( &store->texcoord[i], 0, VB->Size, 32 ); + + store->tmp_f = (GLfloat (*)[3]) MALLOC(VB->Size * sizeof(GLfloat) * 3); + store->tmp_m = (GLfloat *) MALLOC(VB->Size * sizeof(GLfloat)); + + return GL_TRUE; +} + + +static void free_texgen_data( struct tnl_pipeline_stage *stage ) + +{ + struct texgen_stage_data *store = TEXGEN_STAGE_DATA(stage); + GLuint i; + + if (store) { + for (i = 0 ; i < MAX_TEXTURE_COORD_UNITS ; i++) + if (store->texcoord[i].data) + _mesa_vector4f_free( &store->texcoord[i] ); + + + if (store->tmp_f) FREE( store->tmp_f ); + if (store->tmp_m) FREE( store->tmp_m ); + FREE( store ); + stage->privatePtr = NULL; + } +} + + + +const struct tnl_pipeline_stage _tnl_texgen_stage = +{ + "texgen", /* name */ + NULL, /* private data */ + alloc_texgen_data, /* destructor */ + free_texgen_data, /* destructor */ + validate_texgen_stage, /* check */ + run_texgen_stage /* run -- initially set to alloc data */ +}; diff --git a/mesalib/src/mesa/tnl/t_vb_texmat.c b/mesalib/src/mesa/tnl/t_vb_texmat.c index 7b30946c8..38aa51fc4 100644 --- a/mesalib/src/mesa/tnl/t_vb_texmat.c +++ b/mesalib/src/mesa/tnl/t_vb_texmat.c @@ -1,128 +1,128 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/macros.h" -#include "main/imports.h" -#include "main/mtypes.h" - -#include "math/m_xform.h" - -#include "t_context.h" -#include "t_pipeline.h" - -/* Is there any real benefit seperating texmat from texgen? It means - * we need two lots of intermediate storage. Any changes to - * _NEW_TEXTURE will invalidate both sets -- it's only on changes to - * *only* _NEW_TEXTURE_MATRIX that texgen survives but texmat doesn't. - * - * However, the seperation of this code from the complex texgen stuff - * is very appealing. - */ -struct texmat_stage_data { - GLvector4f texcoord[MAX_TEXTURE_COORD_UNITS]; -}; - -#define TEXMAT_STAGE_DATA(stage) ((struct texmat_stage_data *)stage->privatePtr) - - - -static GLboolean run_texmat_stage( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - struct texmat_stage_data *store = TEXMAT_STAGE_DATA(stage); - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - GLuint i; - - if (!ctx->Texture._TexMatEnabled || ctx->VertexProgram._Current) - return GL_TRUE; - - /* ENABLE_TEXMAT implies that the texture matrix is not the - * identity, so we don't have to check that here. - */ - for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) { - if (ctx->Texture._TexMatEnabled & ENABLE_TEXMAT(i)) { - (void) TransformRaw( &store->texcoord[i], - ctx->TextureMatrixStack[i].Top, - VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]); - - VB->AttribPtr[VERT_ATTRIB_TEX0+i] = &store->texcoord[i]; - } - } - - return GL_TRUE; -} - - -/* Called the first time stage->run() is invoked. - */ -static GLboolean alloc_texmat_data( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - struct texmat_stage_data *store; - GLuint i; - - stage->privatePtr = CALLOC(sizeof(*store)); - store = TEXMAT_STAGE_DATA(stage); - if (!store) - return GL_FALSE; - - for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) - _mesa_vector4f_alloc( &store->texcoord[i], 0, VB->Size, 32 ); - - return GL_TRUE; -} - - -static void free_texmat_data( struct tnl_pipeline_stage *stage ) -{ - struct texmat_stage_data *store = TEXMAT_STAGE_DATA(stage); - GLuint i; - - if (store) { - for (i = 0; i < MAX_TEXTURE_COORD_UNITS; i++) - if (store->texcoord[i].data) - _mesa_vector4f_free( &store->texcoord[i] ); - FREE( store ); - stage->privatePtr = NULL; - } -} - - - -const struct tnl_pipeline_stage _tnl_texture_transform_stage = -{ - "texture transform", /* name */ - NULL, /* private data */ - alloc_texmat_data, - free_texmat_data, /* destructor */ - NULL, - run_texmat_stage, -}; +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/macros.h" +#include "main/imports.h" +#include "main/mtypes.h" + +#include "math/m_xform.h" + +#include "t_context.h" +#include "t_pipeline.h" + +/* Is there any real benefit seperating texmat from texgen? It means + * we need two lots of intermediate storage. Any changes to + * _NEW_TEXTURE will invalidate both sets -- it's only on changes to + * *only* _NEW_TEXTURE_MATRIX that texgen survives but texmat doesn't. + * + * However, the seperation of this code from the complex texgen stuff + * is very appealing. + */ +struct texmat_stage_data { + GLvector4f texcoord[MAX_TEXTURE_COORD_UNITS]; +}; + +#define TEXMAT_STAGE_DATA(stage) ((struct texmat_stage_data *)stage->privatePtr) + + + +static GLboolean run_texmat_stage( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + struct texmat_stage_data *store = TEXMAT_STAGE_DATA(stage); + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + GLuint i; + + if (!ctx->Texture._TexMatEnabled || ctx->VertexProgram._Current) + return GL_TRUE; + + /* ENABLE_TEXMAT implies that the texture matrix is not the + * identity, so we don't have to check that here. + */ + for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) { + if (ctx->Texture._TexMatEnabled & ENABLE_TEXMAT(i)) { + (void) TransformRaw( &store->texcoord[i], + ctx->TextureMatrixStack[i].Top, + VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]); + + VB->AttribPtr[VERT_ATTRIB_TEX0+i] = &store->texcoord[i]; + } + } + + return GL_TRUE; +} + + +/* Called the first time stage->run() is invoked. + */ +static GLboolean alloc_texmat_data( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + struct texmat_stage_data *store; + GLuint i; + + stage->privatePtr = CALLOC(sizeof(*store)); + store = TEXMAT_STAGE_DATA(stage); + if (!store) + return GL_FALSE; + + for (i = 0 ; i < ctx->Const.MaxTextureCoordUnits ; i++) + _mesa_vector4f_alloc( &store->texcoord[i], 0, VB->Size, 32 ); + + return GL_TRUE; +} + + +static void free_texmat_data( struct tnl_pipeline_stage *stage ) +{ + struct texmat_stage_data *store = TEXMAT_STAGE_DATA(stage); + GLuint i; + + if (store) { + for (i = 0; i < MAX_TEXTURE_COORD_UNITS; i++) + if (store->texcoord[i].data) + _mesa_vector4f_free( &store->texcoord[i] ); + FREE( store ); + stage->privatePtr = NULL; + } +} + + + +const struct tnl_pipeline_stage _tnl_texture_transform_stage = +{ + "texture transform", /* name */ + NULL, /* private data */ + alloc_texmat_data, + free_texmat_data, /* destructor */ + NULL, + run_texmat_stage, +}; diff --git a/mesalib/src/mesa/tnl/t_vb_vertex.c b/mesalib/src/mesa/tnl/t_vb_vertex.c index a77959410..26e8ae065 100644 --- a/mesalib/src/mesa/tnl/t_vb_vertex.c +++ b/mesalib/src/mesa/tnl/t_vb_vertex.c @@ -1,283 +1,283 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - - -#include "main/glheader.h" -#include "main/colormac.h" -#include "main/macros.h" -#include "main/imports.h" -#include "main/mtypes.h" - -#include "math/m_xform.h" - -#include "t_context.h" -#include "t_pipeline.h" - - - -struct vertex_stage_data { - GLvector4f eye; - GLvector4f clip; - GLvector4f proj; - GLubyte *clipmask; - GLubyte ormask; - GLubyte andmask; -}; - -#define VERTEX_STAGE_DATA(stage) ((struct vertex_stage_data *)stage->privatePtr) - - - - -/* This function implements cliptesting for user-defined clip planes. - * The clipping of primitives to these planes is implemented in - * t_render_clip.h. - */ -#define USER_CLIPTEST(NAME, SZ) \ -static void NAME( struct gl_context *ctx, \ - GLvector4f *clip, \ - GLubyte *clipmask, \ - GLubyte *clipormask, \ - GLubyte *clipandmask ) \ -{ \ - GLuint p; \ - \ - for (p = 0; p < ctx->Const.MaxClipPlanes; p++) \ - if (ctx->Transform.ClipPlanesEnabled & (1 << p)) { \ - GLuint nr, i; \ - const GLfloat a = ctx->Transform._ClipUserPlane[p][0]; \ - const GLfloat b = ctx->Transform._ClipUserPlane[p][1]; \ - const GLfloat c = ctx->Transform._ClipUserPlane[p][2]; \ - const GLfloat d = ctx->Transform._ClipUserPlane[p][3]; \ - GLfloat *coord = (GLfloat *)clip->data; \ - GLuint stride = clip->stride; \ - GLuint count = clip->count; \ - \ - for (nr = 0, i = 0 ; i < count ; i++) { \ - GLfloat dp = coord[0] * a + coord[1] * b; \ - if (SZ > 2) dp += coord[2] * c; \ - if (SZ > 3) dp += coord[3] * d; else dp += d; \ - \ - if (dp < 0) { \ - nr++; \ - clipmask[i] |= CLIP_USER_BIT; \ - } \ - \ - STRIDE_F(coord, stride); \ - } \ - \ - if (nr > 0) { \ - *clipormask |= CLIP_USER_BIT; \ - if (nr == count) { \ - *clipandmask |= CLIP_USER_BIT; \ - return; \ - } \ - } \ - } \ -} - - -USER_CLIPTEST(userclip2, 2) -USER_CLIPTEST(userclip3, 3) -USER_CLIPTEST(userclip4, 4) - -static void (*(usercliptab[5]))( struct gl_context *, - GLvector4f *, GLubyte *, - GLubyte *, GLubyte * ) = -{ - NULL, - NULL, - userclip2, - userclip3, - userclip4 -}; - - -void -tnl_clip_prepare(struct gl_context *ctx) -{ - /* Neither the x86 nor sparc asm cliptest functions have been updated - * for ARB_depth_clamp, so force the C paths. - */ - if (ctx->Transform.DepthClamp) { - static GLboolean c_funcs_installed = GL_FALSE; - if (!c_funcs_installed) { - init_c_cliptest(); - c_funcs_installed = GL_TRUE; - } - } -} - - - -static GLboolean run_vertex_stage( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - struct vertex_stage_data *store = (struct vertex_stage_data *)stage->privatePtr; - TNLcontext *tnl = TNL_CONTEXT(ctx); - struct vertex_buffer *VB = &tnl->vb; - - if (ctx->VertexProgram._Current) - return GL_TRUE; - - tnl_clip_prepare(ctx); - - if (ctx->_NeedEyeCoords) { - /* Separate modelview transformation: - * Use combined ModelProject to avoid some depth artifacts - */ - if (ctx->ModelviewMatrixStack.Top->type == MATRIX_IDENTITY) - VB->EyePtr = VB->AttribPtr[_TNL_ATTRIB_POS]; - else - VB->EyePtr = TransformRaw( &store->eye, - ctx->ModelviewMatrixStack.Top, - VB->AttribPtr[_TNL_ATTRIB_POS]); - } - - VB->ClipPtr = TransformRaw( &store->clip, - &ctx->_ModelProjectMatrix, - VB->AttribPtr[_TNL_ATTRIB_POS] ); - - /* Drivers expect this to be clean to element 4... - */ - switch (VB->ClipPtr->size) { - case 1: - /* impossible */ - case 2: - _mesa_vector4f_clean_elem( VB->ClipPtr, VB->Count, 2 ); - /* fall-through */ - case 3: - _mesa_vector4f_clean_elem( VB->ClipPtr, VB->Count, 3 ); - /* fall-through */ - case 4: - break; - } - - - /* Cliptest and perspective divide. Clip functions must clear - * the clipmask. - */ - store->ormask = 0; - store->andmask = CLIP_FRUSTUM_BITS; - - if (tnl->NeedNdcCoords) { - VB->NdcPtr = - _mesa_clip_tab[VB->ClipPtr->size]( VB->ClipPtr, - &store->proj, - store->clipmask, - &store->ormask, - &store->andmask, - !ctx->Transform.DepthClamp ); - } - else { - VB->NdcPtr = NULL; - _mesa_clip_np_tab[VB->ClipPtr->size]( VB->ClipPtr, - NULL, - store->clipmask, - &store->ormask, - &store->andmask, - !ctx->Transform.DepthClamp ); - } - - if (store->andmask) - return GL_FALSE; - - - /* Test userclip planes. This contributes to VB->ClipMask, so - * is essentially required to be in this stage. - */ - if (ctx->Transform.ClipPlanesEnabled) { - usercliptab[VB->ClipPtr->size]( ctx, - VB->ClipPtr, - store->clipmask, - &store->ormask, - &store->andmask ); - - if (store->andmask) - return GL_FALSE; - } - - VB->ClipAndMask = store->andmask; - VB->ClipOrMask = store->ormask; - VB->ClipMask = store->clipmask; - - return GL_TRUE; -} - - -static GLboolean init_vertex_stage( struct gl_context *ctx, - struct tnl_pipeline_stage *stage ) -{ - struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; - struct vertex_stage_data *store; - GLuint size = VB->Size; - - stage->privatePtr = CALLOC(sizeof(*store)); - store = VERTEX_STAGE_DATA(stage); - if (!store) - return GL_FALSE; - - _mesa_vector4f_alloc( &store->eye, 0, size, 32 ); - _mesa_vector4f_alloc( &store->clip, 0, size, 32 ); - _mesa_vector4f_alloc( &store->proj, 0, size, 32 ); - - store->clipmask = (GLubyte *) _mesa_align_malloc(sizeof(GLubyte)*size, 32 ); - - if (!store->clipmask || - !store->eye.data || - !store->clip.data || - !store->proj.data) - return GL_FALSE; - - return GL_TRUE; -} - -static void dtr( struct tnl_pipeline_stage *stage ) -{ - struct vertex_stage_data *store = VERTEX_STAGE_DATA(stage); - - if (store) { - _mesa_vector4f_free( &store->eye ); - _mesa_vector4f_free( &store->clip ); - _mesa_vector4f_free( &store->proj ); - _mesa_align_free( store->clipmask ); - FREE(store); - stage->privatePtr = NULL; - stage->run = init_vertex_stage; - } -} - - -const struct tnl_pipeline_stage _tnl_vertex_transform_stage = -{ - "modelview/project/cliptest/divide", - NULL, /* private data */ - init_vertex_stage, - dtr, /* destructor */ - NULL, - run_vertex_stage /* run -- initially set to init */ -}; +/* + * Mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + + +#include "main/glheader.h" +#include "main/colormac.h" +#include "main/macros.h" +#include "main/imports.h" +#include "main/mtypes.h" + +#include "math/m_xform.h" + +#include "t_context.h" +#include "t_pipeline.h" + + + +struct vertex_stage_data { + GLvector4f eye; + GLvector4f clip; + GLvector4f proj; + GLubyte *clipmask; + GLubyte ormask; + GLubyte andmask; +}; + +#define VERTEX_STAGE_DATA(stage) ((struct vertex_stage_data *)stage->privatePtr) + + + + +/* This function implements cliptesting for user-defined clip planes. + * The clipping of primitives to these planes is implemented in + * t_render_clip.h. + */ +#define USER_CLIPTEST(NAME, SZ) \ +static void NAME( struct gl_context *ctx, \ + GLvector4f *clip, \ + GLubyte *clipmask, \ + GLubyte *clipormask, \ + GLubyte *clipandmask ) \ +{ \ + GLuint p; \ + \ + for (p = 0; p < ctx->Const.MaxClipPlanes; p++) \ + if (ctx->Transform.ClipPlanesEnabled & (1 << p)) { \ + GLuint nr, i; \ + const GLfloat a = ctx->Transform._ClipUserPlane[p][0]; \ + const GLfloat b = ctx->Transform._ClipUserPlane[p][1]; \ + const GLfloat c = ctx->Transform._ClipUserPlane[p][2]; \ + const GLfloat d = ctx->Transform._ClipUserPlane[p][3]; \ + GLfloat *coord = (GLfloat *)clip->data; \ + GLuint stride = clip->stride; \ + GLuint count = clip->count; \ + \ + for (nr = 0, i = 0 ; i < count ; i++) { \ + GLfloat dp = coord[0] * a + coord[1] * b; \ + if (SZ > 2) dp += coord[2] * c; \ + if (SZ > 3) dp += coord[3] * d; else dp += d; \ + \ + if (dp < 0) { \ + nr++; \ + clipmask[i] |= CLIP_USER_BIT; \ + } \ + \ + STRIDE_F(coord, stride); \ + } \ + \ + if (nr > 0) { \ + *clipormask |= CLIP_USER_BIT; \ + if (nr == count) { \ + *clipandmask |= CLIP_USER_BIT; \ + return; \ + } \ + } \ + } \ +} + + +USER_CLIPTEST(userclip2, 2) +USER_CLIPTEST(userclip3, 3) +USER_CLIPTEST(userclip4, 4) + +static void (*(usercliptab[5]))( struct gl_context *, + GLvector4f *, GLubyte *, + GLubyte *, GLubyte * ) = +{ + NULL, + NULL, + userclip2, + userclip3, + userclip4 +}; + + +void +tnl_clip_prepare(struct gl_context *ctx) +{ + /* Neither the x86 nor sparc asm cliptest functions have been updated + * for ARB_depth_clamp, so force the C paths. + */ + if (ctx->Transform.DepthClamp) { + static GLboolean c_funcs_installed = GL_FALSE; + if (!c_funcs_installed) { + init_c_cliptest(); + c_funcs_installed = GL_TRUE; + } + } +} + + + +static GLboolean run_vertex_stage( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + struct vertex_stage_data *store = (struct vertex_stage_data *)stage->privatePtr; + TNLcontext *tnl = TNL_CONTEXT(ctx); + struct vertex_buffer *VB = &tnl->vb; + + if (ctx->VertexProgram._Current) + return GL_TRUE; + + tnl_clip_prepare(ctx); + + if (ctx->_NeedEyeCoords) { + /* Separate modelview transformation: + * Use combined ModelProject to avoid some depth artifacts + */ + if (ctx->ModelviewMatrixStack.Top->type == MATRIX_IDENTITY) + VB->EyePtr = VB->AttribPtr[_TNL_ATTRIB_POS]; + else + VB->EyePtr = TransformRaw( &store->eye, + ctx->ModelviewMatrixStack.Top, + VB->AttribPtr[_TNL_ATTRIB_POS]); + } + + VB->ClipPtr = TransformRaw( &store->clip, + &ctx->_ModelProjectMatrix, + VB->AttribPtr[_TNL_ATTRIB_POS] ); + + /* Drivers expect this to be clean to element 4... + */ + switch (VB->ClipPtr->size) { + case 1: + /* impossible */ + case 2: + _mesa_vector4f_clean_elem( VB->ClipPtr, VB->Count, 2 ); + /* fall-through */ + case 3: + _mesa_vector4f_clean_elem( VB->ClipPtr, VB->Count, 3 ); + /* fall-through */ + case 4: + break; + } + + + /* Cliptest and perspective divide. Clip functions must clear + * the clipmask. + */ + store->ormask = 0; + store->andmask = CLIP_FRUSTUM_BITS; + + if (tnl->NeedNdcCoords) { + VB->NdcPtr = + _mesa_clip_tab[VB->ClipPtr->size]( VB->ClipPtr, + &store->proj, + store->clipmask, + &store->ormask, + &store->andmask, + !ctx->Transform.DepthClamp ); + } + else { + VB->NdcPtr = NULL; + _mesa_clip_np_tab[VB->ClipPtr->size]( VB->ClipPtr, + NULL, + store->clipmask, + &store->ormask, + &store->andmask, + !ctx->Transform.DepthClamp ); + } + + if (store->andmask) + return GL_FALSE; + + + /* Test userclip planes. This contributes to VB->ClipMask, so + * is essentially required to be in this stage. + */ + if (ctx->Transform.ClipPlanesEnabled) { + usercliptab[VB->ClipPtr->size]( ctx, + VB->ClipPtr, + store->clipmask, + &store->ormask, + &store->andmask ); + + if (store->andmask) + return GL_FALSE; + } + + VB->ClipAndMask = store->andmask; + VB->ClipOrMask = store->ormask; + VB->ClipMask = store->clipmask; + + return GL_TRUE; +} + + +static GLboolean init_vertex_stage( struct gl_context *ctx, + struct tnl_pipeline_stage *stage ) +{ + struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb; + struct vertex_stage_data *store; + GLuint size = VB->Size; + + stage->privatePtr = CALLOC(sizeof(*store)); + store = VERTEX_STAGE_DATA(stage); + if (!store) + return GL_FALSE; + + _mesa_vector4f_alloc( &store->eye, 0, size, 32 ); + _mesa_vector4f_alloc( &store->clip, 0, size, 32 ); + _mesa_vector4f_alloc( &store->proj, 0, size, 32 ); + + store->clipmask = (GLubyte *) _mesa_align_malloc(sizeof(GLubyte)*size, 32 ); + + if (!store->clipmask || + !store->eye.data || + !store->clip.data || + !store->proj.data) + return GL_FALSE; + + return GL_TRUE; +} + +static void dtr( struct tnl_pipeline_stage *stage ) +{ + struct vertex_stage_data *store = VERTEX_STAGE_DATA(stage); + + if (store) { + _mesa_vector4f_free( &store->eye ); + _mesa_vector4f_free( &store->clip ); + _mesa_vector4f_free( &store->proj ); + _mesa_align_free( store->clipmask ); + FREE(store); + stage->privatePtr = NULL; + stage->run = init_vertex_stage; + } +} + + +const struct tnl_pipeline_stage _tnl_vertex_transform_stage = +{ + "modelview/project/cliptest/divide", + NULL, /* private data */ + init_vertex_stage, + dtr, /* destructor */ + NULL, + run_vertex_stage /* run -- initially set to init */ +}; diff --git a/mesalib/src/mesa/tnl/t_vertex.h b/mesalib/src/mesa/tnl/t_vertex.h index bdbc88aff..83b0dbceb 100644 --- a/mesalib/src/mesa/tnl/t_vertex.h +++ b/mesalib/src/mesa/tnl/t_vertex.h @@ -1,184 +1,184 @@ -/* - * Copyright 2003 Tungsten Graphics, inc. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * on the rights to use, copy, modify, merge, publish, distribute, sub - * license, and/or sell copies of the Software, and to permit persons to whom - * the Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL - * TUNGSTEN GRAPHICS AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM, - * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR - * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE - * USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - -#ifndef _TNL_VERTEX_H -#define _TNL_VERTEX_H - -#include "main/glheader.h" -#include "t_context.h" - -struct gl_context; -struct tnl_clipspace; - -/* New mechanism to specify hardware vertices so that tnl can build - * and manipulate them directly. - */ - - -/* It will probably be necessary to allow drivers to specify new - * emit-styles to cover all the wierd and wacky things out there. - */ -enum tnl_attr_format { - EMIT_1F, - EMIT_2F, - EMIT_3F, - EMIT_4F, - EMIT_2F_VIEWPORT, /* do viewport transform and emit */ - EMIT_3F_VIEWPORT, /* do viewport transform and emit */ - EMIT_4F_VIEWPORT, /* do viewport transform and emit */ - EMIT_3F_XYW, /* for projective texture */ - EMIT_1UB_1F, /* for fog coordinate */ - EMIT_3UB_3F_RGB, /* for specular color */ - EMIT_3UB_3F_BGR, /* for specular color */ - EMIT_4UB_4F_RGBA, /* for color */ - EMIT_4UB_4F_BGRA, /* for color */ - EMIT_4UB_4F_ARGB, /* for color */ - EMIT_4UB_4F_ABGR, /* for color */ - EMIT_4CHAN_4F_RGBA, /* for swrast color */ - EMIT_PAD, /* leave a hole of 'offset' bytes */ - EMIT_MAX -}; - -struct tnl_attr_map { - GLuint attrib; /* _TNL_ATTRIB_ enum */ - enum tnl_attr_format format; - GLuint offset; -}; - -struct tnl_format_info { - const char *name; - tnl_extract_func extract; - tnl_insert_func insert[4]; - const GLuint attrsize; -}; - -extern const struct tnl_format_info _tnl_format_info[EMIT_MAX]; - - -/* Interpolate between two vertices to produce a third: - */ -extern void _tnl_interp( struct gl_context *ctx, - GLfloat t, - GLuint edst, GLuint eout, GLuint ein, - GLboolean force_boundary ); - -/* Copy colors from one vertex to another: - */ -extern void _tnl_copy_pv( struct gl_context *ctx, GLuint edst, GLuint esrc ); - - -/* Extract a named attribute from a hardware vertex. Will have to - * reverse any viewport transformation, swizzling or other conversions - * which may have been applied: - */ -extern void _tnl_get_attr( struct gl_context *ctx, const void *vertex, GLenum attrib, - GLfloat *dest ); - -/* Complementary to the above. - */ -extern void _tnl_set_attr( struct gl_context *ctx, void *vout, GLenum attrib, - const GLfloat *src ); - - -extern void *_tnl_get_vertex( struct gl_context *ctx, GLuint nr ); - -extern GLuint _tnl_install_attrs( struct gl_context *ctx, - const struct tnl_attr_map *map, - GLuint nr, const GLfloat *vp, - GLuint unpacked_size ); - -extern void _tnl_free_vertices( struct gl_context *ctx ); - -extern void _tnl_init_vertices( struct gl_context *ctx, - GLuint vb_size, - GLuint max_vertex_size ); - -extern void *_tnl_emit_vertices_to_buffer( struct gl_context *ctx, - GLuint start, - GLuint end, - void *dest ); - -/* This function isn't optimal. Check out - * gallium/auxilary/translate for a more comprehensive implementation of - * the same functionality. - */ - -extern void *_tnl_emit_indexed_vertices_to_buffer( struct gl_context *ctx, - const GLuint *elts, - GLuint start, - GLuint end, - void *dest ); - - -extern void _tnl_build_vertices( struct gl_context *ctx, - GLuint start, - GLuint end, - GLuint newinputs ); - -extern void _tnl_invalidate_vertices( struct gl_context *ctx, GLuint newinputs ); - -extern void _tnl_invalidate_vertex_state( struct gl_context *ctx, GLuint new_state ); - -extern void _tnl_notify_pipeline_output_change( struct gl_context *ctx ); - - -#define GET_VERTEX_STATE(ctx) &(TNL_CONTEXT(ctx)->clipspace) - -/* Internal function: - */ -void _tnl_register_fastpath( struct tnl_clipspace *vtx, - GLboolean match_strides ); - - -/* t_vertex_generic.c -- Internal functions for t_vertex.c - */ -void _tnl_generic_copy_pv_extras( struct gl_context *ctx, - GLuint dst, GLuint src ); - -void _tnl_generic_interp_extras( struct gl_context *ctx, - GLfloat t, - GLuint dst, GLuint out, GLuint in, - GLboolean force_boundary ); - -void _tnl_generic_copy_pv( struct gl_context *ctx, GLuint edst, GLuint esrc ); - -void _tnl_generic_interp( struct gl_context *ctx, - GLfloat t, - GLuint edst, GLuint eout, GLuint ein, - GLboolean force_boundary ); - -void _tnl_generic_emit( struct gl_context *ctx, - GLuint count, - GLubyte *v ); - -void _tnl_generate_hardwired_emit( struct gl_context *ctx ); - -/* t_vertex_sse.c -- Internal functions for t_vertex.c - */ -void _tnl_generate_sse_emit( struct gl_context *ctx ); - -#endif +/* + * Copyright 2003 Tungsten Graphics, inc. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * on the rights to use, copy, modify, merge, publish, distribute, sub + * license, and/or sell copies of the Software, and to permit persons to whom + * the Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * TUNGSTEN GRAPHICS AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + +#ifndef _TNL_VERTEX_H +#define _TNL_VERTEX_H + +#include "main/glheader.h" +#include "t_context.h" + +struct gl_context; +struct tnl_clipspace; + +/* New mechanism to specify hardware vertices so that tnl can build + * and manipulate them directly. + */ + + +/* It will probably be necessary to allow drivers to specify new + * emit-styles to cover all the wierd and wacky things out there. + */ +enum tnl_attr_format { + EMIT_1F, + EMIT_2F, + EMIT_3F, + EMIT_4F, + EMIT_2F_VIEWPORT, /* do viewport transform and emit */ + EMIT_3F_VIEWPORT, /* do viewport transform and emit */ + EMIT_4F_VIEWPORT, /* do viewport transform and emit */ + EMIT_3F_XYW, /* for projective texture */ + EMIT_1UB_1F, /* for fog coordinate */ + EMIT_3UB_3F_RGB, /* for specular color */ + EMIT_3UB_3F_BGR, /* for specular color */ + EMIT_4UB_4F_RGBA, /* for color */ + EMIT_4UB_4F_BGRA, /* for color */ + EMIT_4UB_4F_ARGB, /* for color */ + EMIT_4UB_4F_ABGR, /* for color */ + EMIT_4CHAN_4F_RGBA, /* for swrast color */ + EMIT_PAD, /* leave a hole of 'offset' bytes */ + EMIT_MAX +}; + +struct tnl_attr_map { + GLuint attrib; /* _TNL_ATTRIB_ enum */ + enum tnl_attr_format format; + GLuint offset; +}; + +struct tnl_format_info { + const char *name; + tnl_extract_func extract; + tnl_insert_func insert[4]; + const GLuint attrsize; +}; + +extern const struct tnl_format_info _tnl_format_info[EMIT_MAX]; + + +/* Interpolate between two vertices to produce a third: + */ +extern void _tnl_interp( struct gl_context *ctx, + GLfloat t, + GLuint edst, GLuint eout, GLuint ein, + GLboolean force_boundary ); + +/* Copy colors from one vertex to another: + */ +extern void _tnl_copy_pv( struct gl_context *ctx, GLuint edst, GLuint esrc ); + + +/* Extract a named attribute from a hardware vertex. Will have to + * reverse any viewport transformation, swizzling or other conversions + * which may have been applied: + */ +extern void _tnl_get_attr( struct gl_context *ctx, const void *vertex, GLenum attrib, + GLfloat *dest ); + +/* Complementary to the above. + */ +extern void _tnl_set_attr( struct gl_context *ctx, void *vout, GLenum attrib, + const GLfloat *src ); + + +extern void *_tnl_get_vertex( struct gl_context *ctx, GLuint nr ); + +extern GLuint _tnl_install_attrs( struct gl_context *ctx, + const struct tnl_attr_map *map, + GLuint nr, const GLfloat *vp, + GLuint unpacked_size ); + +extern void _tnl_free_vertices( struct gl_context *ctx ); + +extern void _tnl_init_vertices( struct gl_context *ctx, + GLuint vb_size, + GLuint max_vertex_size ); + +extern void *_tnl_emit_vertices_to_buffer( struct gl_context *ctx, + GLuint start, + GLuint end, + void *dest ); + +/* This function isn't optimal. Check out + * gallium/auxilary/translate for a more comprehensive implementation of + * the same functionality. + */ + +extern void *_tnl_emit_indexed_vertices_to_buffer( struct gl_context *ctx, + const GLuint *elts, + GLuint start, + GLuint end, + void *dest ); + + +extern void _tnl_build_vertices( struct gl_context *ctx, + GLuint start, + GLuint end, + GLuint newinputs ); + +extern void _tnl_invalidate_vertices( struct gl_context *ctx, GLuint newinputs ); + +extern void _tnl_invalidate_vertex_state( struct gl_context *ctx, GLuint new_state ); + +extern void _tnl_notify_pipeline_output_change( struct gl_context *ctx ); + + +#define GET_VERTEX_STATE(ctx) &(TNL_CONTEXT(ctx)->clipspace) + +/* Internal function: + */ +void _tnl_register_fastpath( struct tnl_clipspace *vtx, + GLboolean match_strides ); + + +/* t_vertex_generic.c -- Internal functions for t_vertex.c + */ +void _tnl_generic_copy_pv_extras( struct gl_context *ctx, + GLuint dst, GLuint src ); + +void _tnl_generic_interp_extras( struct gl_context *ctx, + GLfloat t, + GLuint dst, GLuint out, GLuint in, + GLboolean force_boundary ); + +void _tnl_generic_copy_pv( struct gl_context *ctx, GLuint edst, GLuint esrc ); + +void _tnl_generic_interp( struct gl_context *ctx, + GLfloat t, + GLuint edst, GLuint eout, GLuint ein, + GLboolean force_boundary ); + +void _tnl_generic_emit( struct gl_context *ctx, + GLuint count, + GLubyte *v ); + +void _tnl_generate_hardwired_emit( struct gl_context *ctx ); + +/* t_vertex_sse.c -- Internal functions for t_vertex.c + */ +void _tnl_generate_sse_emit( struct gl_context *ctx ); + +#endif diff --git a/mesalib/src/mesa/tnl/t_vp_build.c b/mesalib/src/mesa/tnl/t_vp_build.c index 913cba676..70492a435 100644 --- a/mesalib/src/mesa/tnl/t_vp_build.c +++ b/mesalib/src/mesa/tnl/t_vp_build.c @@ -1,60 +1,60 @@ -/* - * Mesa 3-D graphics library - * Version: 7.1 - * - * Copyright (C) 2007 Tungsten Graphics All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * TUNGSTEN GRAPHICS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, - * WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file t_vp_build.c - * Create a vertex program to execute the current fixed function T&L pipeline. - * \author Keith Whitwell - */ - - -#include "main/glheader.h" -#include "main/ffvertex_prog.h" -#include "main/mtypes.h" -#include "t_vp_build.h" - - -/** - * XXX This should go away someday, but still referenced by some drivers... - */ -void _tnl_UpdateFixedFunctionProgram( struct gl_context *ctx ) -{ - const struct gl_vertex_program *prev = ctx->VertexProgram._Current; - - if (!ctx->VertexProgram._Current || - ctx->VertexProgram._Current == ctx->VertexProgram._TnlProgram) { - ctx->VertexProgram._Current - = ctx->VertexProgram._TnlProgram - = _mesa_get_fixed_func_vertex_program(ctx); - } - - /* Tell the driver about the change. Could define a new target for - * this? - */ - if (ctx->VertexProgram._Current != prev && ctx->Driver.BindProgram) { - ctx->Driver.BindProgram(ctx, GL_VERTEX_PROGRAM_ARB, - (struct gl_program *) ctx->VertexProgram._Current); - } -} +/* + * Mesa 3-D graphics library + * Version: 7.1 + * + * Copyright (C) 2007 Tungsten Graphics All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * TUNGSTEN GRAPHICS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + * WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file t_vp_build.c + * Create a vertex program to execute the current fixed function T&L pipeline. + * \author Keith Whitwell + */ + + +#include "main/glheader.h" +#include "main/ffvertex_prog.h" +#include "main/mtypes.h" +#include "t_vp_build.h" + + +/** + * XXX This should go away someday, but still referenced by some drivers... + */ +void _tnl_UpdateFixedFunctionProgram( struct gl_context *ctx ) +{ + const struct gl_vertex_program *prev = ctx->VertexProgram._Current; + + if (!ctx->VertexProgram._Current || + ctx->VertexProgram._Current == ctx->VertexProgram._TnlProgram) { + ctx->VertexProgram._Current + = ctx->VertexProgram._TnlProgram + = _mesa_get_fixed_func_vertex_program(ctx); + } + + /* Tell the driver about the change. Could define a new target for + * this? + */ + if (ctx->VertexProgram._Current != prev && ctx->Driver.BindProgram) { + ctx->Driver.BindProgram(ctx, GL_VERTEX_PROGRAM_ARB, + (struct gl_program *) ctx->VertexProgram._Current); + } +} diff --git a/mesalib/src/mesa/tnl/t_vp_build.h b/mesalib/src/mesa/tnl/t_vp_build.h index 6cba2ce5a..e9f6de5a9 100644 --- a/mesalib/src/mesa/tnl/t_vp_build.h +++ b/mesalib/src/mesa/tnl/t_vp_build.h @@ -1,42 +1,42 @@ -/* - * Mesa 3-D graphics library - * Version: 6.3 - * - * Copyright (C) 2005 Tungsten Graphics All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * TUNGSTEN GRAPHICS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, - * WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef T_VP_BUILD_H -#define T_VP_BUILD_H - -struct gl_context; - -#define TNL_FIXED_FUNCTION_STATE_FLAGS (_NEW_PROGRAM | \ - _NEW_LIGHT | \ - _NEW_TEXTURE | \ - _NEW_TEXTURE_MATRIX | \ - _NEW_TRANSFORM | \ - _NEW_FOG | \ - _NEW_POINT) - -extern void _tnl_UpdateFixedFunctionProgram( struct gl_context *ctx ); - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.3 + * + * Copyright (C) 2005 Tungsten Graphics All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * TUNGSTEN GRAPHICS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + * WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef T_VP_BUILD_H +#define T_VP_BUILD_H + +struct gl_context; + +#define TNL_FIXED_FUNCTION_STATE_FLAGS (_NEW_PROGRAM | \ + _NEW_LIGHT | \ + _NEW_TEXTURE | \ + _NEW_TEXTURE_MATRIX | \ + _NEW_TRANSFORM | \ + _NEW_FOG | \ + _NEW_POINT) + +extern void _tnl_UpdateFixedFunctionProgram( struct gl_context *ctx ); + +#endif diff --git a/mesalib/src/mesa/vbo/vbo_exec_eval.c b/mesalib/src/mesa/vbo/vbo_exec_eval.c index 0aff01b25..1e8c3c45b 100644 --- a/mesalib/src/mesa/vbo/vbo_exec_eval.c +++ b/mesalib/src/mesa/vbo/vbo_exec_eval.c @@ -1,263 +1,263 @@ -/* - * Mesa 3-D graphics library - * Version: 6.1 - * - * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Keith Whitwell - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "main/macros.h" -#include "math/m_eval.h" -#include "main/dispatch.h" -#include "vbo_exec.h" - - -static void clear_active_eval1( struct vbo_exec_context *exec, GLuint attr ) -{ - assert(attr < Elements(exec->eval.map1)); - exec->eval.map1[attr].map = NULL; -} - -static void clear_active_eval2( struct vbo_exec_context *exec, GLuint attr ) -{ - assert(attr < Elements(exec->eval.map2)); - exec->eval.map2[attr].map = NULL; -} - -static void set_active_eval1( struct vbo_exec_context *exec, GLuint attr, GLuint dim, - struct gl_1d_map *map ) -{ - assert(attr < Elements(exec->eval.map1)); - if (!exec->eval.map1[attr].map) { - exec->eval.map1[attr].map = map; - exec->eval.map1[attr].sz = dim; - } -} - -static void set_active_eval2( struct vbo_exec_context *exec, GLuint attr, GLuint dim, - struct gl_2d_map *map ) -{ - assert(attr < Elements(exec->eval.map2)); - if (!exec->eval.map2[attr].map) { - exec->eval.map2[attr].map = map; - exec->eval.map2[attr].sz = dim; - } -} - -void vbo_exec_eval_update( struct vbo_exec_context *exec ) -{ - struct gl_context *ctx = exec->ctx; - GLuint attr; - - /* Vertex program maps have priority over conventional attribs */ - - for (attr = 0; attr < VBO_ATTRIB_FIRST_MATERIAL; attr++) { - clear_active_eval1( exec, attr ); - clear_active_eval2( exec, attr ); - } - - if (ctx->Eval.Map1Color4) - set_active_eval1( exec, VBO_ATTRIB_COLOR0, 4, &ctx->EvalMap.Map1Color4 ); - - if (ctx->Eval.Map2Color4) - set_active_eval2( exec, VBO_ATTRIB_COLOR0, 4, &ctx->EvalMap.Map2Color4 ); - - if (ctx->Eval.Map1TextureCoord4) - set_active_eval1( exec, VBO_ATTRIB_TEX0, 4, &ctx->EvalMap.Map1Texture4 ); - else if (ctx->Eval.Map1TextureCoord3) - set_active_eval1( exec, VBO_ATTRIB_TEX0, 3, &ctx->EvalMap.Map1Texture3 ); - else if (ctx->Eval.Map1TextureCoord2) - set_active_eval1( exec, VBO_ATTRIB_TEX0, 2, &ctx->EvalMap.Map1Texture2 ); - else if (ctx->Eval.Map1TextureCoord1) - set_active_eval1( exec, VBO_ATTRIB_TEX0, 1, &ctx->EvalMap.Map1Texture1 ); - - if (ctx->Eval.Map2TextureCoord4) - set_active_eval2( exec, VBO_ATTRIB_TEX0, 4, &ctx->EvalMap.Map2Texture4 ); - else if (ctx->Eval.Map2TextureCoord3) - set_active_eval2( exec, VBO_ATTRIB_TEX0, 3, &ctx->EvalMap.Map2Texture3 ); - else if (ctx->Eval.Map2TextureCoord2) - set_active_eval2( exec, VBO_ATTRIB_TEX0, 2, &ctx->EvalMap.Map2Texture2 ); - else if (ctx->Eval.Map2TextureCoord1) - set_active_eval2( exec, VBO_ATTRIB_TEX0, 1, &ctx->EvalMap.Map2Texture1 ); - - if (ctx->Eval.Map1Normal) - set_active_eval1( exec, VBO_ATTRIB_NORMAL, 3, &ctx->EvalMap.Map1Normal ); - - if (ctx->Eval.Map2Normal) - set_active_eval2( exec, VBO_ATTRIB_NORMAL, 3, &ctx->EvalMap.Map2Normal ); - - if (ctx->Eval.Map1Vertex4) - set_active_eval1( exec, VBO_ATTRIB_POS, 4, &ctx->EvalMap.Map1Vertex4 ); - else if (ctx->Eval.Map1Vertex3) - set_active_eval1( exec, VBO_ATTRIB_POS, 3, &ctx->EvalMap.Map1Vertex3 ); - - if (ctx->Eval.Map2Vertex4) - set_active_eval2( exec, VBO_ATTRIB_POS, 4, &ctx->EvalMap.Map2Vertex4 ); - else if (ctx->Eval.Map2Vertex3) - set_active_eval2( exec, VBO_ATTRIB_POS, 3, &ctx->EvalMap.Map2Vertex3 ); - - /* _NEW_PROGRAM */ - if (ctx->VertexProgram._Enabled) { - /* These are the 16 evaluators which GL_NV_vertex_program defines. - * They alias and override the conventional vertex attributs. - */ - for (attr = 0; attr < 16; attr++) { - /* _NEW_EVAL */ - assert(attr < Elements(ctx->Eval.Map1Attrib)); - if (ctx->Eval.Map1Attrib[attr]) - set_active_eval1( exec, attr, 4, &ctx->EvalMap.Map1Attrib[attr] ); - - assert(attr < Elements(ctx->Eval.Map2Attrib)); - if (ctx->Eval.Map2Attrib[attr]) - set_active_eval2( exec, attr, 4, &ctx->EvalMap.Map2Attrib[attr] ); - } - } - - exec->eval.recalculate_maps = 0; -} - - - -void vbo_exec_do_EvalCoord1f(struct vbo_exec_context *exec, GLfloat u) -{ - GLuint attr; - - for (attr = 1; attr <= VBO_ATTRIB_TEX7; attr++) { - struct gl_1d_map *map = exec->eval.map1[attr].map; - if (map) { - GLfloat uu = (u - map->u1) * map->du; - GLfloat data[4]; - - ASSIGN_4V(data, 0, 0, 0, 1); - - _math_horner_bezier_curve(map->Points, data, uu, - exec->eval.map1[attr].sz, - map->Order); - - COPY_SZ_4V( exec->vtx.attrptr[attr], - exec->vtx.attrsz[attr], - data ); - } - } - - /** Vertex -- EvalCoord1f is a noop if this map not enabled: - **/ - if (exec->eval.map1[0].map) { - struct gl_1d_map *map = exec->eval.map1[0].map; - GLfloat uu = (u - map->u1) * map->du; - GLfloat vertex[4]; - - ASSIGN_4V(vertex, 0, 0, 0, 1); - - _math_horner_bezier_curve(map->Points, vertex, uu, - exec->eval.map1[0].sz, - map->Order); - - if (exec->eval.map1[0].sz == 4) - CALL_Vertex4fv(GET_DISPATCH(), ( vertex )); - else - CALL_Vertex3fv(GET_DISPATCH(), ( vertex )); - } -} - - - -void vbo_exec_do_EvalCoord2f( struct vbo_exec_context *exec, - GLfloat u, GLfloat v ) -{ - GLuint attr; - - for (attr = 1; attr <= VBO_ATTRIB_TEX7; attr++) { - struct gl_2d_map *map = exec->eval.map2[attr].map; - if (map) { - GLfloat uu = (u - map->u1) * map->du; - GLfloat vv = (v - map->v1) * map->dv; - GLfloat data[4]; - - ASSIGN_4V(data, 0, 0, 0, 1); - - _math_horner_bezier_surf(map->Points, - data, - uu, vv, - exec->eval.map2[attr].sz, - map->Uorder, map->Vorder); - - COPY_SZ_4V( exec->vtx.attrptr[attr], - exec->vtx.attrsz[attr], - data ); - } - } - - /** Vertex -- EvalCoord2f is a noop if this map not enabled: - **/ - if (exec->eval.map2[0].map) { - struct gl_2d_map *map = exec->eval.map2[0].map; - GLfloat uu = (u - map->u1) * map->du; - GLfloat vv = (v - map->v1) * map->dv; - GLfloat vertex[4]; - - ASSIGN_4V(vertex, 0, 0, 0, 1); - - if (exec->ctx->Eval.AutoNormal) { - GLfloat normal[4]; - GLfloat du[4], dv[4]; - - _math_de_casteljau_surf(map->Points, vertex, du, dv, uu, vv, - exec->eval.map2[0].sz, - map->Uorder, map->Vorder); - - if (exec->eval.map2[0].sz == 4) { - du[0] = du[0]*vertex[3] - du[3]*vertex[0]; - du[1] = du[1]*vertex[3] - du[3]*vertex[1]; - du[2] = du[2]*vertex[3] - du[3]*vertex[2]; - - dv[0] = dv[0]*vertex[3] - dv[3]*vertex[0]; - dv[1] = dv[1]*vertex[3] - dv[3]*vertex[1]; - dv[2] = dv[2]*vertex[3] - dv[3]*vertex[2]; - } - - - CROSS3(normal, du, dv); - NORMALIZE_3FV(normal); - normal[3] = 1.0; - - COPY_SZ_4V( exec->vtx.attrptr[VBO_ATTRIB_NORMAL], - exec->vtx.attrsz[VBO_ATTRIB_NORMAL], - normal ); - - } - else { - _math_horner_bezier_surf(map->Points, vertex, uu, vv, - exec->eval.map2[0].sz, - map->Uorder, map->Vorder); - } - - if (exec->vtx.attrsz[0] == 4) - CALL_Vertex4fv(GET_DISPATCH(), ( vertex )); - else - CALL_Vertex3fv(GET_DISPATCH(), ( vertex )); - } -} - - +/* + * Mesa 3-D graphics library + * Version: 6.1 + * + * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Keith Whitwell + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "main/macros.h" +#include "math/m_eval.h" +#include "main/dispatch.h" +#include "vbo_exec.h" + + +static void clear_active_eval1( struct vbo_exec_context *exec, GLuint attr ) +{ + assert(attr < Elements(exec->eval.map1)); + exec->eval.map1[attr].map = NULL; +} + +static void clear_active_eval2( struct vbo_exec_context *exec, GLuint attr ) +{ + assert(attr < Elements(exec->eval.map2)); + exec->eval.map2[attr].map = NULL; +} + +static void set_active_eval1( struct vbo_exec_context *exec, GLuint attr, GLuint dim, + struct gl_1d_map *map ) +{ + assert(attr < Elements(exec->eval.map1)); + if (!exec->eval.map1[attr].map) { + exec->eval.map1[attr].map = map; + exec->eval.map1[attr].sz = dim; + } +} + +static void set_active_eval2( struct vbo_exec_context *exec, GLuint attr, GLuint dim, + struct gl_2d_map *map ) +{ + assert(attr < Elements(exec->eval.map2)); + if (!exec->eval.map2[attr].map) { + exec->eval.map2[attr].map = map; + exec->eval.map2[attr].sz = dim; + } +} + +void vbo_exec_eval_update( struct vbo_exec_context *exec ) +{ + struct gl_context *ctx = exec->ctx; + GLuint attr; + + /* Vertex program maps have priority over conventional attribs */ + + for (attr = 0; attr < VBO_ATTRIB_FIRST_MATERIAL; attr++) { + clear_active_eval1( exec, attr ); + clear_active_eval2( exec, attr ); + } + + if (ctx->Eval.Map1Color4) + set_active_eval1( exec, VBO_ATTRIB_COLOR0, 4, &ctx->EvalMap.Map1Color4 ); + + if (ctx->Eval.Map2Color4) + set_active_eval2( exec, VBO_ATTRIB_COLOR0, 4, &ctx->EvalMap.Map2Color4 ); + + if (ctx->Eval.Map1TextureCoord4) + set_active_eval1( exec, VBO_ATTRIB_TEX0, 4, &ctx->EvalMap.Map1Texture4 ); + else if (ctx->Eval.Map1TextureCoord3) + set_active_eval1( exec, VBO_ATTRIB_TEX0, 3, &ctx->EvalMap.Map1Texture3 ); + else if (ctx->Eval.Map1TextureCoord2) + set_active_eval1( exec, VBO_ATTRIB_TEX0, 2, &ctx->EvalMap.Map1Texture2 ); + else if (ctx->Eval.Map1TextureCoord1) + set_active_eval1( exec, VBO_ATTRIB_TEX0, 1, &ctx->EvalMap.Map1Texture1 ); + + if (ctx->Eval.Map2TextureCoord4) + set_active_eval2( exec, VBO_ATTRIB_TEX0, 4, &ctx->EvalMap.Map2Texture4 ); + else if (ctx->Eval.Map2TextureCoord3) + set_active_eval2( exec, VBO_ATTRIB_TEX0, 3, &ctx->EvalMap.Map2Texture3 ); + else if (ctx->Eval.Map2TextureCoord2) + set_active_eval2( exec, VBO_ATTRIB_TEX0, 2, &ctx->EvalMap.Map2Texture2 ); + else if (ctx->Eval.Map2TextureCoord1) + set_active_eval2( exec, VBO_ATTRIB_TEX0, 1, &ctx->EvalMap.Map2Texture1 ); + + if (ctx->Eval.Map1Normal) + set_active_eval1( exec, VBO_ATTRIB_NORMAL, 3, &ctx->EvalMap.Map1Normal ); + + if (ctx->Eval.Map2Normal) + set_active_eval2( exec, VBO_ATTRIB_NORMAL, 3, &ctx->EvalMap.Map2Normal ); + + if (ctx->Eval.Map1Vertex4) + set_active_eval1( exec, VBO_ATTRIB_POS, 4, &ctx->EvalMap.Map1Vertex4 ); + else if (ctx->Eval.Map1Vertex3) + set_active_eval1( exec, VBO_ATTRIB_POS, 3, &ctx->EvalMap.Map1Vertex3 ); + + if (ctx->Eval.Map2Vertex4) + set_active_eval2( exec, VBO_ATTRIB_POS, 4, &ctx->EvalMap.Map2Vertex4 ); + else if (ctx->Eval.Map2Vertex3) + set_active_eval2( exec, VBO_ATTRIB_POS, 3, &ctx->EvalMap.Map2Vertex3 ); + + /* _NEW_PROGRAM */ + if (ctx->VertexProgram._Enabled) { + /* These are the 16 evaluators which GL_NV_vertex_program defines. + * They alias and override the conventional vertex attributs. + */ + for (attr = 0; attr < 16; attr++) { + /* _NEW_EVAL */ + assert(attr < Elements(ctx->Eval.Map1Attrib)); + if (ctx->Eval.Map1Attrib[attr]) + set_active_eval1( exec, attr, 4, &ctx->EvalMap.Map1Attrib[attr] ); + + assert(attr < Elements(ctx->Eval.Map2Attrib)); + if (ctx->Eval.Map2Attrib[attr]) + set_active_eval2( exec, attr, 4, &ctx->EvalMap.Map2Attrib[attr] ); + } + } + + exec->eval.recalculate_maps = 0; +} + + + +void vbo_exec_do_EvalCoord1f(struct vbo_exec_context *exec, GLfloat u) +{ + GLuint attr; + + for (attr = 1; attr <= VBO_ATTRIB_TEX7; attr++) { + struct gl_1d_map *map = exec->eval.map1[attr].map; + if (map) { + GLfloat uu = (u - map->u1) * map->du; + GLfloat data[4]; + + ASSIGN_4V(data, 0, 0, 0, 1); + + _math_horner_bezier_curve(map->Points, data, uu, + exec->eval.map1[attr].sz, + map->Order); + + COPY_SZ_4V( exec->vtx.attrptr[attr], + exec->vtx.attrsz[attr], + data ); + } + } + + /** Vertex -- EvalCoord1f is a noop if this map not enabled: + **/ + if (exec->eval.map1[0].map) { + struct gl_1d_map *map = exec->eval.map1[0].map; + GLfloat uu = (u - map->u1) * map->du; + GLfloat vertex[4]; + + ASSIGN_4V(vertex, 0, 0, 0, 1); + + _math_horner_bezier_curve(map->Points, vertex, uu, + exec->eval.map1[0].sz, + map->Order); + + if (exec->eval.map1[0].sz == 4) + CALL_Vertex4fv(GET_DISPATCH(), ( vertex )); + else + CALL_Vertex3fv(GET_DISPATCH(), ( vertex )); + } +} + + + +void vbo_exec_do_EvalCoord2f( struct vbo_exec_context *exec, + GLfloat u, GLfloat v ) +{ + GLuint attr; + + for (attr = 1; attr <= VBO_ATTRIB_TEX7; attr++) { + struct gl_2d_map *map = exec->eval.map2[attr].map; + if (map) { + GLfloat uu = (u - map->u1) * map->du; + GLfloat vv = (v - map->v1) * map->dv; + GLfloat data[4]; + + ASSIGN_4V(data, 0, 0, 0, 1); + + _math_horner_bezier_surf(map->Points, + data, + uu, vv, + exec->eval.map2[attr].sz, + map->Uorder, map->Vorder); + + COPY_SZ_4V( exec->vtx.attrptr[attr], + exec->vtx.attrsz[attr], + data ); + } + } + + /** Vertex -- EvalCoord2f is a noop if this map not enabled: + **/ + if (exec->eval.map2[0].map) { + struct gl_2d_map *map = exec->eval.map2[0].map; + GLfloat uu = (u - map->u1) * map->du; + GLfloat vv = (v - map->v1) * map->dv; + GLfloat vertex[4]; + + ASSIGN_4V(vertex, 0, 0, 0, 1); + + if (exec->ctx->Eval.AutoNormal) { + GLfloat normal[4]; + GLfloat du[4], dv[4]; + + _math_de_casteljau_surf(map->Points, vertex, du, dv, uu, vv, + exec->eval.map2[0].sz, + map->Uorder, map->Vorder); + + if (exec->eval.map2[0].sz == 4) { + du[0] = du[0]*vertex[3] - du[3]*vertex[0]; + du[1] = du[1]*vertex[3] - du[3]*vertex[1]; + du[2] = du[2]*vertex[3] - du[3]*vertex[2]; + + dv[0] = dv[0]*vertex[3] - dv[3]*vertex[0]; + dv[1] = dv[1]*vertex[3] - dv[3]*vertex[1]; + dv[2] = dv[2]*vertex[3] - dv[3]*vertex[2]; + } + + + CROSS3(normal, du, dv); + NORMALIZE_3FV(normal); + normal[3] = 1.0; + + COPY_SZ_4V( exec->vtx.attrptr[VBO_ATTRIB_NORMAL], + exec->vtx.attrsz[VBO_ATTRIB_NORMAL], + normal ); + + } + else { + _math_horner_bezier_surf(map->Points, vertex, uu, vv, + exec->eval.map2[0].sz, + map->Uorder, map->Vorder); + } + + if (exec->vtx.attrsz[0] == 4) + CALL_Vertex4fv(GET_DISPATCH(), ( vertex )); + else + CALL_Vertex3fv(GET_DISPATCH(), ( vertex )); + } +} + + diff --git a/mesalib/src/mesa/vbo/vbo_save_loopback.c b/mesalib/src/mesa/vbo/vbo_save_loopback.c index 6a8860c78..51e598e7e 100644 --- a/mesalib/src/mesa/vbo/vbo_save_loopback.c +++ b/mesalib/src/mesa/vbo/vbo_save_loopback.c @@ -1,196 +1,196 @@ -/************************************************************************** - * - * Copyright 2005 Tungsten Graphics, Inc., Cedar Park, Texas. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF - * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. - * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR - * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, - * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE - * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - **************************************************************************/ - -#include "main/context.h" -#include "main/glheader.h" -#include "main/enums.h" -#include "main/imports.h" -#include "main/mfeatures.h" -#include "main/mtypes.h" -#include "main/dispatch.h" -#include "glapi/glapi.h" - -#include "vbo_context.h" - - -#if FEATURE_dlist - - -typedef void (*attr_func)( struct gl_context *ctx, GLint target, const GLfloat * ); - - -/* This file makes heavy use of the aliasing of NV vertex attributes - * with the legacy attributes, and also with ARB and Material - * attributes as currently implemented. - */ -static void VertexAttrib1fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) -{ - CALL_VertexAttrib1fvNV(ctx->Exec, (target, v)); -} - -static void VertexAttrib2fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) -{ - CALL_VertexAttrib2fvNV(ctx->Exec, (target, v)); -} - -static void VertexAttrib3fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) -{ - CALL_VertexAttrib3fvNV(ctx->Exec, (target, v)); -} - -static void VertexAttrib4fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) -{ - CALL_VertexAttrib4fvNV(ctx->Exec, (target, v)); -} - -static attr_func vert_attrfunc[4] = { - VertexAttrib1fvNV, - VertexAttrib2fvNV, - VertexAttrib3fvNV, - VertexAttrib4fvNV -}; - -struct loopback_attr { - GLint target; - GLint sz; - attr_func func; -}; - -/* Don't emit ends and begins on wrapped primitives. Don't replay - * wrapped vertices. If we get here, it's probably because the - * precalculated wrapping is wrong. - */ -static void loopback_prim( struct gl_context *ctx, - const GLfloat *buffer, - const struct _mesa_prim *prim, - GLuint wrap_count, - GLuint vertex_size, - const struct loopback_attr *la, GLuint nr ) -{ - GLint start = prim->start; - GLint end = start + prim->count; - const GLfloat *data; - GLint j; - GLuint k; - - if (0) - printf("loopback prim %s(%s,%s) verts %d..%d\n", - _mesa_lookup_prim_by_nr(prim->mode), - prim->begin ? "begin" : "..", - prim->end ? "end" : "..", - start, - end); - - if (prim->begin) { - CALL_Begin(GET_DISPATCH(), ( prim->mode )); - } - else { - assert(start == 0); - start += wrap_count; - } - - data = buffer + start * vertex_size; - - for (j = start ; j < end ; j++) { - const GLfloat *tmp = data + la[0].sz; - - for (k = 1 ; k < nr ; k++) { - la[k].func( ctx, la[k].target, tmp ); - tmp += la[k].sz; - } - - /* Fire the vertex - */ - la[0].func( ctx, VBO_ATTRIB_POS, data ); - data = tmp; - } - - if (prim->end) { - CALL_End(GET_DISPATCH(), ()); - } -} - -/* Primitives generated by DrawArrays/DrawElements/Rectf may be - * caught here. If there is no primitive in progress, execute them - * normally, otherwise need to track and discard the generated - * primitives. - */ -static void loopback_weak_prim( struct gl_context *ctx, - const struct _mesa_prim *prim ) -{ - /* Use the prim_weak flag to ensure that if this primitive - * wraps, we don't mistake future vertex_lists for part of the - * surrounding primitive. - * - * While this flag is set, we are simply disposing of data - * generated by an operation now known to be a noop. - */ - if (prim->begin) - ctx->Driver.CurrentExecPrimitive |= VBO_SAVE_PRIM_WEAK; - if (prim->end) - ctx->Driver.CurrentExecPrimitive &= ~VBO_SAVE_PRIM_WEAK; -} - - -void vbo_loopback_vertex_list( struct gl_context *ctx, - const GLfloat *buffer, - const GLubyte *attrsz, - const struct _mesa_prim *prim, - GLuint prim_count, - GLuint wrap_count, - GLuint vertex_size) -{ - struct loopback_attr la[VBO_ATTRIB_MAX]; - GLuint i, nr = 0; - - /* All Legacy, NV, ARB and Material attributes are routed through - * the NV attributes entrypoints: - */ - for (i = 0 ; i < VBO_ATTRIB_MAX ; i++) { - if (attrsz[i]) { - la[nr].target = i; - la[nr].sz = attrsz[i]; - la[nr].func = vert_attrfunc[attrsz[i]-1]; - nr++; - } - } - - for (i = 0 ; i < prim_count ; i++) { - if ((prim[i].mode & VBO_SAVE_PRIM_WEAK) && - (ctx->Driver.CurrentExecPrimitive != PRIM_OUTSIDE_BEGIN_END)) - { - loopback_weak_prim( ctx, &prim[i] ); - } - else - { - loopback_prim( ctx, buffer, &prim[i], wrap_count, vertex_size, la, nr ); - } - } -} - - -#endif /* FEATURE_dlist */ +/************************************************************************** + * + * Copyright 2005 Tungsten Graphics, Inc., Cedar Park, Texas. + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the + * "Software"), to deal in the Software without restriction, including + * without limitation the rights to use, copy, modify, merge, publish, + * distribute, sub license, and/or sell copies of the Software, and to + * permit persons to whom the Software is furnished to do so, subject to + * the following conditions: + * + * The above copyright notice and this permission notice (including the + * next paragraph) shall be included in all copies or substantial portions + * of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF + * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. + * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR + * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, + * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE + * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + **************************************************************************/ + +#include "main/context.h" +#include "main/glheader.h" +#include "main/enums.h" +#include "main/imports.h" +#include "main/mfeatures.h" +#include "main/mtypes.h" +#include "main/dispatch.h" +#include "glapi/glapi.h" + +#include "vbo_context.h" + + +#if FEATURE_dlist + + +typedef void (*attr_func)( struct gl_context *ctx, GLint target, const GLfloat * ); + + +/* This file makes heavy use of the aliasing of NV vertex attributes + * with the legacy attributes, and also with ARB and Material + * attributes as currently implemented. + */ +static void VertexAttrib1fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) +{ + CALL_VertexAttrib1fvNV(ctx->Exec, (target, v)); +} + +static void VertexAttrib2fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) +{ + CALL_VertexAttrib2fvNV(ctx->Exec, (target, v)); +} + +static void VertexAttrib3fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) +{ + CALL_VertexAttrib3fvNV(ctx->Exec, (target, v)); +} + +static void VertexAttrib4fvNV(struct gl_context *ctx, GLint target, const GLfloat *v) +{ + CALL_VertexAttrib4fvNV(ctx->Exec, (target, v)); +} + +static attr_func vert_attrfunc[4] = { + VertexAttrib1fvNV, + VertexAttrib2fvNV, + VertexAttrib3fvNV, + VertexAttrib4fvNV +}; + +struct loopback_attr { + GLint target; + GLint sz; + attr_func func; +}; + +/* Don't emit ends and begins on wrapped primitives. Don't replay + * wrapped vertices. If we get here, it's probably because the + * precalculated wrapping is wrong. + */ +static void loopback_prim( struct gl_context *ctx, + const GLfloat *buffer, + const struct _mesa_prim *prim, + GLuint wrap_count, + GLuint vertex_size, + const struct loopback_attr *la, GLuint nr ) +{ + GLint start = prim->start; + GLint end = start + prim->count; + const GLfloat *data; + GLint j; + GLuint k; + + if (0) + printf("loopback prim %s(%s,%s) verts %d..%d\n", + _mesa_lookup_prim_by_nr(prim->mode), + prim->begin ? "begin" : "..", + prim->end ? "end" : "..", + start, + end); + + if (prim->begin) { + CALL_Begin(GET_DISPATCH(), ( prim->mode )); + } + else { + assert(start == 0); + start += wrap_count; + } + + data = buffer + start * vertex_size; + + for (j = start ; j < end ; j++) { + const GLfloat *tmp = data + la[0].sz; + + for (k = 1 ; k < nr ; k++) { + la[k].func( ctx, la[k].target, tmp ); + tmp += la[k].sz; + } + + /* Fire the vertex + */ + la[0].func( ctx, VBO_ATTRIB_POS, data ); + data = tmp; + } + + if (prim->end) { + CALL_End(GET_DISPATCH(), ()); + } +} + +/* Primitives generated by DrawArrays/DrawElements/Rectf may be + * caught here. If there is no primitive in progress, execute them + * normally, otherwise need to track and discard the generated + * primitives. + */ +static void loopback_weak_prim( struct gl_context *ctx, + const struct _mesa_prim *prim ) +{ + /* Use the prim_weak flag to ensure that if this primitive + * wraps, we don't mistake future vertex_lists for part of the + * surrounding primitive. + * + * While this flag is set, we are simply disposing of data + * generated by an operation now known to be a noop. + */ + if (prim->begin) + ctx->Driver.CurrentExecPrimitive |= VBO_SAVE_PRIM_WEAK; + if (prim->end) + ctx->Driver.CurrentExecPrimitive &= ~VBO_SAVE_PRIM_WEAK; +} + + +void vbo_loopback_vertex_list( struct gl_context *ctx, + const GLfloat *buffer, + const GLubyte *attrsz, + const struct _mesa_prim *prim, + GLuint prim_count, + GLuint wrap_count, + GLuint vertex_size) +{ + struct loopback_attr la[VBO_ATTRIB_MAX]; + GLuint i, nr = 0; + + /* All Legacy, NV, ARB and Material attributes are routed through + * the NV attributes entrypoints: + */ + for (i = 0 ; i < VBO_ATTRIB_MAX ; i++) { + if (attrsz[i]) { + la[nr].target = i; + la[nr].sz = attrsz[i]; + la[nr].func = vert_attrfunc[attrsz[i]-1]; + nr++; + } + } + + for (i = 0 ; i < prim_count ; i++) { + if ((prim[i].mode & VBO_SAVE_PRIM_WEAK) && + (ctx->Driver.CurrentExecPrimitive != PRIM_OUTSIDE_BEGIN_END)) + { + loopback_weak_prim( ctx, &prim[i] ); + } + else + { + loopback_prim( ctx, buffer, &prim[i], wrap_count, vertex_size, la, nr ); + } + } +} + + +#endif /* FEATURE_dlist */ diff --git a/mesalib/src/mesa/vbo/vbo_split.h b/mesalib/src/mesa/vbo/vbo_split.h index 14cf9e49c..b7f0a9c57 100644 --- a/mesalib/src/mesa/vbo/vbo_split.h +++ b/mesalib/src/mesa/vbo/vbo_split.h @@ -1,72 +1,72 @@ -/* - * mesa 3-D graphics library - * Version: 6.5 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file vbo_context.h - * \brief VBO builder module datatypes and definitions. - * \author Keith Whitwell - */ - - -/** - * \mainpage The VBO splitter - * - * This is the private data used internally to the vbo_split_prims() - * helper function. Nobody outside the vbo_split* files needs to - * include or know about this structure. - */ - - -#ifndef _VBO_SPLIT_H -#define _VBO_SPLIT_H - -#include "vbo.h" - - -/* True if a primitive can be split without copying of vertices, false - * otherwise. - */ -GLboolean split_prim_inplace(GLenum mode, GLuint *first, GLuint *incr); - -void vbo_split_inplace( struct gl_context *ctx, - const struct gl_client_array *arrays[], - const struct _mesa_prim *prim, - GLuint nr_prims, - const struct _mesa_index_buffer *ib, - GLuint min_index, - GLuint max_index, - vbo_draw_func draw, - const struct split_limits *limits ); - -/* Requires ib != NULL: - */ -void vbo_split_copy( struct gl_context *ctx, - const struct gl_client_array *arrays[], - const struct _mesa_prim *prim, - GLuint nr_prims, - const struct _mesa_index_buffer *ib, - vbo_draw_func draw, - const struct split_limits *limits ); - -#endif +/* + * mesa 3-D graphics library + * Version: 6.5 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file vbo_context.h + * \brief VBO builder module datatypes and definitions. + * \author Keith Whitwell + */ + + +/** + * \mainpage The VBO splitter + * + * This is the private data used internally to the vbo_split_prims() + * helper function. Nobody outside the vbo_split* files needs to + * include or know about this structure. + */ + + +#ifndef _VBO_SPLIT_H +#define _VBO_SPLIT_H + +#include "vbo.h" + + +/* True if a primitive can be split without copying of vertices, false + * otherwise. + */ +GLboolean split_prim_inplace(GLenum mode, GLuint *first, GLuint *incr); + +void vbo_split_inplace( struct gl_context *ctx, + const struct gl_client_array *arrays[], + const struct _mesa_prim *prim, + GLuint nr_prims, + const struct _mesa_index_buffer *ib, + GLuint min_index, + GLuint max_index, + vbo_draw_func draw, + const struct split_limits *limits ); + +/* Requires ib != NULL: + */ +void vbo_split_copy( struct gl_context *ctx, + const struct gl_client_array *arrays[], + const struct _mesa_prim *prim, + GLuint nr_prims, + const struct _mesa_index_buffer *ib, + vbo_draw_func draw, + const struct split_limits *limits ); + +#endif diff --git a/mesalib/src/mesa/x86/3dnow.c b/mesalib/src/mesa/x86/3dnow.c index 549157c50..de2fb1e2a 100644 --- a/mesalib/src/mesa/x86/3dnow.c +++ b/mesalib/src/mesa/x86/3dnow.c @@ -1,91 +1,91 @@ - -/* - * Mesa 3-D graphics library - * Version: 5.0.1 - * - * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * 3DNow! optimizations contributed by - * Holger Waechtler - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "math/m_xform.h" -#include "tnl/t_context.h" - -#include "3dnow.h" -#include "x86_xform.h" - -#ifdef DEBUG_MATH -#include "math/m_debug.h" -#endif - - -#ifdef USE_3DNOW_ASM -DECLARE_XFORM_GROUP( 3dnow, 2 ) -DECLARE_XFORM_GROUP( 3dnow, 3 ) -DECLARE_XFORM_GROUP( 3dnow, 4 ) - -DECLARE_NORM_GROUP( 3dnow ) - - -extern void _ASMAPI -_mesa_v16_3dnow_general_xform( GLfloat *first_vert, - const GLfloat *m, - const GLfloat *src, - GLuint src_stride, - GLuint count ); - -extern void _ASMAPI -_mesa_3dnow_project_vertices( GLfloat *first, - GLfloat *last, - const GLfloat *m, - GLuint stride ); - -extern void _ASMAPI -_mesa_3dnow_project_clipped_vertices( GLfloat *first, - GLfloat *last, - const GLfloat *m, - GLuint stride, - const GLubyte *clipmask ); -#endif - - -void _mesa_init_3dnow_transform_asm( void ) -{ -#ifdef USE_3DNOW_ASM - ASSIGN_XFORM_GROUP( 3dnow, 2 ); - ASSIGN_XFORM_GROUP( 3dnow, 3 ); - ASSIGN_XFORM_GROUP( 3dnow, 4 ); - - /* There's a bug somewhere in the 3dnow_normal.S file that causes - * bad shading. Disable for now. - ASSIGN_NORM_GROUP( 3dnow ); - */ - -#ifdef DEBUG_MATH - _math_test_all_transform_functions( "3DNow!" ); - _math_test_all_normal_transform_functions( "3DNow!" ); -#endif -#endif -} + +/* + * Mesa 3-D graphics library + * Version: 5.0.1 + * + * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * 3DNow! optimizations contributed by + * Holger Waechtler + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "math/m_xform.h" +#include "tnl/t_context.h" + +#include "3dnow.h" +#include "x86_xform.h" + +#ifdef DEBUG_MATH +#include "math/m_debug.h" +#endif + + +#ifdef USE_3DNOW_ASM +DECLARE_XFORM_GROUP( 3dnow, 2 ) +DECLARE_XFORM_GROUP( 3dnow, 3 ) +DECLARE_XFORM_GROUP( 3dnow, 4 ) + +DECLARE_NORM_GROUP( 3dnow ) + + +extern void _ASMAPI +_mesa_v16_3dnow_general_xform( GLfloat *first_vert, + const GLfloat *m, + const GLfloat *src, + GLuint src_stride, + GLuint count ); + +extern void _ASMAPI +_mesa_3dnow_project_vertices( GLfloat *first, + GLfloat *last, + const GLfloat *m, + GLuint stride ); + +extern void _ASMAPI +_mesa_3dnow_project_clipped_vertices( GLfloat *first, + GLfloat *last, + const GLfloat *m, + GLuint stride, + const GLubyte *clipmask ); +#endif + + +void _mesa_init_3dnow_transform_asm( void ) +{ +#ifdef USE_3DNOW_ASM + ASSIGN_XFORM_GROUP( 3dnow, 2 ); + ASSIGN_XFORM_GROUP( 3dnow, 3 ); + ASSIGN_XFORM_GROUP( 3dnow, 4 ); + + /* There's a bug somewhere in the 3dnow_normal.S file that causes + * bad shading. Disable for now. + ASSIGN_NORM_GROUP( 3dnow ); + */ + +#ifdef DEBUG_MATH + _math_test_all_transform_functions( "3DNow!" ); + _math_test_all_normal_transform_functions( "3DNow!" ); +#endif +#endif +} diff --git a/mesalib/src/mesa/x86/3dnow.h b/mesalib/src/mesa/x86/3dnow.h index 067f8f721..1c1fedcd4 100644 --- a/mesalib/src/mesa/x86/3dnow.h +++ b/mesalib/src/mesa/x86/3dnow.h @@ -1,36 +1,36 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * 3DNow! optimizations contributed by - * Holger Waechtler - */ - -#ifndef __3DNOW_H__ -#define __3DNOW_H__ - -void _mesa_init_3dnow_transform_asm( void ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * 3DNow! optimizations contributed by + * Holger Waechtler + */ + +#ifndef __3DNOW_H__ +#define __3DNOW_H__ + +void _mesa_init_3dnow_transform_asm( void ); + +#endif diff --git a/mesalib/src/mesa/x86/3dnow_normal.S b/mesalib/src/mesa/x86/3dnow_normal.S index bfa316588..7f5f6b357 100644 --- a/mesalib/src/mesa/x86/3dnow_normal.S +++ b/mesalib/src/mesa/x86/3dnow_normal.S @@ -1,852 +1,852 @@ - -/* - * Mesa 3-D graphics library - * Version: 5.1 - * - * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * 3Dnow assembly code by Holger Waechtler - */ - -#ifdef USE_3DNOW_ASM - -#include "assyntax.h" -#include "matypes.h" -#include "norm_args.h" - - SEG_TEXT - -#define M(i) REGOFF(i * 4, ECX) -#define STRIDE REGOFF(12, ESI) - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_transform_normalize_normals) -HIDDEN(_mesa_3dnow_transform_normalize_normals) -GLNAME(_mesa_3dnow_transform_normalize_normals): - -#define FRAME_OFFSET 12 - - PUSH_L ( EDI ) - PUSH_L ( ESI ) - PUSH_L ( EBP ) - - MOV_L ( ARG_LENGTHS, EDI ) - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_DEST, EAX ) - MOV_L ( REGOFF(V4F_COUNT, ESI), EBP ) /* dest->count = in->count */ - MOV_L ( EBP, REGOFF(V4F_COUNT, EAX) ) - MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ - MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ - MOV_L ( ARG_MAT, ECX ) - MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ - - CMP_L ( CONST(0), EBP ) /* count > 0 ?? */ - JE ( LLBL (G3TN_end) ) - - MOV_L ( REGOFF (V4F_COUNT, ESI), EBP ) - FEMMS - - PUSH_L ( EBP ) - PUSH_L ( EAX ) - PUSH_L ( EDX ) /* save counter & pointer for */ - /* the normalize pass */ -#undef FRAME_OFFSET -#define FRAME_OFFSET 24 - - MOVQ ( M(0), MM3 ) /* m1 | m0 */ - MOVQ ( M(4), MM4 ) /* m5 | m4 */ - - MOVD ( M(2), MM5 ) /* | m2 */ - PUNPCKLDQ ( M(6), MM5 ) /* m6 | m2 */ - - MOVQ ( M(8), MM6 ) /* m9 | m8 */ - MOVQ ( M(10), MM7 ) /* | m10 */ - - CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ - JNE ( LLBL (G3TN_scale_end ) ) - - MOVD ( ARG_SCALE, MM0 ) /* | scale */ - PUNPCKLDQ ( MM0, MM0 ) /* scale | scale */ - - PFMUL ( MM0, MM3 ) /* scale * m1 | scale * m0 */ - PFMUL ( MM0, MM4 ) /* scale * m5 | scale * m4 */ - PFMUL ( MM0, MM5 ) /* scale * m6 | scale * m2 */ - PFMUL ( MM0, MM6 ) /* scale * m9 | scale * m8 */ - PFMUL ( MM0, MM7 ) /* | scale * m10 */ - -ALIGNTEXT32 -LLBL (G3TN_scale_end): -LLBL (G3TN_transform): - MOVQ ( REGIND (EDX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF (8, EDX), MM2 ) /* | x2 */ - - MOVQ ( MM0, MM1 ) /* x1 | x0 */ - PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ - - PFMUL ( MM3, MM0 ) /* x1*m1 | x0*m0 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - PREFETCHW ( REGIND(EAX) ) - - PFMUL ( MM4, MM1 ) /* x1*m5 | x0*m4 */ - PFACC ( MM1, MM0 ) /* x0*m4+x1*m5 | x0*m0+x1*m1 */ - - PFMUL ( MM5, MM2 ) /* x2*m6 | x2*m2 */ - PFADD ( MM2, MM0 ) /* x0*m4+x1*m5+x2*m6| x0*m0+...+x2**/ - - MOVQ ( REGIND (EDX), MM1 ) /* x1 | x0 */ - MOVQ ( MM0, REGOFF(-16, EAX) ) /* write r0, r1 */ - - PFMUL ( MM6, MM1 ) /* x1*m9 | x0*m8 */ - MOVD ( REGOFF (8, EDX), MM2 ) /* | x2 */ - - PFMUL ( MM7, MM2 ) /* | x2*m10 */ - PFACC ( MM1, MM1 ) /* *not used* | x0*m8+x1*m9 */ - - PFADD ( MM2, MM1 ) /* *not used* | x0*m8+x1*m9+x2*m*/ - ADD_L ( STRIDE, EDX ) /* next normal */ - - PREFETCH ( REGIND(EDX) ) - - MOVD ( MM1, REGOFF(-8, EAX) ) /* write r2 */ - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - JNZ ( LLBL (G3TN_transform) ) - - - POP_L ( EDX ) /* end of transform --- */ - POP_L ( EAX ) /* now normalizing ... */ - POP_L ( EBP ) - - CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ - JE ( LLBL (G3TN_norm ) ) /* calculate lengths */ - - -ALIGNTEXT32 -LLBL (G3TN_norm_w_lengths): - - PREFETCHW ( REGOFF(12,EAX) ) - - MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ - - MOVD ( REGIND (EDI), MM3 ) /* | length (x) */ - PFMUL ( MM3, MM1 ) /* | x2 (normalize*/ - - PUNPCKLDQ ( MM3, MM3 ) /* length (x) | length (x) */ - PFMUL ( MM3, MM0 ) /* x1 (normalized) | x0 (normalize*/ - - ADD_L ( STRIDE, EDX ) /* next normal */ - ADD_L ( CONST(4), EDI ) /* next length */ - - PREFETCH ( REGIND(EDI) ) - - MOVQ ( MM0, REGIND(EAX) ) /* write new x0, x1 */ - MOVD ( MM1, REGOFF(8, EAX) ) /* write new x2 */ - - ADD_L ( CONST(16), EAX ) /* next r */ - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - - JNZ ( LLBL (G3TN_norm_w_lengths) ) - JMP ( LLBL (G3TN_exit_3dnow) ) - -ALIGNTEXT32 -LLBL (G3TN_norm): - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND (EAX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ - - MOVQ ( MM0, MM3 ) /* x1 | x0 */ - MOVQ ( MM1, MM4 ) /* | x2 */ - - PFMUL ( MM0, MM3 ) /* x1*x1 | x0*x0 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - PFMUL ( MM1, MM4 ) /* | x2*x2 */ - PFADD ( MM4, MM3 ) /* | x0*x0+x2*x2 */ - - PFACC ( MM3, MM3 ) /* **not used** | x0*x0+x1*x1+x2**/ - PFRSQRT ( MM3, MM5 ) /* 1/sqrt (x0*x0+x1*x1+x2*x2) */ - - MOVQ ( MM5, MM4 ) - PUNPCKLDQ ( MM3, MM3 ) - - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - PFMUL ( MM5, MM5 ) - - PFRSQIT1 ( MM3, MM5 ) - PFRCPIT2 ( MM4, MM5 ) - - PFMUL ( MM5, MM0 ) /* x1 (normalized) | x0 (normalize*/ - - MOVQ ( MM0, REGOFF(-16, EAX) ) /* write new x0, x1 */ - PFMUL ( MM5, MM1 ) /* | x2 (normalize*/ - - MOVD ( MM1, REGOFF(-8, EAX) ) /* write new x2 */ - JNZ ( LLBL (G3TN_norm) ) - -LLBL (G3TN_exit_3dnow): - FEMMS - -LLBL (G3TN_end): - POP_L ( EBP ) - POP_L ( ESI ) - POP_L ( EDI ) - RET - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_transform_normalize_normals_no_rot) -HIDDEN(_mesa_3dnow_transform_normalize_normals_no_rot) -GLNAME(_mesa_3dnow_transform_normalize_normals_no_rot): - -#undef FRAME_OFFSET -#define FRAME_OFFSET 12 - - PUSH_L ( EDI ) - PUSH_L ( ESI ) - PUSH_L ( EBP ) - - MOV_L ( ARG_LENGTHS, EDI ) - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_DEST, EAX ) - MOV_L ( REGOFF(V4F_COUNT, ESI), EBP ) /* dest->count = in->count */ - MOV_L ( EBP, REGOFF(V4F_COUNT, EAX) ) - MOV_L ( ARG_MAT, ECX ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ - MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ - MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ - - CMP_L ( CONST(0), EBP ) /* count > 0 ?? */ - JE ( LLBL (G3TNNR_end) ) - - FEMMS - - MOVD ( M(0), MM0 ) /* | m0 */ - PUNPCKLDQ ( M(5), MM0 ) /* m5 | m0 */ - - MOVD ( M(10), MM2 ) /* | m10 */ - PUNPCKLDQ ( MM2, MM2 ) /* m10 | m10 */ - - CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ - JNE ( LLBL (G3TNNR_scale_end ) ) - - MOVD ( ARG_SCALE, MM7 ) /* | scale */ - PUNPCKLDQ ( MM7, MM7 ) /* scale | scale */ - - PFMUL ( MM7, MM0 ) /* scale * m5 | scale * m0 */ - PFMUL ( MM7, MM2 ) /* scale * m10 | scale * m10 */ - -ALIGNTEXT32 -LLBL (G3TNNR_scale_end): - CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ - JE ( LLBL (G3TNNR_norm) ) /* need to calculate lengths */ - - MOVD ( REGIND(EDI), MM3 ) /* | length (x) */ - - -ALIGNTEXT32 -LLBL (G3TNNR_norm_w_lengths): /* use precalculated lengths */ - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND(EDX), MM6 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EDX), MM7 ) /* | x2 */ - - PFMUL ( MM0, MM6 ) /* x1*m5 | x0*m0 */ - ADD_L ( STRIDE, EDX ) /* next normal */ - - PREFETCH ( REGIND(EDX) ) - - PFMUL ( MM2, MM7 ) /* | x2*m10 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - PFMUL ( MM3, MM7 ) /* | x2 (normalized) */ - PUNPCKLDQ ( MM3, MM3 ) /* length (x) | length (x) */ - - ADD_L ( CONST(4), EDI ) /* next length */ - PFMUL ( MM3, MM6 ) /* x1 (normalized) | x0 (normalized) */ - - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - MOVQ ( MM6, REGOFF(-16, EAX) ) /* write r0, r1 */ - - MOVD ( MM7, REGOFF(-8, EAX) ) /* write r2 */ - MOVD ( REGIND(EDI), MM3 ) /* | length (x) */ - - JNZ ( LLBL (G3TNNR_norm_w_lengths) ) - JMP ( LLBL (G3TNNR_exit_3dnow) ) - -ALIGNTEXT32 -LLBL (G3TNNR_norm): /* need to calculate lengths */ - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND(EDX), MM6 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EDX), MM7 ) /* | x2 */ - - PFMUL ( MM0, MM6 ) /* x1*m5 | x0*m0 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - PFMUL ( MM2, MM7 ) /* | x2*m10 */ - MOVQ ( MM6, MM3 ) /* x1 (transformed)| x0 (transformed) */ - - MOVQ ( MM7, MM4 ) /* | x2 (transformed) */ - PFMUL ( MM6, MM3 ) /* x1*x1 | x0*x0 */ - - - PFMUL ( MM7, MM4 ) /* | x2*x2 */ - PFACC ( MM3, MM3 ) /* **not used** | x0*x0+x1*x1 */ - - PFADD ( MM4, MM3 ) /* | x0*x0+x1*x1+x2*x2*/ - ADD_L ( STRIDE, EDX ) /* next normal */ - - PREFETCH ( REGIND(EDX) ) - - PFRSQRT ( MM3, MM5 ) /* 1/sqrt (x0*x0+x1*x1+x2*x2) */ - MOVQ ( MM5, MM4 ) - - PUNPCKLDQ ( MM3, MM3 ) - PFMUL ( MM5, MM5 ) - - PFRSQIT1 ( MM3, MM5 ) - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - - PFRCPIT2 ( MM4, MM5 ) - PFMUL ( MM5, MM6 ) /* x1 (normalized) | x0 (normalized) */ - - MOVQ ( MM6, REGOFF(-16, EAX) ) /* write r0, r1 */ - PFMUL ( MM5, MM7 ) /* | x2 (normalized) */ - - MOVD ( MM7, REGOFF(-8, EAX) ) /* write r2 */ - JNZ ( LLBL (G3TNNR_norm) ) - - -LLBL (G3TNNR_exit_3dnow): - FEMMS - -LLBL (G3TNNR_end): - POP_L ( EBP ) - POP_L ( ESI ) - POP_L ( EDI ) - RET - - - - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_transform_rescale_normals_no_rot) -HIDDEN(_mesa_3dnow_transform_rescale_normals_no_rot) -GLNAME(_mesa_3dnow_transform_rescale_normals_no_rot): - -#undef FRAME_OFFSET -#define FRAME_OFFSET 12 - - PUSH_L ( EDI ) - PUSH_L ( ESI ) - PUSH_L ( EBP ) - - MOV_L ( ARG_IN, EAX ) - MOV_L ( ARG_DEST, EDX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EBP ) /* dest->count = in->count */ - MOV_L ( EBP, REGOFF(V4F_COUNT, EDX) ) - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_MAT, ECX ) - MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ - MOV_L ( REGOFF(V4F_START, EDX), EAX ) /* dest->start */ - MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ - - CMP_L ( CONST(0), EBP ) - JE ( LLBL (G3TRNR_end) ) - - FEMMS - - MOVD ( ARG_SCALE, MM6 ) /* | scale */ - PUNPCKLDQ ( MM6, MM6 ) /* scale | scale */ - - MOVD ( REGIND(ECX), MM0 ) /* | m0 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m5 | m0 */ - - PFMUL ( MM6, MM0 ) /* scale*m5 | scale*m0 */ - MOVD ( REGOFF(40, ECX), MM2 ) /* | m10 */ - - PFMUL ( MM6, MM2 ) /* | scale*m10 */ - -ALIGNTEXT32 -LLBL (G3TRNR_rescale): - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND(EDX), MM4 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EDX), MM5 ) /* | x2 */ - - PFMUL ( MM0, MM4 ) /* x1*m5 | x0*m0 */ - ADD_L ( STRIDE, EDX ) /* next normal */ - - PREFETCH ( REGIND(EDX) ) - - PFMUL ( MM2, MM5 ) /* | x2*m10 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - MOVQ ( MM4, REGOFF(-16, EAX) ) /* write r0, r1 */ - - MOVD ( MM5, REGOFF(-8, EAX) ) /* write r2 */ - JNZ ( LLBL (G3TRNR_rescale) ) /* cnt > 0 ? -> process next normal */ - - FEMMS - -LLBL (G3TRNR_end): - POP_L ( EBP ) - POP_L ( ESI ) - POP_L ( EDI ) - RET - - - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_transform_rescale_normals) -HIDDEN(_mesa_3dnow_transform_rescale_normals) -GLNAME(_mesa_3dnow_transform_rescale_normals): - -#undef FRAME_OFFSET -#define FRAME_OFFSET 8 - - PUSH_L ( EDI ) - PUSH_L ( ESI ) - - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_DEST, EAX ) - MOV_L ( ARG_MAT, ECX ) - MOV_L ( REGOFF(V4F_COUNT, ESI), EDI ) /* dest->count = in->count */ - MOV_L ( EDI, REGOFF(V4F_COUNT, EAX) ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ - MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ - MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ - - CMP_L ( CONST(0), EDI ) - JE ( LLBL (G3TR_end) ) - - FEMMS - - MOVQ ( REGIND(ECX), MM3 ) /* m1 | m0 */ - - MOVQ ( REGOFF(16,ECX), MM4 ) /* m5 | m4 */ - MOVD ( ARG_SCALE, MM0 ) /* scale */ - - MOVD ( REGOFF(8,ECX), MM5 ) /* | m2 */ - PUNPCKLDQ ( MM0, MM0 ) /* scale | scale */ - - PUNPCKLDQ ( REGOFF(24, ECX), MM5 ) - PFMUL ( MM0, MM3 ) /* scale*m1 | scale*m0 */ - - MOVQ ( REGOFF(32, ECX), MM6 ) /* m9 | m8*/ - PFMUL ( MM0, MM4 ) /* scale*m5 | scale*m4 */ - - MOVD ( REGOFF(40, ECX), MM7 ) /* | m10 */ - PFMUL ( MM0, MM5 ) /* scale*m6 | scale*m2 */ - - PFMUL ( MM0, MM6 ) /* scale*m9 | scale*m8 */ - - PFMUL ( MM0, MM7 ) /* | scale*m10 */ - -ALIGNTEXT32 -LLBL (G3TR_rescale): - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND(EDX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ - - MOVQ ( MM0, MM1 ) /* x1 | x0 */ - PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ - - PFMUL ( MM3, MM0 ) /* x1*m1 | x0*m0 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - PFMUL ( MM4, MM1 ) /* x1*m5 | x0*m4 */ - PFACC ( MM1, MM0 ) /* x0*m4+x1*m5 | x0*m0+x1*m1 */ - - MOVQ ( REGIND(EDX), MM1 ) /* x1 | x0 */ - - PFMUL ( MM5, MM2 ) /* x2*m6 | x2*m2 */ - PFADD ( MM2, MM0 ) /* x0*m4...+x2*m6| x0*m0+x1*m1+x2*m2 */ - - MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ - ADD_L ( STRIDE, EDX ) /* next normal */ - - PREFETCH ( REGIND(EDX) ) - - MOVQ ( MM0, REGOFF(-16, EAX) ) /* write r0, r1 */ - PFMUL ( MM6, MM1 ) /* x1*m9 | x0*m8 */ - - PFMUL ( MM7, MM2 ) /* | x2*m10 */ - PFACC ( MM1, MM1 ) /* *not used* | x0*m8+x1*m9 */ - - PFADD ( MM2, MM1 ) /* *not used* | x0*m8+x1*m9+x2*m10 */ - MOVD ( MM1, REGOFF(-8, EAX) ) /* write r2 */ - - SUB_L ( CONST(1), EDI ) /* decrement normal counter */ - JNZ ( LLBL (G3TR_rescale) ) - - FEMMS - -LLBL (G3TR_end): - POP_L ( ESI ) - POP_L ( EDI ) - RET - - - - - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_transform_normals_no_rot) -HIDDEN(_mesa_3dnow_transform_normals_no_rot) -GLNAME(_mesa_3dnow_transform_normals_no_rot): - -#undef FRAME_OFFSET -#define FRAME_OFFSET 8 - - PUSH_L ( EDI ) - PUSH_L ( ESI ) - - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_DEST, EAX ) - MOV_L ( ARG_MAT, ECX ) - MOV_L ( REGOFF(V4F_COUNT, ESI), EDI ) /* dest->count = in->count */ - MOV_L ( EDI, REGOFF(V4F_COUNT, EAX) ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ - MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ - MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ - - CMP_L ( CONST(0), EDI ) - JE ( LLBL (G3TNR_end) ) - - FEMMS - - MOVD ( REGIND(ECX), MM0 ) /* | m0 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m5 | m0 */ - - MOVD ( REGOFF(40, ECX), MM2 ) /* | m10 */ - PUNPCKLDQ ( MM2, MM2 ) /* m10 | m10 */ - -ALIGNTEXT32 -LLBL (G3TNR_transform): - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND(EDX), MM4 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EDX), MM5 ) /* | x2 */ - - PFMUL ( MM0, MM4 ) /* x1*m5 | x0*m0 */ - ADD_L ( STRIDE, EDX) /* next normal */ - - PREFETCH ( REGIND(EDX) ) - - PFMUL ( MM2, MM5 ) /* | x2*m10 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - SUB_L ( CONST(1), EDI ) /* decrement normal counter */ - MOVQ ( MM4, REGOFF(-16, EAX) ) /* write r0, r1 */ - - MOVD ( MM5, REGOFF(-8, EAX) ) /* write r2 */ - JNZ ( LLBL (G3TNR_transform) ) - - FEMMS - -LLBL (G3TNR_end): - POP_L ( ESI ) - POP_L ( EDI ) - RET - - - - - - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_transform_normals) -HIDDEN(_mesa_3dnow_transform_normals) -GLNAME(_mesa_3dnow_transform_normals): - -#undef FRAME_OFFSET -#define FRAME_OFFSET 8 - - PUSH_L ( EDI ) - PUSH_L ( ESI ) - - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_DEST, EAX ) - MOV_L ( ARG_MAT, ECX ) - MOV_L ( REGOFF(V4F_COUNT, ESI), EDI ) /* dest->count = in->count */ - MOV_L ( EDI, REGOFF(V4F_COUNT, EAX) ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ - MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ - MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ - - CMP_L ( CONST(0), EDI ) /* count > 0 ?? */ - JE ( LLBL (G3T_end) ) - - FEMMS - - MOVQ ( REGIND(ECX), MM3 ) /* m1 | m0 */ - MOVQ ( REGOFF(16, ECX), MM4 ) /* m5 | m4 */ - - MOVD ( REGOFF(8, ECX), MM5 ) /* | m2 */ - PUNPCKLDQ ( REGOFF(24, ECX), MM5 ) /* m6 | m2 */ - - MOVQ ( REGOFF(32, ECX), MM6 ) /* m9 | m8 */ - MOVD ( REGOFF(40, ECX), MM7 ) /* | m10 */ - -ALIGNTEXT32 -LLBL (G3T_transform): - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND(EDX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ - - MOVQ ( MM0, MM1 ) /* x1 | x0 */ - PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ - - PFMUL ( MM3, MM0 ) /* x1*m1 | x0*m0 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - PFMUL ( MM4, MM1 ) /* x1*m5 | x0*m4 */ - PFACC ( MM1, MM0 ) /* x0*m4+x1*m5 | x0*m0+x1*m1 */ - - PFMUL ( MM5, MM2 ) /* x2*m6 | x2*m2 */ - PFADD ( MM2, MM0 ) /* x0*m4...+x2*m6| x0*m0+x1*m1+x2*m2 */ - - MOVQ ( REGIND(EDX), MM1 ) /* x1 | x0 */ - MOVQ ( MM0, REGOFF(-16, EAX) ) /* write r0, r1 */ - - PFMUL ( MM6, MM1 ) /* x1*m9 | x0*m8 */ - MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ - - PFMUL ( MM7, MM2 ) /* | x2*m10 */ - ADD_L ( STRIDE, EDX ) /* next normal */ - - PREFETCH ( REGIND(EDX) ) - - PFACC ( MM1, MM1 ) /* *not used* | x0*m8+x1*m9 */ - PFADD ( MM2, MM1 ) /* *not used* | x0*m8+x1*m9+x2*m10 */ - - MOVD ( MM1, REGOFF(-8, EAX) ) /* write r2 */ - SUB_L ( CONST(1), EDI ) /* decrement normal counter */ - - JNZ ( LLBL (G3T_transform) ) - - FEMMS - -LLBL (G3T_end): - POP_L ( ESI ) - POP_L ( EDI ) - RET - - - - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_normalize_normals) -HIDDEN(_mesa_3dnow_normalize_normals) -GLNAME(_mesa_3dnow_normalize_normals): - -#undef FRAME_OFFSET -#define FRAME_OFFSET 12 - - PUSH_L ( EDI ) - PUSH_L ( ESI ) - PUSH_L ( EBP ) - - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_DEST, EAX ) - MOV_L ( REGOFF(V4F_COUNT, ESI), EBP ) /* dest->count = in->count */ - MOV_L ( EBP, REGOFF(V4F_COUNT, EAX) ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ - MOV_L ( REGOFF(V4F_START, ESI), ECX ) /* in->start */ - MOV_L ( ARG_LENGTHS, EDX ) - - CMP_L ( CONST(0), EBP ) /* count > 0 ?? */ - JE ( LLBL (G3N_end) ) - - FEMMS - - CMP_L ( CONST(0), EDX ) /* lengths == 0 ? */ - JE ( LLBL (G3N_norm2) ) /* calculate lengths */ - -ALIGNTEXT32 -LLBL (G3N_norm1): /* use precalculated lengths */ - - PREFETCH ( REGIND(EAX) ) - - MOVQ ( REGIND(ECX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, ECX), MM1 ) /* | x2 */ - - MOVD ( REGIND(EDX), MM3 ) /* | length (x) */ - PFMUL ( MM3, MM1 ) /* | x2 (normalized) */ - - PUNPCKLDQ ( MM3, MM3 ) /* length (x) | length (x) */ - ADD_L ( STRIDE, ECX ) /* next normal */ - - PREFETCH ( REGIND(ECX) ) - - PFMUL ( MM3, MM0 ) /* x1 (normalized) | x0 (normalized) */ - MOVQ ( MM0, REGIND(EAX) ) /* write new x0, x1 */ - - MOVD ( MM1, REGOFF(8, EAX) ) /* write new x2 */ - ADD_L ( CONST(16), EAX ) /* next r */ - - ADD_L ( CONST(4), EDX ) /* next length */ - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - - JNZ ( LLBL (G3N_norm1) ) - - JMP ( LLBL (G3N_end1) ) - -ALIGNTEXT32 -LLBL (G3N_norm2): /* need to calculate lengths */ - - PREFETCHW ( REGIND(EAX) ) - - PREFETCH ( REGIND(ECX) ) - - MOVQ ( REGIND(ECX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, ECX), MM1 ) /* | x2 */ - - MOVQ ( MM0, MM3 ) /* x1 | x0 */ - ADD_L ( STRIDE, ECX ) /* next normal */ - - PFMUL ( MM0, MM3 ) /* x1*x1 | x0*x0 */ - MOVQ ( MM1, MM4 ) /* | x2 */ - - ADD_L ( CONST(16), EAX ) /* next r */ - PFMUL ( MM1, MM4 ) /* | x2*x2 */ - - PFADD ( MM4, MM3 ) /* | x0*x0+x2*x2 */ - PFACC ( MM3, MM3 ) /* x0*x0+...+x2*x2 | x0*x0+x1*x1+x2*x2*/ - - PFRSQRT ( MM3, MM5 ) /* 1/sqrt (x0*x0+x1*x1+x2*x2) */ - MOVQ ( MM5, MM4 ) - - PUNPCKLDQ ( MM3, MM3 ) - PFMUL ( MM5, MM5 ) - - PFRSQIT1 ( MM3, MM5 ) - SUB_L ( CONST(1), EBP ) /* decrement normal counter */ - - PFRCPIT2 ( MM4, MM5 ) - - PFMUL ( MM5, MM0 ) /* x1 (normalized) | x0 (normalized) */ - MOVQ ( MM0, REGOFF(-16, EAX) ) /* write new x0, x1 */ - - PFMUL ( MM5, MM1 ) /* | x2 (normalized) */ - MOVD ( MM1, REGOFF(-8, EAX) ) /* write new x2 */ - - JNZ ( LLBL (G3N_norm2) ) - -LLBL (G3N_end1): - FEMMS - -LLBL (G3N_end): - POP_L ( EBP ) - POP_L ( ESI ) - POP_L ( EDI ) - RET - - - - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_3dnow_rescale_normals) -HIDDEN(_mesa_3dnow_rescale_normals) -GLNAME(_mesa_3dnow_rescale_normals): - -#undef FRAME_OFFSET -#define FRAME_OFFSET 8 - PUSH_L ( EDI ) - PUSH_L ( ESI ) - - MOV_L ( ARG_IN, ESI ) - MOV_L ( ARG_DEST, EAX ) - MOV_L ( REGOFF(V4F_COUNT, ESI), EDX ) /* dest->count = in->count */ - MOV_L ( EDX, REGOFF(V4F_COUNT, EAX) ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ - MOV_L ( REGOFF(V4F_START, ESI), ECX ) /* in->start */ - - CMP_L ( CONST(0), EDX ) - JE ( LLBL (G3R_end) ) - - FEMMS - - MOVD ( ARG_SCALE, MM0 ) /* scale */ - PUNPCKLDQ ( MM0, MM0 ) - -ALIGNTEXT32 -LLBL (G3R_rescale): - - PREFETCHW ( REGIND(EAX) ) - - MOVQ ( REGIND(ECX), MM1 ) /* x1 | x0 */ - MOVD ( REGOFF(8, ECX), MM2 ) /* | x2 */ - - PFMUL ( MM0, MM1 ) /* x1*scale | x0*scale */ - ADD_L ( STRIDE, ECX ) /* next normal */ - - PREFETCH ( REGIND(ECX) ) - - PFMUL ( MM0, MM2 ) /* | x2*scale */ - ADD_L ( CONST(16), EAX ) /* next r */ - - MOVQ ( MM1, REGOFF(-16, EAX) ) /* write r0, r1 */ - MOVD ( MM2, REGOFF(-8, EAX) ) /* write r2 */ - - SUB_L ( CONST(1), EDX ) /* decrement normal counter */ - JNZ ( LLBL (G3R_rescale) ) - - FEMMS - -LLBL (G3R_end): - POP_L ( ESI ) - POP_L ( EDI ) - RET - -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 5.1 + * + * Copyright (C) 1999-2003 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * 3Dnow assembly code by Holger Waechtler + */ + +#ifdef USE_3DNOW_ASM + +#include "assyntax.h" +#include "matypes.h" +#include "norm_args.h" + + SEG_TEXT + +#define M(i) REGOFF(i * 4, ECX) +#define STRIDE REGOFF(12, ESI) + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_transform_normalize_normals) +HIDDEN(_mesa_3dnow_transform_normalize_normals) +GLNAME(_mesa_3dnow_transform_normalize_normals): + +#define FRAME_OFFSET 12 + + PUSH_L ( EDI ) + PUSH_L ( ESI ) + PUSH_L ( EBP ) + + MOV_L ( ARG_LENGTHS, EDI ) + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_DEST, EAX ) + MOV_L ( REGOFF(V4F_COUNT, ESI), EBP ) /* dest->count = in->count */ + MOV_L ( EBP, REGOFF(V4F_COUNT, EAX) ) + MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ + MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ + MOV_L ( ARG_MAT, ECX ) + MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ + + CMP_L ( CONST(0), EBP ) /* count > 0 ?? */ + JE ( LLBL (G3TN_end) ) + + MOV_L ( REGOFF (V4F_COUNT, ESI), EBP ) + FEMMS + + PUSH_L ( EBP ) + PUSH_L ( EAX ) + PUSH_L ( EDX ) /* save counter & pointer for */ + /* the normalize pass */ +#undef FRAME_OFFSET +#define FRAME_OFFSET 24 + + MOVQ ( M(0), MM3 ) /* m1 | m0 */ + MOVQ ( M(4), MM4 ) /* m5 | m4 */ + + MOVD ( M(2), MM5 ) /* | m2 */ + PUNPCKLDQ ( M(6), MM5 ) /* m6 | m2 */ + + MOVQ ( M(8), MM6 ) /* m9 | m8 */ + MOVQ ( M(10), MM7 ) /* | m10 */ + + CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ + JNE ( LLBL (G3TN_scale_end ) ) + + MOVD ( ARG_SCALE, MM0 ) /* | scale */ + PUNPCKLDQ ( MM0, MM0 ) /* scale | scale */ + + PFMUL ( MM0, MM3 ) /* scale * m1 | scale * m0 */ + PFMUL ( MM0, MM4 ) /* scale * m5 | scale * m4 */ + PFMUL ( MM0, MM5 ) /* scale * m6 | scale * m2 */ + PFMUL ( MM0, MM6 ) /* scale * m9 | scale * m8 */ + PFMUL ( MM0, MM7 ) /* | scale * m10 */ + +ALIGNTEXT32 +LLBL (G3TN_scale_end): +LLBL (G3TN_transform): + MOVQ ( REGIND (EDX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF (8, EDX), MM2 ) /* | x2 */ + + MOVQ ( MM0, MM1 ) /* x1 | x0 */ + PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ + + PFMUL ( MM3, MM0 ) /* x1*m1 | x0*m0 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + PREFETCHW ( REGIND(EAX) ) + + PFMUL ( MM4, MM1 ) /* x1*m5 | x0*m4 */ + PFACC ( MM1, MM0 ) /* x0*m4+x1*m5 | x0*m0+x1*m1 */ + + PFMUL ( MM5, MM2 ) /* x2*m6 | x2*m2 */ + PFADD ( MM2, MM0 ) /* x0*m4+x1*m5+x2*m6| x0*m0+...+x2**/ + + MOVQ ( REGIND (EDX), MM1 ) /* x1 | x0 */ + MOVQ ( MM0, REGOFF(-16, EAX) ) /* write r0, r1 */ + + PFMUL ( MM6, MM1 ) /* x1*m9 | x0*m8 */ + MOVD ( REGOFF (8, EDX), MM2 ) /* | x2 */ + + PFMUL ( MM7, MM2 ) /* | x2*m10 */ + PFACC ( MM1, MM1 ) /* *not used* | x0*m8+x1*m9 */ + + PFADD ( MM2, MM1 ) /* *not used* | x0*m8+x1*m9+x2*m*/ + ADD_L ( STRIDE, EDX ) /* next normal */ + + PREFETCH ( REGIND(EDX) ) + + MOVD ( MM1, REGOFF(-8, EAX) ) /* write r2 */ + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + JNZ ( LLBL (G3TN_transform) ) + + + POP_L ( EDX ) /* end of transform --- */ + POP_L ( EAX ) /* now normalizing ... */ + POP_L ( EBP ) + + CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ + JE ( LLBL (G3TN_norm ) ) /* calculate lengths */ + + +ALIGNTEXT32 +LLBL (G3TN_norm_w_lengths): + + PREFETCHW ( REGOFF(12,EAX) ) + + MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ + + MOVD ( REGIND (EDI), MM3 ) /* | length (x) */ + PFMUL ( MM3, MM1 ) /* | x2 (normalize*/ + + PUNPCKLDQ ( MM3, MM3 ) /* length (x) | length (x) */ + PFMUL ( MM3, MM0 ) /* x1 (normalized) | x0 (normalize*/ + + ADD_L ( STRIDE, EDX ) /* next normal */ + ADD_L ( CONST(4), EDI ) /* next length */ + + PREFETCH ( REGIND(EDI) ) + + MOVQ ( MM0, REGIND(EAX) ) /* write new x0, x1 */ + MOVD ( MM1, REGOFF(8, EAX) ) /* write new x2 */ + + ADD_L ( CONST(16), EAX ) /* next r */ + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + + JNZ ( LLBL (G3TN_norm_w_lengths) ) + JMP ( LLBL (G3TN_exit_3dnow) ) + +ALIGNTEXT32 +LLBL (G3TN_norm): + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND (EAX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ + + MOVQ ( MM0, MM3 ) /* x1 | x0 */ + MOVQ ( MM1, MM4 ) /* | x2 */ + + PFMUL ( MM0, MM3 ) /* x1*x1 | x0*x0 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + PFMUL ( MM1, MM4 ) /* | x2*x2 */ + PFADD ( MM4, MM3 ) /* | x0*x0+x2*x2 */ + + PFACC ( MM3, MM3 ) /* **not used** | x0*x0+x1*x1+x2**/ + PFRSQRT ( MM3, MM5 ) /* 1/sqrt (x0*x0+x1*x1+x2*x2) */ + + MOVQ ( MM5, MM4 ) + PUNPCKLDQ ( MM3, MM3 ) + + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + PFMUL ( MM5, MM5 ) + + PFRSQIT1 ( MM3, MM5 ) + PFRCPIT2 ( MM4, MM5 ) + + PFMUL ( MM5, MM0 ) /* x1 (normalized) | x0 (normalize*/ + + MOVQ ( MM0, REGOFF(-16, EAX) ) /* write new x0, x1 */ + PFMUL ( MM5, MM1 ) /* | x2 (normalize*/ + + MOVD ( MM1, REGOFF(-8, EAX) ) /* write new x2 */ + JNZ ( LLBL (G3TN_norm) ) + +LLBL (G3TN_exit_3dnow): + FEMMS + +LLBL (G3TN_end): + POP_L ( EBP ) + POP_L ( ESI ) + POP_L ( EDI ) + RET + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_transform_normalize_normals_no_rot) +HIDDEN(_mesa_3dnow_transform_normalize_normals_no_rot) +GLNAME(_mesa_3dnow_transform_normalize_normals_no_rot): + +#undef FRAME_OFFSET +#define FRAME_OFFSET 12 + + PUSH_L ( EDI ) + PUSH_L ( ESI ) + PUSH_L ( EBP ) + + MOV_L ( ARG_LENGTHS, EDI ) + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_DEST, EAX ) + MOV_L ( REGOFF(V4F_COUNT, ESI), EBP ) /* dest->count = in->count */ + MOV_L ( EBP, REGOFF(V4F_COUNT, EAX) ) + MOV_L ( ARG_MAT, ECX ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ + MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ + MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ + + CMP_L ( CONST(0), EBP ) /* count > 0 ?? */ + JE ( LLBL (G3TNNR_end) ) + + FEMMS + + MOVD ( M(0), MM0 ) /* | m0 */ + PUNPCKLDQ ( M(5), MM0 ) /* m5 | m0 */ + + MOVD ( M(10), MM2 ) /* | m10 */ + PUNPCKLDQ ( MM2, MM2 ) /* m10 | m10 */ + + CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ + JNE ( LLBL (G3TNNR_scale_end ) ) + + MOVD ( ARG_SCALE, MM7 ) /* | scale */ + PUNPCKLDQ ( MM7, MM7 ) /* scale | scale */ + + PFMUL ( MM7, MM0 ) /* scale * m5 | scale * m0 */ + PFMUL ( MM7, MM2 ) /* scale * m10 | scale * m10 */ + +ALIGNTEXT32 +LLBL (G3TNNR_scale_end): + CMP_L ( CONST(0), EDI ) /* lengths == 0 ? */ + JE ( LLBL (G3TNNR_norm) ) /* need to calculate lengths */ + + MOVD ( REGIND(EDI), MM3 ) /* | length (x) */ + + +ALIGNTEXT32 +LLBL (G3TNNR_norm_w_lengths): /* use precalculated lengths */ + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND(EDX), MM6 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EDX), MM7 ) /* | x2 */ + + PFMUL ( MM0, MM6 ) /* x1*m5 | x0*m0 */ + ADD_L ( STRIDE, EDX ) /* next normal */ + + PREFETCH ( REGIND(EDX) ) + + PFMUL ( MM2, MM7 ) /* | x2*m10 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + PFMUL ( MM3, MM7 ) /* | x2 (normalized) */ + PUNPCKLDQ ( MM3, MM3 ) /* length (x) | length (x) */ + + ADD_L ( CONST(4), EDI ) /* next length */ + PFMUL ( MM3, MM6 ) /* x1 (normalized) | x0 (normalized) */ + + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + MOVQ ( MM6, REGOFF(-16, EAX) ) /* write r0, r1 */ + + MOVD ( MM7, REGOFF(-8, EAX) ) /* write r2 */ + MOVD ( REGIND(EDI), MM3 ) /* | length (x) */ + + JNZ ( LLBL (G3TNNR_norm_w_lengths) ) + JMP ( LLBL (G3TNNR_exit_3dnow) ) + +ALIGNTEXT32 +LLBL (G3TNNR_norm): /* need to calculate lengths */ + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND(EDX), MM6 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EDX), MM7 ) /* | x2 */ + + PFMUL ( MM0, MM6 ) /* x1*m5 | x0*m0 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + PFMUL ( MM2, MM7 ) /* | x2*m10 */ + MOVQ ( MM6, MM3 ) /* x1 (transformed)| x0 (transformed) */ + + MOVQ ( MM7, MM4 ) /* | x2 (transformed) */ + PFMUL ( MM6, MM3 ) /* x1*x1 | x0*x0 */ + + + PFMUL ( MM7, MM4 ) /* | x2*x2 */ + PFACC ( MM3, MM3 ) /* **not used** | x0*x0+x1*x1 */ + + PFADD ( MM4, MM3 ) /* | x0*x0+x1*x1+x2*x2*/ + ADD_L ( STRIDE, EDX ) /* next normal */ + + PREFETCH ( REGIND(EDX) ) + + PFRSQRT ( MM3, MM5 ) /* 1/sqrt (x0*x0+x1*x1+x2*x2) */ + MOVQ ( MM5, MM4 ) + + PUNPCKLDQ ( MM3, MM3 ) + PFMUL ( MM5, MM5 ) + + PFRSQIT1 ( MM3, MM5 ) + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + + PFRCPIT2 ( MM4, MM5 ) + PFMUL ( MM5, MM6 ) /* x1 (normalized) | x0 (normalized) */ + + MOVQ ( MM6, REGOFF(-16, EAX) ) /* write r0, r1 */ + PFMUL ( MM5, MM7 ) /* | x2 (normalized) */ + + MOVD ( MM7, REGOFF(-8, EAX) ) /* write r2 */ + JNZ ( LLBL (G3TNNR_norm) ) + + +LLBL (G3TNNR_exit_3dnow): + FEMMS + +LLBL (G3TNNR_end): + POP_L ( EBP ) + POP_L ( ESI ) + POP_L ( EDI ) + RET + + + + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_transform_rescale_normals_no_rot) +HIDDEN(_mesa_3dnow_transform_rescale_normals_no_rot) +GLNAME(_mesa_3dnow_transform_rescale_normals_no_rot): + +#undef FRAME_OFFSET +#define FRAME_OFFSET 12 + + PUSH_L ( EDI ) + PUSH_L ( ESI ) + PUSH_L ( EBP ) + + MOV_L ( ARG_IN, EAX ) + MOV_L ( ARG_DEST, EDX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EBP ) /* dest->count = in->count */ + MOV_L ( EBP, REGOFF(V4F_COUNT, EDX) ) + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_MAT, ECX ) + MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ + MOV_L ( REGOFF(V4F_START, EDX), EAX ) /* dest->start */ + MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ + + CMP_L ( CONST(0), EBP ) + JE ( LLBL (G3TRNR_end) ) + + FEMMS + + MOVD ( ARG_SCALE, MM6 ) /* | scale */ + PUNPCKLDQ ( MM6, MM6 ) /* scale | scale */ + + MOVD ( REGIND(ECX), MM0 ) /* | m0 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m5 | m0 */ + + PFMUL ( MM6, MM0 ) /* scale*m5 | scale*m0 */ + MOVD ( REGOFF(40, ECX), MM2 ) /* | m10 */ + + PFMUL ( MM6, MM2 ) /* | scale*m10 */ + +ALIGNTEXT32 +LLBL (G3TRNR_rescale): + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND(EDX), MM4 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EDX), MM5 ) /* | x2 */ + + PFMUL ( MM0, MM4 ) /* x1*m5 | x0*m0 */ + ADD_L ( STRIDE, EDX ) /* next normal */ + + PREFETCH ( REGIND(EDX) ) + + PFMUL ( MM2, MM5 ) /* | x2*m10 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + MOVQ ( MM4, REGOFF(-16, EAX) ) /* write r0, r1 */ + + MOVD ( MM5, REGOFF(-8, EAX) ) /* write r2 */ + JNZ ( LLBL (G3TRNR_rescale) ) /* cnt > 0 ? -> process next normal */ + + FEMMS + +LLBL (G3TRNR_end): + POP_L ( EBP ) + POP_L ( ESI ) + POP_L ( EDI ) + RET + + + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_transform_rescale_normals) +HIDDEN(_mesa_3dnow_transform_rescale_normals) +GLNAME(_mesa_3dnow_transform_rescale_normals): + +#undef FRAME_OFFSET +#define FRAME_OFFSET 8 + + PUSH_L ( EDI ) + PUSH_L ( ESI ) + + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_DEST, EAX ) + MOV_L ( ARG_MAT, ECX ) + MOV_L ( REGOFF(V4F_COUNT, ESI), EDI ) /* dest->count = in->count */ + MOV_L ( EDI, REGOFF(V4F_COUNT, EAX) ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ + MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ + MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ + + CMP_L ( CONST(0), EDI ) + JE ( LLBL (G3TR_end) ) + + FEMMS + + MOVQ ( REGIND(ECX), MM3 ) /* m1 | m0 */ + + MOVQ ( REGOFF(16,ECX), MM4 ) /* m5 | m4 */ + MOVD ( ARG_SCALE, MM0 ) /* scale */ + + MOVD ( REGOFF(8,ECX), MM5 ) /* | m2 */ + PUNPCKLDQ ( MM0, MM0 ) /* scale | scale */ + + PUNPCKLDQ ( REGOFF(24, ECX), MM5 ) + PFMUL ( MM0, MM3 ) /* scale*m1 | scale*m0 */ + + MOVQ ( REGOFF(32, ECX), MM6 ) /* m9 | m8*/ + PFMUL ( MM0, MM4 ) /* scale*m5 | scale*m4 */ + + MOVD ( REGOFF(40, ECX), MM7 ) /* | m10 */ + PFMUL ( MM0, MM5 ) /* scale*m6 | scale*m2 */ + + PFMUL ( MM0, MM6 ) /* scale*m9 | scale*m8 */ + + PFMUL ( MM0, MM7 ) /* | scale*m10 */ + +ALIGNTEXT32 +LLBL (G3TR_rescale): + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND(EDX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ + + MOVQ ( MM0, MM1 ) /* x1 | x0 */ + PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ + + PFMUL ( MM3, MM0 ) /* x1*m1 | x0*m0 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + PFMUL ( MM4, MM1 ) /* x1*m5 | x0*m4 */ + PFACC ( MM1, MM0 ) /* x0*m4+x1*m5 | x0*m0+x1*m1 */ + + MOVQ ( REGIND(EDX), MM1 ) /* x1 | x0 */ + + PFMUL ( MM5, MM2 ) /* x2*m6 | x2*m2 */ + PFADD ( MM2, MM0 ) /* x0*m4...+x2*m6| x0*m0+x1*m1+x2*m2 */ + + MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ + ADD_L ( STRIDE, EDX ) /* next normal */ + + PREFETCH ( REGIND(EDX) ) + + MOVQ ( MM0, REGOFF(-16, EAX) ) /* write r0, r1 */ + PFMUL ( MM6, MM1 ) /* x1*m9 | x0*m8 */ + + PFMUL ( MM7, MM2 ) /* | x2*m10 */ + PFACC ( MM1, MM1 ) /* *not used* | x0*m8+x1*m9 */ + + PFADD ( MM2, MM1 ) /* *not used* | x0*m8+x1*m9+x2*m10 */ + MOVD ( MM1, REGOFF(-8, EAX) ) /* write r2 */ + + SUB_L ( CONST(1), EDI ) /* decrement normal counter */ + JNZ ( LLBL (G3TR_rescale) ) + + FEMMS + +LLBL (G3TR_end): + POP_L ( ESI ) + POP_L ( EDI ) + RET + + + + + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_transform_normals_no_rot) +HIDDEN(_mesa_3dnow_transform_normals_no_rot) +GLNAME(_mesa_3dnow_transform_normals_no_rot): + +#undef FRAME_OFFSET +#define FRAME_OFFSET 8 + + PUSH_L ( EDI ) + PUSH_L ( ESI ) + + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_DEST, EAX ) + MOV_L ( ARG_MAT, ECX ) + MOV_L ( REGOFF(V4F_COUNT, ESI), EDI ) /* dest->count = in->count */ + MOV_L ( EDI, REGOFF(V4F_COUNT, EAX) ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ + MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ + MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ + + CMP_L ( CONST(0), EDI ) + JE ( LLBL (G3TNR_end) ) + + FEMMS + + MOVD ( REGIND(ECX), MM0 ) /* | m0 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m5 | m0 */ + + MOVD ( REGOFF(40, ECX), MM2 ) /* | m10 */ + PUNPCKLDQ ( MM2, MM2 ) /* m10 | m10 */ + +ALIGNTEXT32 +LLBL (G3TNR_transform): + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND(EDX), MM4 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EDX), MM5 ) /* | x2 */ + + PFMUL ( MM0, MM4 ) /* x1*m5 | x0*m0 */ + ADD_L ( STRIDE, EDX) /* next normal */ + + PREFETCH ( REGIND(EDX) ) + + PFMUL ( MM2, MM5 ) /* | x2*m10 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + SUB_L ( CONST(1), EDI ) /* decrement normal counter */ + MOVQ ( MM4, REGOFF(-16, EAX) ) /* write r0, r1 */ + + MOVD ( MM5, REGOFF(-8, EAX) ) /* write r2 */ + JNZ ( LLBL (G3TNR_transform) ) + + FEMMS + +LLBL (G3TNR_end): + POP_L ( ESI ) + POP_L ( EDI ) + RET + + + + + + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_transform_normals) +HIDDEN(_mesa_3dnow_transform_normals) +GLNAME(_mesa_3dnow_transform_normals): + +#undef FRAME_OFFSET +#define FRAME_OFFSET 8 + + PUSH_L ( EDI ) + PUSH_L ( ESI ) + + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_DEST, EAX ) + MOV_L ( ARG_MAT, ECX ) + MOV_L ( REGOFF(V4F_COUNT, ESI), EDI ) /* dest->count = in->count */ + MOV_L ( EDI, REGOFF(V4F_COUNT, EAX) ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ + MOV_L ( REGOFF(V4F_START, ESI), EDX ) /* in->start */ + MOV_L ( REGOFF(MATRIX_INV, ECX), ECX ) /* mat->inv */ + + CMP_L ( CONST(0), EDI ) /* count > 0 ?? */ + JE ( LLBL (G3T_end) ) + + FEMMS + + MOVQ ( REGIND(ECX), MM3 ) /* m1 | m0 */ + MOVQ ( REGOFF(16, ECX), MM4 ) /* m5 | m4 */ + + MOVD ( REGOFF(8, ECX), MM5 ) /* | m2 */ + PUNPCKLDQ ( REGOFF(24, ECX), MM5 ) /* m6 | m2 */ + + MOVQ ( REGOFF(32, ECX), MM6 ) /* m9 | m8 */ + MOVD ( REGOFF(40, ECX), MM7 ) /* | m10 */ + +ALIGNTEXT32 +LLBL (G3T_transform): + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND(EDX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ + + MOVQ ( MM0, MM1 ) /* x1 | x0 */ + PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ + + PFMUL ( MM3, MM0 ) /* x1*m1 | x0*m0 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + PFMUL ( MM4, MM1 ) /* x1*m5 | x0*m4 */ + PFACC ( MM1, MM0 ) /* x0*m4+x1*m5 | x0*m0+x1*m1 */ + + PFMUL ( MM5, MM2 ) /* x2*m6 | x2*m2 */ + PFADD ( MM2, MM0 ) /* x0*m4...+x2*m6| x0*m0+x1*m1+x2*m2 */ + + MOVQ ( REGIND(EDX), MM1 ) /* x1 | x0 */ + MOVQ ( MM0, REGOFF(-16, EAX) ) /* write r0, r1 */ + + PFMUL ( MM6, MM1 ) /* x1*m9 | x0*m8 */ + MOVD ( REGOFF(8, EDX), MM2 ) /* | x2 */ + + PFMUL ( MM7, MM2 ) /* | x2*m10 */ + ADD_L ( STRIDE, EDX ) /* next normal */ + + PREFETCH ( REGIND(EDX) ) + + PFACC ( MM1, MM1 ) /* *not used* | x0*m8+x1*m9 */ + PFADD ( MM2, MM1 ) /* *not used* | x0*m8+x1*m9+x2*m10 */ + + MOVD ( MM1, REGOFF(-8, EAX) ) /* write r2 */ + SUB_L ( CONST(1), EDI ) /* decrement normal counter */ + + JNZ ( LLBL (G3T_transform) ) + + FEMMS + +LLBL (G3T_end): + POP_L ( ESI ) + POP_L ( EDI ) + RET + + + + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_normalize_normals) +HIDDEN(_mesa_3dnow_normalize_normals) +GLNAME(_mesa_3dnow_normalize_normals): + +#undef FRAME_OFFSET +#define FRAME_OFFSET 12 + + PUSH_L ( EDI ) + PUSH_L ( ESI ) + PUSH_L ( EBP ) + + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_DEST, EAX ) + MOV_L ( REGOFF(V4F_COUNT, ESI), EBP ) /* dest->count = in->count */ + MOV_L ( EBP, REGOFF(V4F_COUNT, EAX) ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ + MOV_L ( REGOFF(V4F_START, ESI), ECX ) /* in->start */ + MOV_L ( ARG_LENGTHS, EDX ) + + CMP_L ( CONST(0), EBP ) /* count > 0 ?? */ + JE ( LLBL (G3N_end) ) + + FEMMS + + CMP_L ( CONST(0), EDX ) /* lengths == 0 ? */ + JE ( LLBL (G3N_norm2) ) /* calculate lengths */ + +ALIGNTEXT32 +LLBL (G3N_norm1): /* use precalculated lengths */ + + PREFETCH ( REGIND(EAX) ) + + MOVQ ( REGIND(ECX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, ECX), MM1 ) /* | x2 */ + + MOVD ( REGIND(EDX), MM3 ) /* | length (x) */ + PFMUL ( MM3, MM1 ) /* | x2 (normalized) */ + + PUNPCKLDQ ( MM3, MM3 ) /* length (x) | length (x) */ + ADD_L ( STRIDE, ECX ) /* next normal */ + + PREFETCH ( REGIND(ECX) ) + + PFMUL ( MM3, MM0 ) /* x1 (normalized) | x0 (normalized) */ + MOVQ ( MM0, REGIND(EAX) ) /* write new x0, x1 */ + + MOVD ( MM1, REGOFF(8, EAX) ) /* write new x2 */ + ADD_L ( CONST(16), EAX ) /* next r */ + + ADD_L ( CONST(4), EDX ) /* next length */ + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + + JNZ ( LLBL (G3N_norm1) ) + + JMP ( LLBL (G3N_end1) ) + +ALIGNTEXT32 +LLBL (G3N_norm2): /* need to calculate lengths */ + + PREFETCHW ( REGIND(EAX) ) + + PREFETCH ( REGIND(ECX) ) + + MOVQ ( REGIND(ECX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, ECX), MM1 ) /* | x2 */ + + MOVQ ( MM0, MM3 ) /* x1 | x0 */ + ADD_L ( STRIDE, ECX ) /* next normal */ + + PFMUL ( MM0, MM3 ) /* x1*x1 | x0*x0 */ + MOVQ ( MM1, MM4 ) /* | x2 */ + + ADD_L ( CONST(16), EAX ) /* next r */ + PFMUL ( MM1, MM4 ) /* | x2*x2 */ + + PFADD ( MM4, MM3 ) /* | x0*x0+x2*x2 */ + PFACC ( MM3, MM3 ) /* x0*x0+...+x2*x2 | x0*x0+x1*x1+x2*x2*/ + + PFRSQRT ( MM3, MM5 ) /* 1/sqrt (x0*x0+x1*x1+x2*x2) */ + MOVQ ( MM5, MM4 ) + + PUNPCKLDQ ( MM3, MM3 ) + PFMUL ( MM5, MM5 ) + + PFRSQIT1 ( MM3, MM5 ) + SUB_L ( CONST(1), EBP ) /* decrement normal counter */ + + PFRCPIT2 ( MM4, MM5 ) + + PFMUL ( MM5, MM0 ) /* x1 (normalized) | x0 (normalized) */ + MOVQ ( MM0, REGOFF(-16, EAX) ) /* write new x0, x1 */ + + PFMUL ( MM5, MM1 ) /* | x2 (normalized) */ + MOVD ( MM1, REGOFF(-8, EAX) ) /* write new x2 */ + + JNZ ( LLBL (G3N_norm2) ) + +LLBL (G3N_end1): + FEMMS + +LLBL (G3N_end): + POP_L ( EBP ) + POP_L ( ESI ) + POP_L ( EDI ) + RET + + + + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_3dnow_rescale_normals) +HIDDEN(_mesa_3dnow_rescale_normals) +GLNAME(_mesa_3dnow_rescale_normals): + +#undef FRAME_OFFSET +#define FRAME_OFFSET 8 + PUSH_L ( EDI ) + PUSH_L ( ESI ) + + MOV_L ( ARG_IN, ESI ) + MOV_L ( ARG_DEST, EAX ) + MOV_L ( REGOFF(V4F_COUNT, ESI), EDX ) /* dest->count = in->count */ + MOV_L ( EDX, REGOFF(V4F_COUNT, EAX) ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) /* dest->start */ + MOV_L ( REGOFF(V4F_START, ESI), ECX ) /* in->start */ + + CMP_L ( CONST(0), EDX ) + JE ( LLBL (G3R_end) ) + + FEMMS + + MOVD ( ARG_SCALE, MM0 ) /* scale */ + PUNPCKLDQ ( MM0, MM0 ) + +ALIGNTEXT32 +LLBL (G3R_rescale): + + PREFETCHW ( REGIND(EAX) ) + + MOVQ ( REGIND(ECX), MM1 ) /* x1 | x0 */ + MOVD ( REGOFF(8, ECX), MM2 ) /* | x2 */ + + PFMUL ( MM0, MM1 ) /* x1*scale | x0*scale */ + ADD_L ( STRIDE, ECX ) /* next normal */ + + PREFETCH ( REGIND(ECX) ) + + PFMUL ( MM0, MM2 ) /* | x2*scale */ + ADD_L ( CONST(16), EAX ) /* next r */ + + MOVQ ( MM1, REGOFF(-16, EAX) ) /* write r0, r1 */ + MOVD ( MM2, REGOFF(-8, EAX) ) /* write r2 */ + + SUB_L ( CONST(1), EDX ) /* decrement normal counter */ + JNZ ( LLBL (G3R_rescale) ) + + FEMMS + +LLBL (G3R_end): + POP_L ( ESI ) + POP_L ( EDI ) + RET + +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/3dnow_xform1.S b/mesalib/src/mesa/x86/3dnow_xform1.S index 9719ba2c1..a73301a8d 100644 --- a/mesalib/src/mesa/x86/3dnow_xform1.S +++ b/mesalib/src/mesa/x86/3dnow_xform1.S @@ -1,437 +1,437 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#ifdef USE_3DNOW_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FRAME_OFFSET 4 - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points1_general ) -HIDDEN(_mesa_3dnow_transform_points1_general) -GLNAME( _mesa_3dnow_transform_points1_general ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(4, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPGR_3 ) ) - - MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ - MOVQ ( REGOFF(8, ECX), MM1 ) /* m03 | m02 */ - - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - MOVQ ( REGOFF(56, ECX), MM3 ) /* m33 | m32 */ - -ALIGNTEXT16 -LLBL( G3TPGR_2 ): - - MOVD ( REGIND(EAX), MM4 ) /* | x0 */ - PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ - - MOVQ ( MM4, MM5 ) /* x0 | x0 */ - PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ - - PFMUL ( MM1, MM5 ) /* x0*m03 | x0*m02 */ - PFADD ( MM2, MM4 ) /* x0*m01+m31 | x0*m00+m30 */ - - PFADD ( MM3, MM5 ) /* x0*m03+m33 | x0*m02+m32 */ - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - - MOVQ ( MM5, REGOFF(8, EDX) ) /* write r3, r2 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TPGR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPGR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points1_identity ) -HIDDEN(_mesa_3dnow_transform_points1_identity) -GLNAME( _mesa_3dnow_transform_points1_identity ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(1), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_1), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(4, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPIR_4) ) - -ALIGNTEXT16 -LLBL( G3TPIR_3 ): - - MOVD ( REGIND(EAX), MM0 ) /* | x0 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - MOVD ( MM0, REGIND(EDX) ) /* | r0 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPIR_3 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPIR_4 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points1_3d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points1_3d_no_rot) -GLNAME( _mesa_3dnow_transform_points1_3d_no_rot ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(4, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3NRR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - - MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ - -ALIGNTEXT16 -LLBL( G3TP3NRR_2 ): - - MOVD ( REGIND(EAX), MM4 ) /* | x0 */ - PFMUL ( MM0, MM4 ) /* | x0*m00 */ - - PFADD ( MM2, MM4 ) /* m31 | x0*m00+m30 */ - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - - MOVD ( MM3, REGOFF(8, EDX) ) /* write r2 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TP3NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3NRR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points1_perspective ) -HIDDEN(_mesa_3dnow_transform_points1_perspective) -GLNAME( _mesa_3dnow_transform_points1_perspective ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(4, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPPR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ - -ALIGNTEXT16 -LLBL( G3TPPR_2 ): - - MOVD ( REGIND(EAX), MM4 ) /* 0 | x0 */ - PFMUL ( MM0, MM4 ) /* 0 | x0*m00 */ - - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - MOVQ ( MM3, REGOFF(8, EDX) ) /* write r2 (=m32), r3 (=0) */ - - ADD_L ( EDI, EAX ) /* next vertex */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPPR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPPR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points1_2d ) -HIDDEN(_mesa_3dnow_transform_points1_2d) -GLNAME( _mesa_3dnow_transform_points1_2d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(4, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2R_3 ) ) - - MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP2R_2 ): - - MOVD ( REGIND(EAX), MM4 ) /* | x0 */ - PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ - - PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ - PFADD ( MM2, MM4 ) /* x0*m01+m31 | x0*m00+m30 */ - - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TP2R_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2R_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points1_2d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points1_2d_no_rot) -GLNAME( _mesa_3dnow_transform_points1_2d_no_rot ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(4, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2NRR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP2NRR_2 ): - - MOVD ( REGIND(EAX), MM4 ) /* | x0 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - PFMUL ( MM0, MM4 ) /* | x0*m00 */ - PFADD ( MM2, MM4 ) /* m31 | x0*m00+m30 */ - - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP2NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2NRR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points1_3d ) -HIDDEN(_mesa_3dnow_transform_points1_3d) -GLNAME( _mesa_3dnow_transform_points1_3d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(4, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3R_3 ) ) - - MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ - MOVD ( REGOFF(8, ECX), MM1 ) /* | m02 */ - - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ - -ALIGNTEXT16 -LLBL( G3TP3R_2 ): - - MOVD ( REGIND(EAX), MM4 ) /* | x0 */ - PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ - - MOVQ ( MM4, MM5 ) /* | x0 */ - PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ - - PFMUL ( MM1, MM5 ) /* | x0*m02 */ - PFADD ( MM2, MM4 ) /* x0*m01+m31 | x0*m00+m30 */ - - PFADD ( MM3, MM5 ) /* | x0*m02+m32 */ - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - - MOVD ( MM5, REGOFF(8, EDX) ) /* write r2 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TP3R_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3R_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifdef USE_3DNOW_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FRAME_OFFSET 4 + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points1_general ) +HIDDEN(_mesa_3dnow_transform_points1_general) +GLNAME( _mesa_3dnow_transform_points1_general ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(4, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPGR_3 ) ) + + MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ + MOVQ ( REGOFF(8, ECX), MM1 ) /* m03 | m02 */ + + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + MOVQ ( REGOFF(56, ECX), MM3 ) /* m33 | m32 */ + +ALIGNTEXT16 +LLBL( G3TPGR_2 ): + + MOVD ( REGIND(EAX), MM4 ) /* | x0 */ + PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ + + MOVQ ( MM4, MM5 ) /* x0 | x0 */ + PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ + + PFMUL ( MM1, MM5 ) /* x0*m03 | x0*m02 */ + PFADD ( MM2, MM4 ) /* x0*m01+m31 | x0*m00+m30 */ + + PFADD ( MM3, MM5 ) /* x0*m03+m33 | x0*m02+m32 */ + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + + MOVQ ( MM5, REGOFF(8, EDX) ) /* write r3, r2 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TPGR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPGR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points1_identity ) +HIDDEN(_mesa_3dnow_transform_points1_identity) +GLNAME( _mesa_3dnow_transform_points1_identity ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(1), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_1), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(4, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPIR_4) ) + +ALIGNTEXT16 +LLBL( G3TPIR_3 ): + + MOVD ( REGIND(EAX), MM0 ) /* | x0 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + MOVD ( MM0, REGIND(EDX) ) /* | r0 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPIR_3 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPIR_4 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points1_3d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points1_3d_no_rot) +GLNAME( _mesa_3dnow_transform_points1_3d_no_rot ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(4, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3NRR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + + MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ + +ALIGNTEXT16 +LLBL( G3TP3NRR_2 ): + + MOVD ( REGIND(EAX), MM4 ) /* | x0 */ + PFMUL ( MM0, MM4 ) /* | x0*m00 */ + + PFADD ( MM2, MM4 ) /* m31 | x0*m00+m30 */ + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + + MOVD ( MM3, REGOFF(8, EDX) ) /* write r2 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TP3NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3NRR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points1_perspective ) +HIDDEN(_mesa_3dnow_transform_points1_perspective) +GLNAME( _mesa_3dnow_transform_points1_perspective ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(4, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPPR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ + +ALIGNTEXT16 +LLBL( G3TPPR_2 ): + + MOVD ( REGIND(EAX), MM4 ) /* 0 | x0 */ + PFMUL ( MM0, MM4 ) /* 0 | x0*m00 */ + + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + MOVQ ( MM3, REGOFF(8, EDX) ) /* write r2 (=m32), r3 (=0) */ + + ADD_L ( EDI, EAX ) /* next vertex */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPPR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPPR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points1_2d ) +HIDDEN(_mesa_3dnow_transform_points1_2d) +GLNAME( _mesa_3dnow_transform_points1_2d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(4, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2R_3 ) ) + + MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP2R_2 ): + + MOVD ( REGIND(EAX), MM4 ) /* | x0 */ + PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ + + PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ + PFADD ( MM2, MM4 ) /* x0*m01+m31 | x0*m00+m30 */ + + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TP2R_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2R_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points1_2d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points1_2d_no_rot) +GLNAME( _mesa_3dnow_transform_points1_2d_no_rot ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(4, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2NRR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP2NRR_2 ): + + MOVD ( REGIND(EAX), MM4 ) /* | x0 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + PFMUL ( MM0, MM4 ) /* | x0*m00 */ + PFADD ( MM2, MM4 ) /* m31 | x0*m00+m30 */ + + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP2NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2NRR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points1_3d ) +HIDDEN(_mesa_3dnow_transform_points1_3d) +GLNAME( _mesa_3dnow_transform_points1_3d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(4, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3R_3 ) ) + + MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ + MOVD ( REGOFF(8, ECX), MM1 ) /* | m02 */ + + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ + +ALIGNTEXT16 +LLBL( G3TP3R_2 ): + + MOVD ( REGIND(EAX), MM4 ) /* | x0 */ + PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ + + MOVQ ( MM4, MM5 ) /* | x0 */ + PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ + + PFMUL ( MM1, MM5 ) /* | x0*m02 */ + PFADD ( MM2, MM4 ) /* x0*m01+m31 | x0*m00+m30 */ + + PFADD ( MM3, MM5 ) /* | x0*m02+m32 */ + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + + MOVD ( MM5, REGOFF(8, EDX) ) /* write r2 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TP3R_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3R_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/3dnow_xform2.S b/mesalib/src/mesa/x86/3dnow_xform2.S index 8c706074a..2988fb7bf 100644 --- a/mesalib/src/mesa/x86/3dnow_xform2.S +++ b/mesalib/src/mesa/x86/3dnow_xform2.S @@ -1,477 +1,477 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#ifdef USE_3DNOW_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FRAME_OFFSET 4 - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points2_general ) -HIDDEN(_mesa_3dnow_transform_points2_general) -GLNAME( _mesa_3dnow_transform_points2_general ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPGR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ - - MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ - - MOVD ( REGOFF(8, ECX), MM2 ) /* | m02 */ - PUNPCKLDQ ( REGOFF(24, ECX), MM2 ) /* m12 | m02 */ - - MOVD ( REGOFF(12, ECX), MM3 ) /* | m03 */ - PUNPCKLDQ ( REGOFF(28, ECX), MM3 ) /* m13 | m03 */ - - MOVQ ( REGOFF(48, ECX), MM4 ) /* m31 | m30 */ - MOVQ ( REGOFF(56, ECX), MM5 ) /* m33 | m32 */ - -ALIGNTEXT16 -LLBL( G3TPGR_2 ): - - MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ - MOVQ ( MM6, MM7 ) /* x1 | x0 */ - - PFMUL ( MM0, MM6 ) /* x1*m10 | x0*m00 */ - PFMUL ( MM1, MM7 ) /* x1*m11 | x0*m01 */ - - PFACC ( MM7, MM6 ) /* x0*m01+x1*m11 | x0*x00+x1*m10 */ - PFADD ( MM4, MM6 ) /* x0*...*m11+m31 | x0*...*m10+m30 */ - - MOVQ ( MM6, REGIND(EDX) ) /* write r1, r0 */ - MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ - - MOVQ ( MM6, MM7 ) /* x1 | x0 */ - PFMUL ( MM2, MM6 ) /* x1*m12 | x0*m02 */ - - PFMUL ( MM3, MM7 ) /* x1*m13 | x0*m03 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - PFACC ( MM7, MM6 ) /* x0*m03+x1*m13 | x0*x02+x1*m12 */ - PFADD ( MM5, MM6 ) /* x0*...*m13+m33 | x0*...*m12+m32 */ - - MOVQ ( MM6, REGOFF(8, EDX) ) /* write r3, r2 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPGR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPGR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points2_perspective ) -HIDDEN(_mesa_3dnow_transform_points2_perspective) -GLNAME( _mesa_3dnow_transform_points2_perspective ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPPR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ - -ALIGNTEXT16 -LLBL( G3TPPR_2 ): - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - MOVQ ( MM3, REGOFF(8, EDX) ) /* write r2 (=m32), r3 (=0) */ - - ADD_L ( EDI, EAX ) /* next vertex */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPPR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPPR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points2_3d ) -HIDDEN(_mesa_3dnow_transform_points2_3d) -GLNAME( _mesa_3dnow_transform_points2_3d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3 ), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3R_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ - - MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ - - MOVD ( REGOFF(8, ECX), MM2 ) /* | m02 */ - PUNPCKLDQ ( REGOFF(24, ECX), MM2 ) /* m12 | m02 */ - - MOVQ ( REGOFF(48, ECX), MM4 ) /* m31 | m30 */ - MOVD ( REGOFF(56, ECX), MM5 ) /* | m32 */ - -ALIGNTEXT16 -LLBL( G3TP3R_2 ): - - MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ - MOVQ ( MM6, MM7 ) /* x1 | x0 */ - - PFMUL ( MM0, MM6 ) /* x1*m10 | x0*m00 */ - PFMUL ( MM1, MM7 ) /* x1*m11 | x0*m01 */ - - PFACC ( MM7, MM6 ) /* x0*m01+x1*m11 | x0*x00+x1*m10 */ - PFADD ( MM4, MM6 ) /* x0*...*m11+m31 | x0*...*m10+m30 */ - - MOVQ ( MM6, REGIND(EDX) ) /* write r1, r0 */ - MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ - - MOVQ ( MM6, MM7 ) /* x1 | x0 */ - PFMUL ( MM2, MM6 ) /* x1*m12 | x0*m02 */ - - PFACC ( MM7, MM6 ) /* ***trash*** | x0*x02+x1*m12 */ - PFADD ( MM5, MM6 ) /* ***trash*** | x0*...*m12+m32 */ - - MOVD ( MM6, REGOFF(8, EDX) ) /* write r2 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TP3R_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3R_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points2_3d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points2_3d_no_rot) -GLNAME( _mesa_3dnow_transform_points2_3d_no_rot ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3 ), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3NRR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ - -ALIGNTEXT16 -LLBL( G3TP3NRR_2 ): - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - - PFADD ( MM2, MM4 ) /* x1*m11+m31 | x0*m00+m30 */ - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - - MOVD ( MM3, REGOFF(8, EDX) ) /* write r2 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TP3NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3NRR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points2_2d ) -HIDDEN(_mesa_3dnow_transform_points2_2d) -GLNAME( _mesa_3dnow_transform_points2_2d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2R_3 ) ) - - MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ - MOVQ ( REGOFF(16, ECX), MM1 ) /* m11 | m10 */ - - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP2R_2 ): - - MOVD ( REGIND(EAX), MM4 ) /* | x0 */ - MOVD ( REGOFF(4, EAX), MM5 ) /* | x1 */ - - PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ - PUNPCKLDQ ( MM5, MM5 ) /* x1 | x1 */ - - PFMUL ( MM1, MM5 ) /* x1*m11 | x1*m10 */ - PFADD ( MM2, MM4 ) /* x...x1*m11+31 | x0*..*m10+m30 */ - - PFADD ( MM5, MM4 ) /* x0*m01+x1*m11 | x0*m00+x1*m10 */ - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TP2R_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2R_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points2_2d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points2_2d_no_rot) -GLNAME( _mesa_3dnow_transform_points2_2d_no_rot ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2NRR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP2NRR_2 ): - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - PFADD ( MM2, MM4 ) /* m31 | x0*m00+m30 */ - - MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP2NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2NRR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points2_identity ) -HIDDEN(_mesa_3dnow_transform_points2_identity) -GLNAME( _mesa_3dnow_transform_points2_identity ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPIR_3 ) ) - -ALIGNTEXT16 -LLBL( G3TPIR_3 ): - - MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ - ADD_L ( EDI, EAX ) /* next vertex */ - - MOVQ ( MM0, REGIND(EDX) ) /* r1 | r0 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPIR_3 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPIR_4 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifdef USE_3DNOW_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FRAME_OFFSET 4 + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points2_general ) +HIDDEN(_mesa_3dnow_transform_points2_general) +GLNAME( _mesa_3dnow_transform_points2_general ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPGR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ + + MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ + + MOVD ( REGOFF(8, ECX), MM2 ) /* | m02 */ + PUNPCKLDQ ( REGOFF(24, ECX), MM2 ) /* m12 | m02 */ + + MOVD ( REGOFF(12, ECX), MM3 ) /* | m03 */ + PUNPCKLDQ ( REGOFF(28, ECX), MM3 ) /* m13 | m03 */ + + MOVQ ( REGOFF(48, ECX), MM4 ) /* m31 | m30 */ + MOVQ ( REGOFF(56, ECX), MM5 ) /* m33 | m32 */ + +ALIGNTEXT16 +LLBL( G3TPGR_2 ): + + MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ + MOVQ ( MM6, MM7 ) /* x1 | x0 */ + + PFMUL ( MM0, MM6 ) /* x1*m10 | x0*m00 */ + PFMUL ( MM1, MM7 ) /* x1*m11 | x0*m01 */ + + PFACC ( MM7, MM6 ) /* x0*m01+x1*m11 | x0*x00+x1*m10 */ + PFADD ( MM4, MM6 ) /* x0*...*m11+m31 | x0*...*m10+m30 */ + + MOVQ ( MM6, REGIND(EDX) ) /* write r1, r0 */ + MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ + + MOVQ ( MM6, MM7 ) /* x1 | x0 */ + PFMUL ( MM2, MM6 ) /* x1*m12 | x0*m02 */ + + PFMUL ( MM3, MM7 ) /* x1*m13 | x0*m03 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + PFACC ( MM7, MM6 ) /* x0*m03+x1*m13 | x0*x02+x1*m12 */ + PFADD ( MM5, MM6 ) /* x0*...*m13+m33 | x0*...*m12+m32 */ + + MOVQ ( MM6, REGOFF(8, EDX) ) /* write r3, r2 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPGR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPGR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points2_perspective ) +HIDDEN(_mesa_3dnow_transform_points2_perspective) +GLNAME( _mesa_3dnow_transform_points2_perspective ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPPR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ + +ALIGNTEXT16 +LLBL( G3TPPR_2 ): + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + MOVQ ( MM3, REGOFF(8, EDX) ) /* write r2 (=m32), r3 (=0) */ + + ADD_L ( EDI, EAX ) /* next vertex */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPPR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPPR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points2_3d ) +HIDDEN(_mesa_3dnow_transform_points2_3d) +GLNAME( _mesa_3dnow_transform_points2_3d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3 ), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3R_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ + + MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ + + MOVD ( REGOFF(8, ECX), MM2 ) /* | m02 */ + PUNPCKLDQ ( REGOFF(24, ECX), MM2 ) /* m12 | m02 */ + + MOVQ ( REGOFF(48, ECX), MM4 ) /* m31 | m30 */ + MOVD ( REGOFF(56, ECX), MM5 ) /* | m32 */ + +ALIGNTEXT16 +LLBL( G3TP3R_2 ): + + MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ + MOVQ ( MM6, MM7 ) /* x1 | x0 */ + + PFMUL ( MM0, MM6 ) /* x1*m10 | x0*m00 */ + PFMUL ( MM1, MM7 ) /* x1*m11 | x0*m01 */ + + PFACC ( MM7, MM6 ) /* x0*m01+x1*m11 | x0*x00+x1*m10 */ + PFADD ( MM4, MM6 ) /* x0*...*m11+m31 | x0*...*m10+m30 */ + + MOVQ ( MM6, REGIND(EDX) ) /* write r1, r0 */ + MOVQ ( REGIND(EAX), MM6 ) /* x1 | x0 */ + + MOVQ ( MM6, MM7 ) /* x1 | x0 */ + PFMUL ( MM2, MM6 ) /* x1*m12 | x0*m02 */ + + PFACC ( MM7, MM6 ) /* ***trash*** | x0*x02+x1*m12 */ + PFADD ( MM5, MM6 ) /* ***trash*** | x0*...*m12+m32 */ + + MOVD ( MM6, REGOFF(8, EDX) ) /* write r2 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TP3R_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3R_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points2_3d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points2_3d_no_rot) +GLNAME( _mesa_3dnow_transform_points2_3d_no_rot ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3 ), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3NRR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ + +ALIGNTEXT16 +LLBL( G3TP3NRR_2 ): + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + + PFADD ( MM2, MM4 ) /* x1*m11+m31 | x0*m00+m30 */ + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + + MOVD ( MM3, REGOFF(8, EDX) ) /* write r2 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TP3NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3NRR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points2_2d ) +HIDDEN(_mesa_3dnow_transform_points2_2d) +GLNAME( _mesa_3dnow_transform_points2_2d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2R_3 ) ) + + MOVQ ( REGIND(ECX), MM0 ) /* m01 | m00 */ + MOVQ ( REGOFF(16, ECX), MM1 ) /* m11 | m10 */ + + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP2R_2 ): + + MOVD ( REGIND(EAX), MM4 ) /* | x0 */ + MOVD ( REGOFF(4, EAX), MM5 ) /* | x1 */ + + PUNPCKLDQ ( MM4, MM4 ) /* x0 | x0 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + PFMUL ( MM0, MM4 ) /* x0*m01 | x0*m00 */ + PUNPCKLDQ ( MM5, MM5 ) /* x1 | x1 */ + + PFMUL ( MM1, MM5 ) /* x1*m11 | x1*m10 */ + PFADD ( MM2, MM4 ) /* x...x1*m11+31 | x0*..*m10+m30 */ + + PFADD ( MM5, MM4 ) /* x0*m01+x1*m11 | x0*m00+x1*m10 */ + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TP2R_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2R_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points2_2d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points2_2d_no_rot) +GLNAME( _mesa_3dnow_transform_points2_2d_no_rot ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2NRR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP2NRR_2 ): + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + PFADD ( MM2, MM4 ) /* m31 | x0*m00+m30 */ + + MOVQ ( MM4, REGIND(EDX) ) /* write r1, r0 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP2NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2NRR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points2_identity ) +HIDDEN(_mesa_3dnow_transform_points2_identity) +GLNAME( _mesa_3dnow_transform_points2_identity ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(2), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPIR_3 ) ) + +ALIGNTEXT16 +LLBL( G3TPIR_3 ): + + MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ + ADD_L ( EDI, EAX ) /* next vertex */ + + MOVQ ( MM0, REGIND(EDX) ) /* r1 | r0 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPIR_3 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPIR_4 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/3dnow_xform3.S b/mesalib/src/mesa/x86/3dnow_xform3.S index d119fe821..a356aaee7 100644 --- a/mesalib/src/mesa/x86/3dnow_xform3.S +++ b/mesalib/src/mesa/x86/3dnow_xform3.S @@ -1,561 +1,561 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#ifdef USE_3DNOW_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FRAME_OFFSET 4 - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points3_general ) -HIDDEN(_mesa_3dnow_transform_points3_general) -GLNAME( _mesa_3dnow_transform_points3_general ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPGR_2 ) ) - - PREFETCHW ( REGIND(EDX) ) - -ALIGNTEXT16 -LLBL( G3TPGR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM2 ) /* | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - MOVQ ( MM0, MM1 ) /* x1 | x0 */ - PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ - - PUNPCKLDQ ( MM0, MM0 ) /* x0 | x0 */ - MOVQ ( MM2, MM5 ) /* x2 | x2 */ - - PUNPCKHDQ ( MM1, MM1 ) /* x1 | x1 */ - PFMUL ( REGOFF(32, ECX), MM2 ) /* x2*m9 | x2*m8 */ - - MOVQ ( MM0, MM3 ) /* x0 | x0 */ - PFMUL ( REGOFF(40, ECX), MM5 ) /* x2*m11 | x2*m10 */ - - MOVQ ( MM1, MM4 ) /* x1 | x1 */ - PFMUL ( REGIND(ECX), MM0 ) /* x0*m1 | x0*m0 */ - - PFADD ( REGOFF(48, ECX), MM2 ) /* x2*m9+m13 | x2*m8+m12 */ - PFMUL ( REGOFF(16, ECX), MM1 ) /* x1*m5 | x1*m4 */ - - PFADD ( REGOFF(56, ECX), MM5 ) /* x2*m11+m15 | x2*m10+m14 */ - PFADD ( MM0, MM1 ) /* x0*m1+x1*m5 | x0*m0+x1*m4 */ - - PFMUL ( REGOFF(8, ECX), MM3 ) /* x0*m3 | x0*m2 */ - PFADD ( MM1, MM2 ) /* r1 | r0 */ - - PFMUL ( REGOFF(24, ECX), MM4 ) /* x1*m7 | x1*m6 */ - ADD_L ( CONST(16), EDX ) /* next output vertex */ - - PFADD ( MM3, MM4 ) /* x0*m3+x1*m7 | x0*m2+x1*m6 */ - MOVQ ( MM2, REGOFF(-16, EDX) ) /* write r0, r1 */ - - PFADD ( MM4, MM5 ) /* r3 | r2 */ - MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPGR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPGR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points3_perspective ) -HIDDEN(_mesa_3dnow_transform_points3_perspective) -GLNAME( _mesa_3dnow_transform_points3_perspective ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPPR_2 ) ) - - PREFETCH ( REGIND(EAX) ) - PREFETCHW ( REGIND(EDX) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVQ ( REGOFF(32, ECX), MM1 ) /* m21 | m20 */ - MOVD ( REGOFF(40, ECX), MM2 ) /* | m22 */ - - MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ - -ALIGNTEXT16 -LLBL( G3TPPR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - PXOR ( MM7, MM7 ) /* 0 | 0 */ - MOVQ ( MM5, MM6 ) /* | x2 */ - - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - PFSUB ( MM5, MM7 ) /* | -x2 */ - - PFMUL ( MM2, MM6 ) /* | x2*m22 */ - PUNPCKLDQ ( MM5, MM5 ) /* x2 | x2 */ - - ADD_L ( CONST(16), EDX ) /* next r */ - PFMUL ( MM1, MM5 ) /* x2*m21 | x2*m20 */ - - PFADD ( MM3, MM6 ) /* | x2*m22+m32 */ - PFADD ( MM4, MM5 ) /* x1*m11+x2*m21 | x0*m00+x2*m20 */ - - MOVQ ( MM5, REGOFF(-16, EDX) ) /* write r0, r1 */ - MOVD ( MM6, REGOFF(-8, EDX) ) /* write r2 */ - - MOVD ( MM7, REGOFF(-4, EDX) ) /* write r3 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPPR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPPR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points3_3d ) -HIDDEN(_mesa_3dnow_transform_points3_3d) -GLNAME( _mesa_3dnow_transform_points3_3d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3R_2 ) ) - - PREFETCH ( REGIND(EAX) ) - PREFETCH ( REGIND(EDX) ) - - MOVD ( REGOFF(8, ECX), MM7 ) /* | m2 */ - PUNPCKLDQ ( REGOFF(24, ECX), MM7 ) /* m6 | m2 */ - - -ALIGNTEXT16 -LLBL( G3TP3R_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - MOVQ ( MM0, MM2 ) /* x1 | x0 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - PUNPCKLDQ ( MM2, MM2 ) /* x0 | x0 */ - MOVQ ( MM0, MM3 ) /* x1 | x0 */ - - PFMUL ( REGIND(ECX), MM2 ) /* x0*m1 | x0*m0 */ - PUNPCKHDQ ( MM3, MM3 ) /* x1 | x1 */ - - MOVQ ( MM1, MM4 ) /* | x2 */ - PFMUL ( REGOFF(16, ECX), MM3 ) /* x1*m5 | x1*m4 */ - - PUNPCKLDQ ( MM4, MM4 ) /* x2 | x2 */ - PFADD ( MM2, MM3 ) /* x0*m1+x1*m5 | x0*m0+x1*m4 */ - - PFMUL ( REGOFF(32, ECX), MM4 ) /* x2*m9 | x2*m8 */ - PFADD ( REGOFF(48, ECX), MM3 ) /* x0*m1+...+m11 | x0*m0+x1*m4+m12 */ - - PFMUL ( MM7, MM0 ) /* x1*m6 | x0*m2 */ - PFADD ( MM4, MM3 ) /* r1 | r0 */ - - PFMUL ( REGOFF(40, ECX), MM1 ) /* | x2*m10 */ - PUNPCKLDQ ( REGOFF(56, ECX), MM1 ) /* m14 | x2*m10 */ - - PFACC ( MM0, MM1 ) - - MOVQ ( MM3, REGOFF(-16, EDX) ) /* write r0, r1 */ - PFACC ( MM1, MM1 ) /* | r2 */ - - MOVD ( MM1, REGOFF(-8, EDX) ) /* write r2 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP3R_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3R_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points3_3d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points3_3d_no_rot) -GLNAME( _mesa_3dnow_transform_points3_3d_no_rot ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3NRR_2 ) ) - - PREFETCH ( REGIND(EAX) ) - PREFETCHW ( REGIND(EDX) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVD ( REGOFF(40, ECX), MM2 ) /* | m22 */ - PUNPCKLDQ ( MM2, MM2 ) /* m22 | m22 */ - - MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ - MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ - - PUNPCKLDQ ( MM3, MM3 ) /* m32 | m32 */ - - -ALIGNTEXT16 -LLBL( G3TP3NRR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCHW ( REGIND(EAX) ) - - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - - PFADD ( MM1, MM4 ) /* x1*m11+m31 | x0*m00+m30 */ - PFMUL ( MM2, MM5 ) /* | x2*m22 */ - - PFADD ( MM3, MM5 ) /* | x2*m22+m32 */ - MOVQ ( MM4, REGIND(EDX) ) /* write r0, r1 */ - - ADD_L ( CONST(16), EDX ) /* next r */ - DEC_L ( ESI ) /* decrement vertex counter */ - - MOVD ( MM5, REGOFF(-8, EDX) ) /* write r2 */ - JNZ ( LLBL( G3TP3NRR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3NRR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points3_2d ) -HIDDEN(_mesa_3dnow_transform_points3_2d) -GLNAME( _mesa_3dnow_transform_points3_2d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2R_3) ) - - PREFETCH ( REGIND(EAX) ) - PREFETCHW ( REGIND(EDX) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ - - MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ - - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP2R_2 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM3 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - MOVQ ( MM3, MM4 ) /* x1 | x0 */ - PFMUL ( MM0, MM3 ) /* x1*m10 | x0*m00 */ - - ADD_L ( CONST(16), EDX ) /* next r */ - PFMUL ( MM1, MM4 ) /* x1*m11 | x0*m01 */ - - PFACC ( MM4, MM3 ) /* x0*m00+x1*m10 | x0*m01+x1*m11 */ - MOVD ( MM5, REGOFF(-8, EDX) ) /* write r2 (=x2) */ - - PFADD ( MM2, MM3 ) /* x0*...*m10+m30 | x0*...*m11+m31 */ - MOVQ ( MM3, REGOFF(-16, EDX) ) /* write r0, r1 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP2R_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2R_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points3_2d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points3_2d_no_rot) -GLNAME( _mesa_3dnow_transform_points3_2d_no_rot ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2NRR_2 ) ) - - PREFETCH ( REGIND(EAX) ) - PREFETCHW ( REGIND(EDX) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ - - -ALIGNTEXT16 -LLBL( G3TP2NRR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - PFADD ( MM1, MM4 ) /* x1*m11+m31 | x0*m00+m30 */ - - MOVQ ( MM4, REGOFF(-16, EDX) ) /* write r0, r1 */ - MOVD ( MM5, REGOFF(-8, EDX) ) /* write r2 (=x2) */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP2NRR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2NRR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points3_identity ) -HIDDEN(_mesa_3dnow_transform_points3_identity) -GLNAME( _mesa_3dnow_transform_points3_identity ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPIR_2 ) ) - - PREFETCHW ( REGIND(EDX) ) - -ALIGNTEXT16 -LLBL( G3TPIR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) - - MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ - MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - ADD_L ( CONST(16), EDX ) /* next r */ - - DEC_L ( ESI ) /* decrement vertex counter */ - MOVQ ( MM0, REGOFF(-16, EDX) ) /* r1 | r0 */ - - MOVD ( MM1, REGOFF(-8, EDX) ) /* | r2 */ - JNZ ( LLBL( G3TPIR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPIR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifdef USE_3DNOW_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FRAME_OFFSET 4 + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points3_general ) +HIDDEN(_mesa_3dnow_transform_points3_general) +GLNAME( _mesa_3dnow_transform_points3_general ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPGR_2 ) ) + + PREFETCHW ( REGIND(EDX) ) + +ALIGNTEXT16 +LLBL( G3TPGR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM2 ) /* | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + MOVQ ( MM0, MM1 ) /* x1 | x0 */ + PUNPCKLDQ ( MM2, MM2 ) /* x2 | x2 */ + + PUNPCKLDQ ( MM0, MM0 ) /* x0 | x0 */ + MOVQ ( MM2, MM5 ) /* x2 | x2 */ + + PUNPCKHDQ ( MM1, MM1 ) /* x1 | x1 */ + PFMUL ( REGOFF(32, ECX), MM2 ) /* x2*m9 | x2*m8 */ + + MOVQ ( MM0, MM3 ) /* x0 | x0 */ + PFMUL ( REGOFF(40, ECX), MM5 ) /* x2*m11 | x2*m10 */ + + MOVQ ( MM1, MM4 ) /* x1 | x1 */ + PFMUL ( REGIND(ECX), MM0 ) /* x0*m1 | x0*m0 */ + + PFADD ( REGOFF(48, ECX), MM2 ) /* x2*m9+m13 | x2*m8+m12 */ + PFMUL ( REGOFF(16, ECX), MM1 ) /* x1*m5 | x1*m4 */ + + PFADD ( REGOFF(56, ECX), MM5 ) /* x2*m11+m15 | x2*m10+m14 */ + PFADD ( MM0, MM1 ) /* x0*m1+x1*m5 | x0*m0+x1*m4 */ + + PFMUL ( REGOFF(8, ECX), MM3 ) /* x0*m3 | x0*m2 */ + PFADD ( MM1, MM2 ) /* r1 | r0 */ + + PFMUL ( REGOFF(24, ECX), MM4 ) /* x1*m7 | x1*m6 */ + ADD_L ( CONST(16), EDX ) /* next output vertex */ + + PFADD ( MM3, MM4 ) /* x0*m3+x1*m7 | x0*m2+x1*m6 */ + MOVQ ( MM2, REGOFF(-16, EDX) ) /* write r0, r1 */ + + PFADD ( MM4, MM5 ) /* r3 | r2 */ + MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPGR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPGR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points3_perspective ) +HIDDEN(_mesa_3dnow_transform_points3_perspective) +GLNAME( _mesa_3dnow_transform_points3_perspective ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPPR_2 ) ) + + PREFETCH ( REGIND(EAX) ) + PREFETCHW ( REGIND(EDX) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVQ ( REGOFF(32, ECX), MM1 ) /* m21 | m20 */ + MOVD ( REGOFF(40, ECX), MM2 ) /* | m22 */ + + MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ + +ALIGNTEXT16 +LLBL( G3TPPR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + PXOR ( MM7, MM7 ) /* 0 | 0 */ + MOVQ ( MM5, MM6 ) /* | x2 */ + + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + PFSUB ( MM5, MM7 ) /* | -x2 */ + + PFMUL ( MM2, MM6 ) /* | x2*m22 */ + PUNPCKLDQ ( MM5, MM5 ) /* x2 | x2 */ + + ADD_L ( CONST(16), EDX ) /* next r */ + PFMUL ( MM1, MM5 ) /* x2*m21 | x2*m20 */ + + PFADD ( MM3, MM6 ) /* | x2*m22+m32 */ + PFADD ( MM4, MM5 ) /* x1*m11+x2*m21 | x0*m00+x2*m20 */ + + MOVQ ( MM5, REGOFF(-16, EDX) ) /* write r0, r1 */ + MOVD ( MM6, REGOFF(-8, EDX) ) /* write r2 */ + + MOVD ( MM7, REGOFF(-4, EDX) ) /* write r3 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPPR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPPR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points3_3d ) +HIDDEN(_mesa_3dnow_transform_points3_3d) +GLNAME( _mesa_3dnow_transform_points3_3d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3R_2 ) ) + + PREFETCH ( REGIND(EAX) ) + PREFETCH ( REGIND(EDX) ) + + MOVD ( REGOFF(8, ECX), MM7 ) /* | m2 */ + PUNPCKLDQ ( REGOFF(24, ECX), MM7 ) /* m6 | m2 */ + + +ALIGNTEXT16 +LLBL( G3TP3R_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + MOVQ ( MM0, MM2 ) /* x1 | x0 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + PUNPCKLDQ ( MM2, MM2 ) /* x0 | x0 */ + MOVQ ( MM0, MM3 ) /* x1 | x0 */ + + PFMUL ( REGIND(ECX), MM2 ) /* x0*m1 | x0*m0 */ + PUNPCKHDQ ( MM3, MM3 ) /* x1 | x1 */ + + MOVQ ( MM1, MM4 ) /* | x2 */ + PFMUL ( REGOFF(16, ECX), MM3 ) /* x1*m5 | x1*m4 */ + + PUNPCKLDQ ( MM4, MM4 ) /* x2 | x2 */ + PFADD ( MM2, MM3 ) /* x0*m1+x1*m5 | x0*m0+x1*m4 */ + + PFMUL ( REGOFF(32, ECX), MM4 ) /* x2*m9 | x2*m8 */ + PFADD ( REGOFF(48, ECX), MM3 ) /* x0*m1+...+m11 | x0*m0+x1*m4+m12 */ + + PFMUL ( MM7, MM0 ) /* x1*m6 | x0*m2 */ + PFADD ( MM4, MM3 ) /* r1 | r0 */ + + PFMUL ( REGOFF(40, ECX), MM1 ) /* | x2*m10 */ + PUNPCKLDQ ( REGOFF(56, ECX), MM1 ) /* m14 | x2*m10 */ + + PFACC ( MM0, MM1 ) + + MOVQ ( MM3, REGOFF(-16, EDX) ) /* write r0, r1 */ + PFACC ( MM1, MM1 ) /* | r2 */ + + MOVD ( MM1, REGOFF(-8, EDX) ) /* write r2 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP3R_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3R_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points3_3d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points3_3d_no_rot) +GLNAME( _mesa_3dnow_transform_points3_3d_no_rot ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3NRR_2 ) ) + + PREFETCH ( REGIND(EAX) ) + PREFETCHW ( REGIND(EDX) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVD ( REGOFF(40, ECX), MM2 ) /* | m22 */ + PUNPCKLDQ ( MM2, MM2 ) /* m22 | m22 */ + + MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ + MOVD ( REGOFF(56, ECX), MM3 ) /* | m32 */ + + PUNPCKLDQ ( MM3, MM3 ) /* m32 | m32 */ + + +ALIGNTEXT16 +LLBL( G3TP3NRR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCHW ( REGIND(EAX) ) + + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + + PFADD ( MM1, MM4 ) /* x1*m11+m31 | x0*m00+m30 */ + PFMUL ( MM2, MM5 ) /* | x2*m22 */ + + PFADD ( MM3, MM5 ) /* | x2*m22+m32 */ + MOVQ ( MM4, REGIND(EDX) ) /* write r0, r1 */ + + ADD_L ( CONST(16), EDX ) /* next r */ + DEC_L ( ESI ) /* decrement vertex counter */ + + MOVD ( MM5, REGOFF(-8, EDX) ) /* write r2 */ + JNZ ( LLBL( G3TP3NRR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3NRR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points3_2d ) +HIDDEN(_mesa_3dnow_transform_points3_2d) +GLNAME( _mesa_3dnow_transform_points3_2d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2R_3) ) + + PREFETCH ( REGIND(EAX) ) + PREFETCHW ( REGIND(EDX) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ + + MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ + + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP2R_2 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM3 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + MOVQ ( MM3, MM4 ) /* x1 | x0 */ + PFMUL ( MM0, MM3 ) /* x1*m10 | x0*m00 */ + + ADD_L ( CONST(16), EDX ) /* next r */ + PFMUL ( MM1, MM4 ) /* x1*m11 | x0*m01 */ + + PFACC ( MM4, MM3 ) /* x0*m00+x1*m10 | x0*m01+x1*m11 */ + MOVD ( MM5, REGOFF(-8, EDX) ) /* write r2 (=x2) */ + + PFADD ( MM2, MM3 ) /* x0*...*m10+m30 | x0*...*m11+m31 */ + MOVQ ( MM3, REGOFF(-16, EDX) ) /* write r0, r1 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP2R_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2R_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points3_2d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points3_2d_no_rot) +GLNAME( _mesa_3dnow_transform_points3_2d_no_rot ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2NRR_2 ) ) + + PREFETCH ( REGIND(EAX) ) + PREFETCHW ( REGIND(EDX) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ + + +ALIGNTEXT16 +LLBL( G3TP2NRR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM5 ) /* | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + PFADD ( MM1, MM4 ) /* x1*m11+m31 | x0*m00+m30 */ + + MOVQ ( MM4, REGOFF(-16, EDX) ) /* write r0, r1 */ + MOVD ( MM5, REGOFF(-8, EDX) ) /* write r2 (=x2) */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP2NRR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2NRR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points3_identity ) +HIDDEN(_mesa_3dnow_transform_points3_identity) +GLNAME( _mesa_3dnow_transform_points3_identity ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(3), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPIR_2 ) ) + + PREFETCHW ( REGIND(EDX) ) + +ALIGNTEXT16 +LLBL( G3TPIR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) + + MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ + MOVD ( REGOFF(8, EAX), MM1 ) /* | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + ADD_L ( CONST(16), EDX ) /* next r */ + + DEC_L ( ESI ) /* decrement vertex counter */ + MOVQ ( MM0, REGOFF(-16, EDX) ) /* r1 | r0 */ + + MOVD ( MM1, REGOFF(-8, EDX) ) /* | r2 */ + JNZ ( LLBL( G3TPIR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPIR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/3dnow_xform4.S b/mesalib/src/mesa/x86/3dnow_xform4.S index 3a9b97160..b2b7c64f2 100644 --- a/mesalib/src/mesa/x86/3dnow_xform4.S +++ b/mesalib/src/mesa/x86/3dnow_xform4.S @@ -1,570 +1,570 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#ifdef USE_3DNOW_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FRAME_OFFSET 4 - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points4_general ) -HIDDEN(_mesa_3dnow_transform_points4_general) -GLNAME( _mesa_3dnow_transform_points4_general ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPGR_2 ) ) - - PREFETCHW ( REGIND(EDX) ) - -ALIGNTEXT16 -LLBL( G3TPGR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ - MOVQ ( REGOFF(8, EAX), MM4 ) /* x3 | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - MOVQ ( MM0, MM2 ) /* x1 | x0 */ - MOVQ ( MM4, MM6 ) /* x3 | x2 */ - - PUNPCKLDQ ( MM0, MM0 ) /* x0 | x0 */ - PUNPCKHDQ ( MM2, MM2 ) /* x1 | x1 */ - - MOVQ ( MM0, MM1 ) /* x0 | x0 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - PFMUL ( REGIND(ECX), MM0 ) /* x0*m1 | x0*m0 */ - MOVQ ( MM2, MM3 ) /* x1 | x1 */ - - PFMUL ( REGOFF(8, ECX), MM1 ) /* x0*m3 | x0*m2 */ - PUNPCKLDQ ( MM4, MM4 ) /* x2 | x2 */ - - PFMUL ( REGOFF(16, ECX), MM2 ) /* x1*m5 | x1*m4 */ - MOVQ ( MM4, MM5 ) /* x2 | x2 */ - - PFMUL ( REGOFF(24, ECX), MM3 ) /* x1*m7 | x1*m6 */ - PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ - - PFMUL ( REGOFF(32, ECX), MM4 ) /* x2*m9 | x2*m8 */ - MOVQ ( MM6, MM7 ) /* x3 | x3 */ - - PFMUL ( REGOFF(40, ECX), MM5 ) /* x2*m11 | x2*m10 */ - PFADD ( MM0, MM2 ) - - PFMUL ( REGOFF(48, ECX), MM6 ) /* x3*m13 | x3*m12 */ - PFADD ( MM1, MM3 ) - - PFMUL ( REGOFF(56, ECX), MM7 ) /* x3*m15 | x3*m14 */ - PFADD ( MM4, MM6 ) - - PFADD ( MM5, MM7 ) - PFADD ( MM2, MM6 ) - - PFADD ( MM3, MM7 ) - MOVQ ( MM6, REGOFF(-16, EDX) ) - - MOVQ ( MM7, REGOFF(-8, EDX) ) - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPGR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPGR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points4_perspective ) -HIDDEN(_mesa_3dnow_transform_points4_perspective) -GLNAME( _mesa_3dnow_transform_points4_perspective ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPPR_2 ) ) - - PREFETCH ( REGIND(EAX) ) - PREFETCHW ( REGIND(EDX) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVD ( REGOFF(40, ECX), MM1 ) /* | m22 */ - PUNPCKLDQ ( REGOFF(56, ECX), MM1 ) /* m32 | m22 */ - - MOVQ ( REGOFF(32, ECX), MM2 ) /* m21 | m20 */ - PXOR ( MM7, MM7 ) /* 0 | 0 */ - -ALIGNTEXT16 -LLBL( G3TPPR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ - MOVD ( REGOFF(8, EAX), MM3 ) /* | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGOFF(32, EAX) ) /* hopefully stride is zero */ - - MOVQ ( MM5, MM6 ) /* x3 | x2 */ - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - - PUNPCKLDQ ( MM5, MM5 ) /* x2 | x2 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - PFMUL ( MM2, MM5 ) /* x2*m21 | x2*m20 */ - PFSUBR ( MM7, MM3 ) /* | -x2 */ - - PFMUL ( MM1, MM6 ) /* x3*m32 | x2*m22 */ - PFADD ( MM4, MM5 ) /* x1*m11+x2*m21 | x0*m00+x2*m20 */ - - PFACC ( MM3, MM6 ) /* -x2 | x2*m22+x3*m32 */ - MOVQ ( MM5, REGOFF(-16, EDX) ) /* write r0, r1 */ - - MOVQ ( MM6, REGOFF(-8, EDX) ) /* write r2, r3 */ - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TPPR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPPR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points4_3d ) -HIDDEN(_mesa_3dnow_transform_points4_3d) -GLNAME( _mesa_3dnow_transform_points4_3d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3R_2 ) ) - - MOVD ( REGOFF(8, ECX), MM6 ) /* | m2 */ - PUNPCKLDQ ( REGOFF(24, ECX), MM6 ) /* m6 | m2 */ - - MOVD ( REGOFF(40, ECX), MM7 ) /* | m10 */ - PUNPCKLDQ ( REGOFF(56, ECX), MM7 ) /* m14 | m10 */ - -ALIGNTEXT16 -LLBL( G3TP3R_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - PREFETCH ( REGOFF(32, EAX) ) /* hopefully array is tightly packed */ - - MOVQ ( REGIND(EAX), MM2 ) /* x1 | x0 */ - MOVQ ( REGOFF(8, EAX), MM3 ) /* x3 | x2 */ - - MOVQ ( MM2, MM0 ) /* x1 | x0 */ - MOVQ ( MM3, MM4 ) /* x3 | x2 */ - - MOVQ ( MM0, MM1 ) /* x1 | x0 */ - MOVQ ( MM4, MM5 ) /* x3 | x2 */ - - PUNPCKLDQ ( MM0, MM0 ) /* x0 | x0 */ - PUNPCKHDQ ( MM1, MM1 ) /* x1 | x1 */ - - PFMUL ( REGIND(ECX), MM0 ) /* x0*m1 | x0*m0 */ - PUNPCKLDQ ( MM3, MM3 ) /* x2 | x2 */ - - PFMUL ( REGOFF(16, ECX), MM1 ) /* x1*m5 | x1*m4 */ - PUNPCKHDQ ( MM4, MM4 ) /* x3 | x3 */ - - PFMUL ( MM6, MM2 ) /* x1*m6 | x0*m2 */ - PFADD ( MM0, MM1 ) /* x0*m1+x1*m5 | x0*m0+x1*m4 */ - - PFMUL ( REGOFF(32, ECX), MM3 ) /* x2*m9 | x2*m8 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - PFMUL ( REGOFF(48, ECX), MM4 ) /* x3*m13 | x3*m12 */ - PFADD ( MM1, MM3 ) /* x0*m1+..+x2*m9 | x0*m0+...+x2*m8 */ - - PFMUL ( MM7, MM5 ) /* x3*m14 | x2*m10 */ - PFADD ( MM3, MM4 ) /* r1 | r0 */ - - PFACC ( MM2, MM5 ) /* x0*m2+x1*m6 | x2*m10+x3*m14 */ - MOVD ( REGOFF(12, EAX), MM0 ) /* | x3 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PFACC ( MM0, MM5 ) /* r3 | r2 */ - - MOVQ ( MM4, REGOFF(-16, EDX) ) /* write r0, r1 */ - MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP3R_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3R_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points4_3d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points4_3d_no_rot) -GLNAME( _mesa_3dnow_transform_points4_3d_no_rot ): - - PUSH_L ( ESI ) - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP3NRR_2 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVD ( REGOFF(40, ECX), MM2 ) /* | m22 */ - PUNPCKLDQ ( REGOFF(56, ECX), MM2 ) /* m32 | m22 */ - - MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP3NRR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ - MOVD ( REGOFF(12, EAX), MM7 ) /* | x3 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGOFF(32, EAX) ) /* hopefully stride is zero */ - - MOVQ ( MM5, MM6 ) /* x3 | x2 */ - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - - PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ - PFMUL ( MM2, MM5 ) /* x3*m32 | x2*m22 */ - - PFMUL ( MM1, MM6 ) /* x3*m31 | x3*m30 */ - PFACC ( MM7, MM5 ) /* x3 | x2*m22+x3*m32 */ - - PFADD ( MM6, MM4 ) /* x1*m11+x3*m31 | x0*m00+x3*m30 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - MOVQ ( MM4, REGOFF(-16, EDX) ) /* write r0, r1 */ - MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP3NRR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP3NRR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points4_2d ) -HIDDEN(_mesa_3dnow_transform_points4_2d) -GLNAME( _mesa_3dnow_transform_points4_2d ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2R_2 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ - - MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ - - MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP2R_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM3 ) /* x1 | x0 */ - MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - MOVQ ( MM3, MM4 ) /* x1 | x0 */ - MOVQ ( MM5, MM6 ) /* x3 | x2 */ - - PFMUL ( MM1, MM4 ) /* x1*m11 | x0*m01 */ - PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ - - PFMUL ( MM0, MM3 ) /* x1*m10 | x0*m00 */ - ADD_L ( CONST(16), EDX ) /* next r */ - - PFACC ( MM4, MM3 ) /* x0*m01+x1*m11 | x0*m00+x1*m10 */ - PFMUL ( MM2, MM6 ) /* x3*m31 | x3*m30 */ - - PFADD ( MM6, MM3 ) /* r1 | r0 */ - MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ - - MOVQ ( MM3, REGOFF(-16, EDX) ) /* write r0, r1 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TP2R_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2R_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points4_2d_no_rot ) -HIDDEN(_mesa_3dnow_transform_points4_2d_no_rot) -GLNAME( _mesa_3dnow_transform_points4_2d_no_rot ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TP2NRR_3 ) ) - - MOVD ( REGIND(ECX), MM0 ) /* | m00 */ - PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ - - MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ - -ALIGNTEXT16 -LLBL( G3TP2NRR_2 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ - MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ - MOVQ ( MM5, MM6 ) /* x3 | x2 */ - - ADD_L ( CONST(16), EDX ) /* next r */ - PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ - - PFMUL ( MM1, MM6 ) /* x3*m31 | x3*m30 */ - PFADD ( MM4, MM6 ) /* x1*m11+x3*m31 | x0*m00+x3*m30 */ - - MOVQ ( MM6, REGOFF(-16, EDX) ) /* write r0, r1 */ - MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - - JNZ ( LLBL( G3TP2NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TP2NRR_3 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_3dnow_transform_points4_identity ) -HIDDEN(_mesa_3dnow_transform_points4_identity) -GLNAME( _mesa_3dnow_transform_points4_identity ): - - PUSH_L ( ESI ) - - MOV_L ( ARG_DEST, ECX ) - MOV_L ( ARG_MATRIX, ESI ) - MOV_L ( ARG_SOURCE, EAX ) - MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) - OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) - MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) - MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) - - PUSH_L ( EDI ) - - MOV_L ( REGOFF(V4F_START, ECX), EDX ) - MOV_L ( ESI, ECX ) - MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) - MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) - MOV_L ( REGOFF(V4F_START, EAX), EAX ) - - TEST_L ( ESI, ESI ) - JZ ( LLBL( G3TPIR_2 ) ) - -ALIGNTEXT16 -LLBL( G3TPIR_1 ): - - PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ - - MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ - MOVQ ( REGOFF(8, EAX), MM1 ) /* x3 | x2 */ - - ADD_L ( EDI, EAX ) /* next vertex */ - PREFETCH ( REGIND(EAX) ) - - ADD_L ( CONST(16), EDX ) /* next r */ - MOVQ ( MM0, REGOFF(-16, EDX) ) /* r1 | r0 */ - - MOVQ ( MM1, REGOFF(-8, EDX) ) /* r3 | r2 */ - - DEC_L ( ESI ) /* decrement vertex counter */ - JNZ ( LLBL( G3TPIR_1 ) ) /* cnt > 0 ? -> process next vertex */ - -LLBL( G3TPIR_2 ): - - FEMMS - POP_L ( EDI ) - POP_L ( ESI ) - RET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifdef USE_3DNOW_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FRAME_OFFSET 4 + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points4_general ) +HIDDEN(_mesa_3dnow_transform_points4_general) +GLNAME( _mesa_3dnow_transform_points4_general ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPGR_2 ) ) + + PREFETCHW ( REGIND(EDX) ) + +ALIGNTEXT16 +LLBL( G3TPGR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ + MOVQ ( REGOFF(8, EAX), MM4 ) /* x3 | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + MOVQ ( MM0, MM2 ) /* x1 | x0 */ + MOVQ ( MM4, MM6 ) /* x3 | x2 */ + + PUNPCKLDQ ( MM0, MM0 ) /* x0 | x0 */ + PUNPCKHDQ ( MM2, MM2 ) /* x1 | x1 */ + + MOVQ ( MM0, MM1 ) /* x0 | x0 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + PFMUL ( REGIND(ECX), MM0 ) /* x0*m1 | x0*m0 */ + MOVQ ( MM2, MM3 ) /* x1 | x1 */ + + PFMUL ( REGOFF(8, ECX), MM1 ) /* x0*m3 | x0*m2 */ + PUNPCKLDQ ( MM4, MM4 ) /* x2 | x2 */ + + PFMUL ( REGOFF(16, ECX), MM2 ) /* x1*m5 | x1*m4 */ + MOVQ ( MM4, MM5 ) /* x2 | x2 */ + + PFMUL ( REGOFF(24, ECX), MM3 ) /* x1*m7 | x1*m6 */ + PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ + + PFMUL ( REGOFF(32, ECX), MM4 ) /* x2*m9 | x2*m8 */ + MOVQ ( MM6, MM7 ) /* x3 | x3 */ + + PFMUL ( REGOFF(40, ECX), MM5 ) /* x2*m11 | x2*m10 */ + PFADD ( MM0, MM2 ) + + PFMUL ( REGOFF(48, ECX), MM6 ) /* x3*m13 | x3*m12 */ + PFADD ( MM1, MM3 ) + + PFMUL ( REGOFF(56, ECX), MM7 ) /* x3*m15 | x3*m14 */ + PFADD ( MM4, MM6 ) + + PFADD ( MM5, MM7 ) + PFADD ( MM2, MM6 ) + + PFADD ( MM3, MM7 ) + MOVQ ( MM6, REGOFF(-16, EDX) ) + + MOVQ ( MM7, REGOFF(-8, EDX) ) + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPGR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPGR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points4_perspective ) +HIDDEN(_mesa_3dnow_transform_points4_perspective) +GLNAME( _mesa_3dnow_transform_points4_perspective ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPPR_2 ) ) + + PREFETCH ( REGIND(EAX) ) + PREFETCHW ( REGIND(EDX) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVD ( REGOFF(40, ECX), MM1 ) /* | m22 */ + PUNPCKLDQ ( REGOFF(56, ECX), MM1 ) /* m32 | m22 */ + + MOVQ ( REGOFF(32, ECX), MM2 ) /* m21 | m20 */ + PXOR ( MM7, MM7 ) /* 0 | 0 */ + +ALIGNTEXT16 +LLBL( G3TPPR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ + MOVD ( REGOFF(8, EAX), MM3 ) /* | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGOFF(32, EAX) ) /* hopefully stride is zero */ + + MOVQ ( MM5, MM6 ) /* x3 | x2 */ + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + + PUNPCKLDQ ( MM5, MM5 ) /* x2 | x2 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + PFMUL ( MM2, MM5 ) /* x2*m21 | x2*m20 */ + PFSUBR ( MM7, MM3 ) /* | -x2 */ + + PFMUL ( MM1, MM6 ) /* x3*m32 | x2*m22 */ + PFADD ( MM4, MM5 ) /* x1*m11+x2*m21 | x0*m00+x2*m20 */ + + PFACC ( MM3, MM6 ) /* -x2 | x2*m22+x3*m32 */ + MOVQ ( MM5, REGOFF(-16, EDX) ) /* write r0, r1 */ + + MOVQ ( MM6, REGOFF(-8, EDX) ) /* write r2, r3 */ + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TPPR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPPR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points4_3d ) +HIDDEN(_mesa_3dnow_transform_points4_3d) +GLNAME( _mesa_3dnow_transform_points4_3d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3R_2 ) ) + + MOVD ( REGOFF(8, ECX), MM6 ) /* | m2 */ + PUNPCKLDQ ( REGOFF(24, ECX), MM6 ) /* m6 | m2 */ + + MOVD ( REGOFF(40, ECX), MM7 ) /* | m10 */ + PUNPCKLDQ ( REGOFF(56, ECX), MM7 ) /* m14 | m10 */ + +ALIGNTEXT16 +LLBL( G3TP3R_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + PREFETCH ( REGOFF(32, EAX) ) /* hopefully array is tightly packed */ + + MOVQ ( REGIND(EAX), MM2 ) /* x1 | x0 */ + MOVQ ( REGOFF(8, EAX), MM3 ) /* x3 | x2 */ + + MOVQ ( MM2, MM0 ) /* x1 | x0 */ + MOVQ ( MM3, MM4 ) /* x3 | x2 */ + + MOVQ ( MM0, MM1 ) /* x1 | x0 */ + MOVQ ( MM4, MM5 ) /* x3 | x2 */ + + PUNPCKLDQ ( MM0, MM0 ) /* x0 | x0 */ + PUNPCKHDQ ( MM1, MM1 ) /* x1 | x1 */ + + PFMUL ( REGIND(ECX), MM0 ) /* x0*m1 | x0*m0 */ + PUNPCKLDQ ( MM3, MM3 ) /* x2 | x2 */ + + PFMUL ( REGOFF(16, ECX), MM1 ) /* x1*m5 | x1*m4 */ + PUNPCKHDQ ( MM4, MM4 ) /* x3 | x3 */ + + PFMUL ( MM6, MM2 ) /* x1*m6 | x0*m2 */ + PFADD ( MM0, MM1 ) /* x0*m1+x1*m5 | x0*m0+x1*m4 */ + + PFMUL ( REGOFF(32, ECX), MM3 ) /* x2*m9 | x2*m8 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + PFMUL ( REGOFF(48, ECX), MM4 ) /* x3*m13 | x3*m12 */ + PFADD ( MM1, MM3 ) /* x0*m1+..+x2*m9 | x0*m0+...+x2*m8 */ + + PFMUL ( MM7, MM5 ) /* x3*m14 | x2*m10 */ + PFADD ( MM3, MM4 ) /* r1 | r0 */ + + PFACC ( MM2, MM5 ) /* x0*m2+x1*m6 | x2*m10+x3*m14 */ + MOVD ( REGOFF(12, EAX), MM0 ) /* | x3 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PFACC ( MM0, MM5 ) /* r3 | r2 */ + + MOVQ ( MM4, REGOFF(-16, EDX) ) /* write r0, r1 */ + MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP3R_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3R_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points4_3d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points4_3d_no_rot) +GLNAME( _mesa_3dnow_transform_points4_3d_no_rot ): + + PUSH_L ( ESI ) + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP3NRR_2 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVD ( REGOFF(40, ECX), MM2 ) /* | m22 */ + PUNPCKLDQ ( REGOFF(56, ECX), MM2 ) /* m32 | m22 */ + + MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP3NRR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ + MOVD ( REGOFF(12, EAX), MM7 ) /* | x3 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGOFF(32, EAX) ) /* hopefully stride is zero */ + + MOVQ ( MM5, MM6 ) /* x3 | x2 */ + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + + PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ + PFMUL ( MM2, MM5 ) /* x3*m32 | x2*m22 */ + + PFMUL ( MM1, MM6 ) /* x3*m31 | x3*m30 */ + PFACC ( MM7, MM5 ) /* x3 | x2*m22+x3*m32 */ + + PFADD ( MM6, MM4 ) /* x1*m11+x3*m31 | x0*m00+x3*m30 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + MOVQ ( MM4, REGOFF(-16, EDX) ) /* write r0, r1 */ + MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP3NRR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP3NRR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points4_2d ) +HIDDEN(_mesa_3dnow_transform_points4_2d) +GLNAME( _mesa_3dnow_transform_points4_2d ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2R_2 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(16, ECX), MM0 ) /* m10 | m00 */ + + MOVD ( REGOFF(4, ECX), MM1 ) /* | m01 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM1 ) /* m11 | m01 */ + + MOVQ ( REGOFF(48, ECX), MM2 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP2R_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM3 ) /* x1 | x0 */ + MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + MOVQ ( MM3, MM4 ) /* x1 | x0 */ + MOVQ ( MM5, MM6 ) /* x3 | x2 */ + + PFMUL ( MM1, MM4 ) /* x1*m11 | x0*m01 */ + PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ + + PFMUL ( MM0, MM3 ) /* x1*m10 | x0*m00 */ + ADD_L ( CONST(16), EDX ) /* next r */ + + PFACC ( MM4, MM3 ) /* x0*m01+x1*m11 | x0*m00+x1*m10 */ + PFMUL ( MM2, MM6 ) /* x3*m31 | x3*m30 */ + + PFADD ( MM6, MM3 ) /* r1 | r0 */ + MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ + + MOVQ ( MM3, REGOFF(-16, EDX) ) /* write r0, r1 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TP2R_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2R_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points4_2d_no_rot ) +HIDDEN(_mesa_3dnow_transform_points4_2d_no_rot) +GLNAME( _mesa_3dnow_transform_points4_2d_no_rot ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TP2NRR_3 ) ) + + MOVD ( REGIND(ECX), MM0 ) /* | m00 */ + PUNPCKLDQ ( REGOFF(20, ECX), MM0 ) /* m11 | m00 */ + + MOVQ ( REGOFF(48, ECX), MM1 ) /* m31 | m30 */ + +ALIGNTEXT16 +LLBL( G3TP2NRR_2 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM4 ) /* x1 | x0 */ + MOVQ ( REGOFF(8, EAX), MM5 ) /* x3 | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + PFMUL ( MM0, MM4 ) /* x1*m11 | x0*m00 */ + MOVQ ( MM5, MM6 ) /* x3 | x2 */ + + ADD_L ( CONST(16), EDX ) /* next r */ + PUNPCKHDQ ( MM6, MM6 ) /* x3 | x3 */ + + PFMUL ( MM1, MM6 ) /* x3*m31 | x3*m30 */ + PFADD ( MM4, MM6 ) /* x1*m11+x3*m31 | x0*m00+x3*m30 */ + + MOVQ ( MM6, REGOFF(-16, EDX) ) /* write r0, r1 */ + MOVQ ( MM5, REGOFF(-8, EDX) ) /* write r2, r3 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + + JNZ ( LLBL( G3TP2NRR_2 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TP2NRR_3 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_3dnow_transform_points4_identity ) +HIDDEN(_mesa_3dnow_transform_points4_identity) +GLNAME( _mesa_3dnow_transform_points4_identity ): + + PUSH_L ( ESI ) + + MOV_L ( ARG_DEST, ECX ) + MOV_L ( ARG_MATRIX, ESI ) + MOV_L ( ARG_SOURCE, EAX ) + MOV_L ( CONST(4), REGOFF(V4F_SIZE, ECX) ) + OR_B ( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, ECX) ) + MOV_L ( REGOFF(V4F_COUNT, EAX), EDX ) + MOV_L ( EDX, REGOFF(V4F_COUNT, ECX) ) + + PUSH_L ( EDI ) + + MOV_L ( REGOFF(V4F_START, ECX), EDX ) + MOV_L ( ESI, ECX ) + MOV_L ( REGOFF(V4F_COUNT, EAX), ESI ) + MOV_L ( REGOFF(V4F_STRIDE, EAX), EDI ) + MOV_L ( REGOFF(V4F_START, EAX), EAX ) + + TEST_L ( ESI, ESI ) + JZ ( LLBL( G3TPIR_2 ) ) + +ALIGNTEXT16 +LLBL( G3TPIR_1 ): + + PREFETCHW ( REGOFF(32, EDX) ) /* prefetch 2 vertices ahead */ + + MOVQ ( REGIND(EAX), MM0 ) /* x1 | x0 */ + MOVQ ( REGOFF(8, EAX), MM1 ) /* x3 | x2 */ + + ADD_L ( EDI, EAX ) /* next vertex */ + PREFETCH ( REGIND(EAX) ) + + ADD_L ( CONST(16), EDX ) /* next r */ + MOVQ ( MM0, REGOFF(-16, EDX) ) /* r1 | r0 */ + + MOVQ ( MM1, REGOFF(-8, EDX) ) /* r3 | r2 */ + + DEC_L ( ESI ) /* decrement vertex counter */ + JNZ ( LLBL( G3TPIR_1 ) ) /* cnt > 0 ? -> process next vertex */ + +LLBL( G3TPIR_2 ): + + FEMMS + POP_L ( EDI ) + POP_L ( ESI ) + RET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/Makefile b/mesalib/src/mesa/x86/Makefile index 8cbe26015..9716dc27f 100644 --- a/mesalib/src/mesa/x86/Makefile +++ b/mesalib/src/mesa/x86/Makefile @@ -1,46 +1,46 @@ -# src/mesa/x86/Makefile - -TOP = ../../.. -include $(TOP)/configs/current - - -INCLUDE_DIRS = \ - -I$(TOP)/include/GL \ - -I$(TOP)/include \ - -I$(TOP)/src/mapi \ - -I.. \ - -I../main \ - -I../math \ - -I../tnl - - -default: gen_matypes matypes.h - -clean: - -rm -f matypes.h gen_matypes - - -gen_matypes: gen_matypes.c - $(HOST_CC) $(ARCH_FLAGS) $(INCLUDE_DIRS) $(HOST_CFLAGS) gen_matypes.c -o gen_matypes - -# need some special rules here, unfortunately -matypes.h: ../main/mtypes.h ../tnl/t_context.h gen_matypes - ./gen_matypes > matypes.h - -common_x86_asm.o: matypes.h -3dnow_normal.o: matypes.h -3dnow_xform1.o: matypes.h -3dnow_xform2.o: matypes.h -3dnow_xform3.o: matypes.h -3dnow_xform4.o: matypes.h -mmx_blend.o: matypes.h -sse_normal.o: matypes.h -sse_xform1.o: matypes.h -sse_xform2.o: matypes.h -sse_xform3.o: matypes.h -sse_xform4.o: matypes.h -x86_cliptest.o: matypes.h -x86_xform2.o: matypes.h -x86_xform3.o: matypes.h -x86_xform4.o: matypes.h - +# src/mesa/x86/Makefile + +TOP = ../../.. +include $(TOP)/configs/current + + +INCLUDE_DIRS = \ + -I$(TOP)/include/GL \ + -I$(TOP)/include \ + -I$(TOP)/src/mapi \ + -I.. \ + -I../main \ + -I../math \ + -I../tnl + + +default: gen_matypes matypes.h + +clean: + -rm -f matypes.h gen_matypes + + +gen_matypes: gen_matypes.c + $(HOST_CC) $(ARCH_FLAGS) $(INCLUDE_DIRS) $(HOST_CFLAGS) gen_matypes.c -o gen_matypes + +# need some special rules here, unfortunately +matypes.h: ../main/mtypes.h ../tnl/t_context.h gen_matypes + ./gen_matypes > matypes.h + +common_x86_asm.o: matypes.h +3dnow_normal.o: matypes.h +3dnow_xform1.o: matypes.h +3dnow_xform2.o: matypes.h +3dnow_xform3.o: matypes.h +3dnow_xform4.o: matypes.h +mmx_blend.o: matypes.h +sse_normal.o: matypes.h +sse_xform1.o: matypes.h +sse_xform2.o: matypes.h +sse_xform3.o: matypes.h +sse_xform4.o: matypes.h +x86_cliptest.o: matypes.h +x86_xform2.o: matypes.h +x86_xform3.o: matypes.h +x86_xform4.o: matypes.h + diff --git a/mesalib/src/mesa/x86/assyntax.h b/mesalib/src/mesa/x86/assyntax.h index 1e3ea66e2..4a41812f6 100644 --- a/mesalib/src/mesa/x86/assyntax.h +++ b/mesalib/src/mesa/x86/assyntax.h @@ -1,1747 +1,1747 @@ - -#ifndef __ASSYNTAX_H__ -#define __ASSYNTAX_H__ - -/* - * Copyright 1992 Vrije Universiteit, The Netherlands - * - * Permission to use, copy, modify, and distribute this software and its - * documentation for any purpose and without fee is hereby granted, provided - * that the above copyright notice appear in all copies and that both that - * copyright notice and this permission notice appear in supporting - * documentation, and that the name of the Vrije Universiteit not be used in - * advertising or publicity pertaining to distribution of the software without - * specific, written prior permission. The Vrije Universiteit makes no - * representations about the suitability of this software for any purpose. - * It is provided "as is" without express or implied warranty. - * - * The Vrije Universiteit DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS - * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, - * IN NO EVENT SHALL The Vrije Universiteit BE LIABLE FOR ANY SPECIAL, - * INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM - * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE - * OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR - * PERFORMANCE OF THIS SOFTWARE. - */ - -/* - * assyntax.h - * - * Select the syntax appropriate to the 386 assembler being used - * To add support for more assemblers add more columns to the CHOICE - * macro. Note that register names must also have uppercase names - * to avoid macro recursion. e.g., #define ah %ah recurses! - * - * NB 1. Some of the macros for certain assemblers imply that the code is to - * run in protected mode!! Caveat emptor. - * - * NB 2. 486 specific instructions are not included. This is to discourage - * their accidental use in code that is intended to run on 386 and 486 - * systems. - * - * Supported assemblers: - * - * (a) AT&T SysVr4 as(1): define ATT_ASSEMBLER - * (b) GNU Assembler gas: define GNU_ASSEMBLER (default) - * (c) Amsterdam Compiler kit: define ACK_ASSEMBLER - * (d) The Netwide Assembler: define NASM_ASSEMBLER - * (e) Microsoft Assembler: define MASM_ASSEMBLER (UNTESTED!) - * - * The following naming conventions have been used to identify the various - * data types: - * _SR = segment register version - * Integer: - * _Q = quadword = 64 bits - * _L = long = 32 bits - * _W = short = 16 bits - * _B = byte = 8 bits - * Floating-point: - * _X = m80real = 80 bits - * _D = double = 64 bits - * _S = single = 32 bits - * - * Author: Gregory J. Sharp, Sept 1992 - * Vrije Universiteit, Amsterdam, The Netherlands - * - * [support for Intel syntax added by Josh Vanderhoof, 1999] - */ - -#if !(defined(NASM_ASSEMBLER) || defined(MASM_ASSEMBLER)) - -/* Default to ATT_ASSEMBLER when SVR4 or SYSV are defined */ -#if (defined(SVR4) || defined(SYSV)) && !defined(GNU_ASSEMBLER) -#define ATT_ASSEMBLER -#endif - -#if !defined(ATT_ASSEMBLER) && !defined(GNU_ASSEMBLER) && !defined(ACK_ASSEMBLER) -#define GNU_ASSEMBLER -#endif - -#if (defined(__STDC__) && !defined(UNIXCPP)) || (defined (sun) && defined (i386) && defined (SVR4) && defined (__STDC__) && !defined (__GNUC__)) -#define CONCAT(x, y) x ## y -#define CONCAT3(x, y, z) x ## y ## z -#else -#define CONCAT(x, y) x/**/y -#define CONCAT3(x, y, z) x/**/y/**/z -#endif - -#ifdef ACK_ASSEMBLER - -/* Assume we write code for 32-bit protected mode! */ - -/* Redefine register names for GAS & AT&T assemblers */ -#define AL al -#define AH ah -#define AX ax -#define EAX ax -#define BL bl -#define BH bh -#define BX bx -#define EBX bx -#define CL cl -#define CH ch -#define CX cx -#define ECX cx -#define DL dl -#define DH dh -#define DX dx -#define EDX dx -#define BP bp -#define EBP bp -#define SI si -#define ESI si -#define DI di -#define EDI di -#define SP sp -#define ESP sp -#define CS cs -#define SS ss -#define DS ds -#define ES es -#define FS fs -#define GS gs -/* Control Registers */ -#define CR0 cr0 -#define CR1 cr1 -#define CR2 cr2 -#define CR3 cr3 -/* Debug Registers */ -#define DR0 dr0 -#define DR1 dr1 -#define DR2 dr2 -#define DR3 dr3 -#define DR4 dr4 -#define DR5 dr5 -#define DR6 dr6 -#define DR7 dr7 -/* Floating-point Stack */ -#define ST st - -#define AS_BEGIN .sect .text; .sect .rom; .sect .data; .sect .bss; .sect .text - - -#define _WTOG o16 /* word toggle for _W instructions */ -#define _LTOG /* long toggle for _L instructions */ -#define ADDR_TOGGLE a16 -#define OPSZ_TOGGLE o16 -#define USE16 .use16 -#define USE32 .use32 - -#define CHOICE(a,b,c) c - -#else /* AT&T or GAS */ - -/* Redefine register names for GAS & AT&T assemblers */ -#define AL %al -#define AH %ah -#define AX %ax -#define EAX %eax -#define BL %bl -#define BH %bh -#define BX %bx -#define EBX %ebx -#define CL %cl -#define CH %ch -#define CX %cx -#define ECX %ecx -#define DL %dl -#define DH %dh -#define DX %dx -#define EDX %edx -#define BP %bp -#define EBP %ebp -#define SI %si -#define ESI %esi -#define DI %di -#define EDI %edi -#define SP %sp -#define ESP %esp -#define CS %cs -#define SS %ss -#define DS %ds -#define ES %es -#define FS %fs -#define GS %gs -/* Control Registers */ -#define CR0 %cr0 -#define CR1 %cr1 -#define CR2 %cr2 -#define CR3 %cr3 -/* Debug Registers */ -#define DR0 %db0 -#define DR1 %db1 -#define DR2 %db2 -#define DR3 %db3 -#define DR4 %db4 -#define DR5 %db5 -#define DR6 %db6 -#define DR7 %db7 -/* Floating-point Stack */ -#define _STX0 %st(0) -#define _STX1 %st(1) -#define _STX2 %st(2) -#define _STX3 %st(3) -#define _STX4 %st(4) -#define _STX5 %st(5) -#define _STX6 %st(6) -#define _STX7 %st(7) -#define ST(x) CONCAT(_STX,x) -#ifdef GNU_ASSEMBLER -#define ST0 %st(0) -#else -#define ST0 %st -#endif -/* MMX Registers */ -#define MM0 %mm0 -#define MM1 %mm1 -#define MM2 %mm2 -#define MM3 %mm3 -#define MM4 %mm4 -#define MM5 %mm5 -#define MM6 %mm6 -#define MM7 %mm7 -/* SSE Registers */ -#define XMM0 %xmm0 -#define XMM1 %xmm1 -#define XMM2 %xmm2 -#define XMM3 %xmm3 -#define XMM4 %xmm4 -#define XMM5 %xmm5 -#define XMM6 %xmm6 -#define XMM7 %xmm7 - -#define AS_BEGIN -#define USE16 -#define USE32 - -#ifdef GNU_ASSEMBLER - -#define ADDR_TOGGLE aword -#define OPSZ_TOGGLE word - -#define CHOICE(a,b,c) b - -#else -/* - * AT&T ASSEMBLER SYNTAX - * ********************* - */ -#define CHOICE(a,b,c) a - -#define ADDR_TOGGLE addr16 -#define OPSZ_TOGGLE data16 - -#endif /* GNU_ASSEMBLER */ -#endif /* ACK_ASSEMBLER */ - - -#if defined(__QNX__) || defined(Lynx) || (defined(SYSV) || defined(SVR4)) && !defined(ACK_ASSEMBLER) || defined(__ELF__) || defined(__GNU__) || defined(__GNUC__) && !defined(__DJGPP__) && !defined(__MINGW32__) -#define GLNAME(a) a -#else -#define GLNAME(a) CONCAT(_,a) -#endif - - - /****************************************/ - /* */ - /* Select the various choices */ - /* */ - /****************************************/ - - -/* Redefine assembler directives */ -/*********************************/ -#define GLOBL CHOICE(.globl, .globl, .extern) -#define GLOBAL GLOBL -#define EXTERN GLOBL -#ifndef __AOUT__ -#define ALIGNTEXT32 CHOICE(.align 32, .balign 32, .align 32) -#define ALIGNTEXT16 CHOICE(.align 16, .balign 16, .align 16) -#define ALIGNTEXT8 CHOICE(.align 8, .balign 8, .align 8) -#define ALIGNTEXT4 CHOICE(.align 4, .balign 4, .align 4) -#define ALIGNTEXT2 CHOICE(.align 2, .balign 2, .align 2) -/* ALIGNTEXT4ifNOP is the same as ALIGNTEXT4, but only if the space is - * guaranteed to be filled with NOPs. Otherwise it does nothing. - */ -#define ALIGNTEXT32ifNOP CHOICE(.align 32, .balign ARG2(32,0x90), /*can't do it*/) -#define ALIGNTEXT16ifNOP CHOICE(.align 16, .balign ARG2(16,0x90), /*can't do it*/) -#define ALIGNTEXT8ifNOP CHOICE(.align 8, .balign ARG2(8,0x90), /*can't do it*/) -#define ALIGNTEXT4ifNOP CHOICE(.align 4, .balign ARG2(4,0x90), /*can't do it*/) -#define ALIGNDATA32 CHOICE(.align 32, .balign ARG2(32,0x0), .align 32) -#define ALIGNDATA16 CHOICE(.align 16, .balign ARG2(16,0x0), .align 16) -#define ALIGNDATA8 CHOICE(.align 8, .balign ARG2(8,0x0), .align 8) -#define ALIGNDATA4 CHOICE(.align 4, .balign ARG2(4,0x0), .align 4) -#define ALIGNDATA2 CHOICE(.align 2, .balign ARG2(2,0x0), .align 2) -#else -/* 'as -aout' on FreeBSD doesn't have .balign */ -#define ALIGNTEXT32 CHOICE(.align 32, .align ARG2(5,0x90), .align 32) -#define ALIGNTEXT16 CHOICE(.align 16, .align ARG2(4,0x90), .align 16) -#define ALIGNTEXT8 CHOICE(.align 8, .align ARG2(3,0x90), .align 8) -#define ALIGNTEXT4 CHOICE(.align 4, .align ARG2(2,0x90), .align 4) -#define ALIGNTEXT2 CHOICE(.align 2, .align ARG2(1,0x90), .align 2) -/* ALIGNTEXT4ifNOP is the same as ALIGNTEXT4, but only if the space is - * guaranteed to be filled with NOPs. Otherwise it does nothing. - */ -#define ALIGNTEXT32ifNOP CHOICE(.align 32, .align ARG2(5,0x90), /*can't do it*/) -#define ALIGNTEXT16ifNOP CHOICE(.align 16, .align ARG2(4,0x90), /*can't do it*/) -#define ALIGNTEXT8ifNOP CHOICE(.align 8, .align ARG2(3,0x90), /*can't do it*/) -#define ALIGNTEXT4ifNOP CHOICE(.align 4, .align ARG2(2,0x90), /*can't do it*/) -#define ALIGNDATA32 CHOICE(.align 32, .align ARG2(5,0x0), .align 32) -#define ALIGNDATA16 CHOICE(.align 16, .align ARG2(4,0x0), .align 16) -#define ALIGNDATA8 CHOICE(.align 8, .align ARG2(3,0x0), .align 8) -#define ALIGNDATA4 CHOICE(.align 4, .align ARG2(2,0x0), .align 4) -#define ALIGNDATA2 CHOICE(.align 2, .align ARG2(1,0x0), .align 2) -#endif /* __AOUT__ */ -#define FILE(s) CHOICE(.file s, .file s, .file s) -#define STRING(s) CHOICE(.string s, .asciz s, .asciz s) -#define D_LONG CHOICE(.long, .long, .data4) -#define D_WORD CHOICE(.value, .short, .data2) -#define D_BYTE CHOICE(.byte, .byte, .data1) -#define SPACE CHOICE(.comm, .space, .space) -#define COMM CHOICE(.comm, .comm, .comm) -#define SEG_DATA CHOICE(.data, .data, .sect .data) -#define SEG_TEXT CHOICE(.text, .text, .sect .text) -#define SEG_BSS CHOICE(.bss, .bss, .sect .bss) - -#ifdef GNU_ASSEMBLER -#define D_SPACE(n) . = . + n -#else -#define D_SPACE(n) .space n -#endif - -/* Addressing Modes */ -/* Immediate Mode */ -#define ADDR(a) CHOICE(CONCAT($,a), $a, a) -#define CONST(a) CHOICE(CONCAT($,a), $a, a) - -/* Indirect Mode */ -#define CONTENT(a) CHOICE(a, a, (a)) /* take contents of variable */ -#define REGIND(a) CHOICE((a), (a), (a)) /* Register a indirect */ -/* Register b indirect plus displacement a */ -#define REGOFF(a, b) CHOICE(a(b), a(b), a(b)) -/* Reg indirect Base + Index + Displacement - this is mainly for 16-bit mode - * which has no scaling - */ -#define REGBID(b,i,d) CHOICE(d(b,i), d(b,i), d(b)(i)) -/* Reg indirect Base + (Index * Scale) */ -#define REGBIS(b,i,s) CHOICE((b,i,s), (b,i,s), (b)(i*s)) -/* Reg indirect Base + (Index * Scale) + Displacement */ -#define REGBISD(b,i,s,d) CHOICE(d(b,i,s), d(b,i,s), d(b)(i*s)) -/* Displaced Scaled Index: */ -#define REGDIS(d,i,s) CHOICE(d(,i,s), d(,i,s), d(i * s)) -/* Indexed Base: */ -#define REGBI(b,i) CHOICE((b,i), (b,i), (b)(i)) -/* Displaced Base: */ -#define REGDB(d,b) CHOICE(d(b), d(b), d(b)) -/* Variable indirect: */ -#define VARINDIRECT(var) CHOICE(*var, *var, (var)) -/* Use register contents as jump/call target: */ -#define CODEPTR(reg) CHOICE(*reg, *reg, reg) - -/* For expressions requiring bracketing - * eg. (CRT0_PM | CRT_EM) - */ - -#define EXPR(a) CHOICE([a], (a), [a]) -#define ENOT(a) CHOICE(0!a, ~a, ~a) -#define EMUL(a,b) CHOICE(a\*b, a*b, a*b) -#define EDIV(a,b) CHOICE(a\/b, a/b, a/b) - -/* - * We have to beat the problem of commas within arguments to choice. - * eg. choice (add a,b, add b,a) will get argument mismatch. Luckily ANSI - * and other known cpp definitions evaluate arguments before substitution - * so the following works. - */ -#define ARG2(a, b) a,b -#define ARG3(a,b,c) a,b,c - -/* Redefine assembler commands */ -#define AAA CHOICE(aaa, aaa, aaa) -#define AAD CHOICE(aad, aad, aad) -#define AAM CHOICE(aam, aam, aam) -#define AAS CHOICE(aas, aas, aas) -#define ADC_L(a, b) CHOICE(adcl ARG2(a,b), adcl ARG2(a,b), _LTOG adc ARG2(b,a)) -#define ADC_W(a, b) CHOICE(adcw ARG2(a,b), adcw ARG2(a,b), _WTOG adc ARG2(b,a)) -#define ADC_B(a, b) CHOICE(adcb ARG2(a,b), adcb ARG2(a,b), adcb ARG2(b,a)) -#define ADD_L(a, b) CHOICE(addl ARG2(a,b), addl ARG2(a,b), _LTOG add ARG2(b,a)) -#define ADD_W(a, b) CHOICE(addw ARG2(a,b), addw ARG2(a,b), _WTOG add ARG2(b,a)) -#define ADD_B(a, b) CHOICE(addb ARG2(a,b), addb ARG2(a,b), addb ARG2(b,a)) -#define AND_L(a, b) CHOICE(andl ARG2(a,b), andl ARG2(a,b), _LTOG and ARG2(b,a)) -#define AND_W(a, b) CHOICE(andw ARG2(a,b), andw ARG2(a,b), _WTOG and ARG2(b,a)) -#define AND_B(a, b) CHOICE(andb ARG2(a,b), andb ARG2(a,b), andb ARG2(b,a)) -#define ARPL(a,b) CHOICE(arpl ARG2(a,b), arpl ARG2(a,b), arpl ARG2(b,a)) -#define BOUND_L(a, b) CHOICE(boundl ARG2(a,b), boundl ARG2(b,a), _LTOG bound ARG2(b,a)) -#define BOUND_W(a, b) CHOICE(boundw ARG2(a,b), boundw ARG2(b,a), _WTOG bound ARG2(b,a)) -#define BSF_L(a, b) CHOICE(bsfl ARG2(a,b), bsfl ARG2(a,b), _LTOG bsf ARG2(b,a)) -#define BSF_W(a, b) CHOICE(bsfw ARG2(a,b), bsfw ARG2(a,b), _WTOG bsf ARG2(b,a)) -#define BSR_L(a, b) CHOICE(bsrl ARG2(a,b), bsrl ARG2(a,b), _LTOG bsr ARG2(b,a)) -#define BSR_W(a, b) CHOICE(bsrw ARG2(a,b), bsrw ARG2(a,b), _WTOG bsr ARG2(b,a)) -#define BT_L(a, b) CHOICE(btl ARG2(a,b), btl ARG2(a,b), _LTOG bt ARG2(b,a)) -#define BT_W(a, b) CHOICE(btw ARG2(a,b), btw ARG2(a,b), _WTOG bt ARG2(b,a)) -#define BTC_L(a, b) CHOICE(btcl ARG2(a,b), btcl ARG2(a,b), _LTOG btc ARG2(b,a)) -#define BTC_W(a, b) CHOICE(btcw ARG2(a,b), btcw ARG2(a,b), _WTOG btc ARG2(b,a)) -#define BTR_L(a, b) CHOICE(btrl ARG2(a,b), btrl ARG2(a,b), _LTOG btr ARG2(b,a)) -#define BTR_W(a, b) CHOICE(btrw ARG2(a,b), btrw ARG2(a,b), _WTOG btr ARG2(b,a)) -#define BTS_L(a, b) CHOICE(btsl ARG2(a,b), btsl ARG2(a,b), _LTOG bts ARG2(b,a)) -#define BTS_W(a, b) CHOICE(btsw ARG2(a,b), btsw ARG2(a,b), _WTOG bts ARG2(b,a)) -#define CALL(a) CHOICE(call a, call a, call a) -#define CALLF(s,a) CHOICE(lcall ARG2(s,a), lcall ARG2(s,a), callf s:a) -#define CBW CHOICE(cbtw, cbw, cbw) -#define CWDE CHOICE(cwtd, cwde, cwde) -#define CLC CHOICE(clc, clc, clc) -#define CLD CHOICE(cld, cld, cld) -#define CLI CHOICE(cli, cli, cli) -#define CLTS CHOICE(clts, clts, clts) -#define CMC CHOICE(cmc, cmc, cmc) -#define CMP_L(a, b) CHOICE(cmpl ARG2(a,b), cmpl ARG2(a,b), _LTOG cmp ARG2(b,a)) -#define CMP_W(a, b) CHOICE(cmpw ARG2(a,b), cmpw ARG2(a,b), _WTOG cmp ARG2(b,a)) -#define CMP_B(a, b) CHOICE(cmpb ARG2(a,b), cmpb ARG2(a,b), cmpb ARG2(b,a)) -#define CMPS_L CHOICE(cmpsl, cmpsl, _LTOG cmps) -#define CMPS_W CHOICE(cmpsw, cmpsw, _WTOG cmps) -#define CMPS_B CHOICE(cmpsb, cmpsb, cmpsb) -#define CWD CHOICE(cwtl, cwd, cwd) -#define CDQ CHOICE(cltd, cdq, cdq) -#define DAA CHOICE(daa, daa, daa) -#define DAS CHOICE(das, das, das) -#define DEC_L(a) CHOICE(decl a, decl a, _LTOG dec a) -#define DEC_W(a) CHOICE(decw a, decw a, _WTOG dec a) -#define DEC_B(a) CHOICE(decb a, decb a, decb a) -#define DIV_L(a) CHOICE(divl a, divl a, div a) -#define DIV_W(a) CHOICE(divw a, divw a, div a) -#define DIV_B(a) CHOICE(divb a, divb a, divb a) -#define ENTER(a,b) CHOICE(enter ARG2(a,b), enter ARG2(a,b), enter ARG2(b,a)) -#define HLT CHOICE(hlt, hlt, hlt) -#define IDIV_L(a) CHOICE(idivl a, idivl a, _LTOG idiv a) -#define IDIV_W(a) CHOICE(idivw a, idivw a, _WTOG idiv a) -#define IDIV_B(a) CHOICE(idivb a, idivb a, idivb a) -/* More forms than this for imul!! */ -#define IMUL_L(a, b) CHOICE(imull ARG2(a,b), imull ARG2(a,b), _LTOG imul ARG2(b,a)) -#define IMUL_W(a, b) CHOICE(imulw ARG2(a,b), imulw ARG2(a,b), _WTOG imul ARG2(b,a)) -#define IMUL_B(a) CHOICE(imulb a, imulb a, imulb a) -#define IN_L CHOICE(inl (DX), inl ARG2(DX,EAX), _LTOG in DX) -#define IN_W CHOICE(inw (DX), inw ARG2(DX,AX), _WTOG in DX) -#define IN_B CHOICE(inb (DX), inb ARG2(DX,AL), inb DX) -/* Please AS code writer: use the following ONLY, if you refer to ports<256 - * directly, but not in IN1_W(DX), for instance, even if IN1_ looks nicer - */ -#if defined (sun) -#define IN1_L(a) CHOICE(inl (a), inl ARG2(a,EAX), _LTOG in a) -#define IN1_W(a) CHOICE(inw (a), inw ARG2(a,AX), _WTOG in a) -#define IN1_B(a) CHOICE(inb (a), inb ARG2(a,AL), inb a) -#else -#define IN1_L(a) CHOICE(inl a, inl ARG2(a,EAX), _LTOG in a) -#define IN1_W(a) CHOICE(inw a, inw ARG2(a,AX), _WTOG in a) -#define IN1_B(a) CHOICE(inb a, inb ARG2(a,AL), inb a) -#endif -#define INC_L(a) CHOICE(incl a, incl a, _LTOG inc a) -#define INC_W(a) CHOICE(incw a, incw a, _WTOG inc a) -#define INC_B(a) CHOICE(incb a, incb a, incb a) -#define INS_L CHOICE(insl, insl, _LTOG ins) -#define INS_W CHOICE(insw, insw, _WTOG ins) -#define INS_B CHOICE(insb, insb, insb) -#define INT(a) CHOICE(int a, int a, int a) -#define INT3 CHOICE(int CONST(3), int3, int CONST(3)) -#define INTO CHOICE(into, into, into) -#define IRET CHOICE(iret, iret, iret) -#define IRETD CHOICE(iret, iret, iretd) -#define JA(a) CHOICE(ja a, ja a, ja a) -#define JAE(a) CHOICE(jae a, jae a, jae a) -#define JB(a) CHOICE(jb a, jb a, jb a) -#define JBE(a) CHOICE(jbe a, jbe a, jbe a) -#define JC(a) CHOICE(jc a, jc a, jc a) -#define JE(a) CHOICE(je a, je a, je a) -#define JG(a) CHOICE(jg a, jg a, jg a) -#define JGE(a) CHOICE(jge a, jge a, jge a) -#define JL(a) CHOICE(jl a, jl a, jl a) -#define JLE(a) CHOICE(jle a, jle a, jle a) -#define JNA(a) CHOICE(jna a, jna a, jna a) -#define JNAE(a) CHOICE(jnae a, jnae a, jnae a) -#define JNB(a) CHOICE(jnb a, jnb a, jnb a) -#define JNBE(a) CHOICE(jnbe a, jnbe a, jnbe a) -#define JNC(a) CHOICE(jnc a, jnc a, jnc a) -#define JNE(a) CHOICE(jne a, jne a, jne a) -#define JNG(a) CHOICE(jng a, jng a, jng a) -#define JNGE(a) CHOICE(jnge a, jnge a, jnge a) -#define JNL(a) CHOICE(jnl a, jnl a, jnl a) -#define JNLE(a) CHOICE(jnle a, jnle a, jnle a) -#define JNO(a) CHOICE(jno a, jno a, jno a) -#define JNP(a) CHOICE(jnp a, jnp a, jnp a) -#define JNS(a) CHOICE(jns a, jns a, jns a) -#define JNZ(a) CHOICE(jnz a, jnz a, jnz a) -#define JO(a) CHOICE(jo a, jo a, jo a) -#define JP(a) CHOICE(jp a, jp a, jp a) -#define JPE(a) CHOICE(jpe a, jpe a, jpe a) -#define JPO(a) CHOICE(jpo a, jpo a, jpo a) -#define JS(a) CHOICE(js a, js a, js a) -#define JZ(a) CHOICE(jz a, jz a, jz a) -#define JMP(a) CHOICE(jmp a, jmp a, jmp a) -#define JMPF(s,a) CHOICE(ljmp ARG2(s,a), ljmp ARG2(s,a), jmpf s:a) -#define LAHF CHOICE(lahf, lahf, lahf) -#if !defined(_REAL_MODE) && !defined(_V86_MODE) -#define LAR(a, b) CHOICE(lar ARG2(a, b), lar ARG2(a, b), lar ARG2(b, a)) -#endif -#define LEA_L(a, b) CHOICE(leal ARG2(a,b), leal ARG2(a,b), _LTOG lea ARG2(b,a)) -#define LEA_W(a, b) CHOICE(leaw ARG2(a,b), leaw ARG2(a,b), _WTOG lea ARG2(b,a)) -#define LEAVE CHOICE(leave, leave, leave) -#define LGDT(a) CHOICE(lgdt a, lgdt a, lgdt a) -#define LIDT(a) CHOICE(lidt a, lidt a, lidt a) -#define LDS(a, b) CHOICE(ldsl ARG2(a,b), lds ARG2(a,b), lds ARG2(b,a)) -#define LES(a, b) CHOICE(lesl ARG2(a,b), les ARG2(a,b), les ARG2(b,a)) -#define LFS(a, b) CHOICE(lfsl ARG2(a,b), lfs ARG2(a,b), lfs ARG2(b,a)) -#define LGS(a, b) CHOICE(lgsl ARG2(a,b), lgs ARG2(a,b), lgs ARG2(b,a)) -#define LSS(a, b) CHOICE(lssl ARG2(a,b), lss ARG2(a,b), lss ARG2(b,a)) -#define LLDT(a) CHOICE(lldt a, lldt a, lldt a) -#define LMSW(a) CHOICE(lmsw a, lmsw a, lmsw a) -#define LOCK CHOICE(lock, lock, lock) -#define LODS_L CHOICE(lodsl, lodsl, _LTOG lods) -#define LODS_W CHOICE(lodsw, lodsw, _WTOG lods) -#define LODS_B CHOICE(lodsb, lodsb, lodsb) -#define LOOP(a) CHOICE(loop a, loop a, loop a) -#define LOOPE(a) CHOICE(loope a, loope a, loope a) -#define LOOPZ(a) CHOICE(loopz a, loopz a, loopz a) -#define LOOPNE(a) CHOICE(loopne a, loopne a, loopne a) -#define LOOPNZ(a) CHOICE(loopnz a, loopnz a, loopnz a) -#if !defined(_REAL_MODE) && !defined(_V86_MODE) -#define LSL(a, b) CHOICE(lsl ARG2(a,b), lsl ARG2(a,b), lsl ARG2(b,a)) -#endif -#define LTR(a) CHOICE(ltr a, ltr a, ltr a) -#define MOV_SR(a, b) CHOICE(movw ARG2(a,b), mov ARG2(a,b), mov ARG2(b,a)) -#define MOV_L(a, b) CHOICE(movl ARG2(a,b), movl ARG2(a,b), _LTOG mov ARG2(b,a)) -#define MOV_W(a, b) CHOICE(movw ARG2(a,b), movw ARG2(a,b), _WTOG mov ARG2(b,a)) -#define MOV_B(a, b) CHOICE(movb ARG2(a,b), movb ARG2(a,b), movb ARG2(b,a)) -#define MOVS_L CHOICE(movsl, movsl, _LTOG movs) -#define MOVS_W CHOICE(movsw, movsw, _WTOG movs) -#define MOVS_B CHOICE(movsb, movsb, movsb) -#define MOVSX_BL(a, b) CHOICE(movsbl ARG2(a,b), movsbl ARG2(a,b), movsx ARG2(b,a)) -#define MOVSX_BW(a, b) CHOICE(movsbw ARG2(a,b), movsbw ARG2(a,b), movsx ARG2(b,a)) -#define MOVSX_WL(a, b) CHOICE(movswl ARG2(a,b), movswl ARG2(a,b), movsx ARG2(b,a)) -#define MOVZX_BL(a, b) CHOICE(movzbl ARG2(a,b), movzbl ARG2(a,b), movzx ARG2(b,a)) -#define MOVZX_BW(a, b) CHOICE(movzbw ARG2(a,b), movzbw ARG2(a,b), movzx ARG2(b,a)) -#define MOVZX_WL(a, b) CHOICE(movzwl ARG2(a,b), movzwl ARG2(a,b), movzx ARG2(b,a)) -#define MUL_L(a) CHOICE(mull a, mull a, _LTOG mul a) -#define MUL_W(a) CHOICE(mulw a, mulw a, _WTOG mul a) -#define MUL_B(a) CHOICE(mulb a, mulb a, mulb a) -#define NEG_L(a) CHOICE(negl a, negl a, _LTOG neg a) -#define NEG_W(a) CHOICE(negw a, negw a, _WTOG neg a) -#define NEG_B(a) CHOICE(negb a, negb a, negb a) -#define NOP CHOICE(nop, nop, nop) -#define NOT_L(a) CHOICE(notl a, notl a, _LTOG not a) -#define NOT_W(a) CHOICE(notw a, notw a, _WTOG not a) -#define NOT_B(a) CHOICE(notb a, notb a, notb a) -#define OR_L(a,b) CHOICE(orl ARG2(a,b), orl ARG2(a,b), _LTOG or ARG2(b,a)) -#define OR_W(a,b) CHOICE(orw ARG2(a,b), orw ARG2(a,b), _WTOG or ARG2(b,a)) -#define OR_B(a,b) CHOICE(orb ARG2(a,b), orb ARG2(a,b), orb ARG2(b,a)) -#define OUT_L CHOICE(outl (DX), outl ARG2(EAX,DX), _LTOG out DX) -#define OUT_W CHOICE(outw (DX), outw ARG2(AX,DX), _WTOG out DX) -#define OUT_B CHOICE(outb (DX), outb ARG2(AL,DX), outb DX) -/* Please AS code writer: use the following ONLY, if you refer to ports<256 - * directly, but not in OUT1_W(DX), for instance, even if OUT1_ looks nicer - */ -#define OUT1_L(a) CHOICE(outl (a), outl ARG2(EAX,a), _LTOG out a) -#define OUT1_W(a) CHOICE(outw (a), outw ARG2(AX,a), _WTOG out a) -#define OUT1_B(a) CHOICE(outb (a), outb ARG2(AL,a), outb a) -#define OUTS_L CHOICE(outsl, outsl, _LTOG outs) -#define OUTS_W CHOICE(outsw, outsw, _WTOG outs) -#define OUTS_B CHOICE(outsb, outsb, outsb) -#define POP_SR(a) CHOICE(pop a, pop a, pop a) -#define POP_L(a) CHOICE(popl a, popl a, _LTOG pop a) -#define POP_W(a) CHOICE(popw a, popw a, _WTOG pop a) -#define POPA_L CHOICE(popal, popal, _LTOG popa) -#define POPA_W CHOICE(popaw, popaw, _WTOG popa) -#define POPF_L CHOICE(popfl, popfl, _LTOG popf) -#define POPF_W CHOICE(popfw, popfw, _WTOG popf) -#define PUSH_SR(a) CHOICE(push a, push a, push a) -#define PUSH_L(a) CHOICE(pushl a, pushl a, _LTOG push a) -#define PUSH_W(a) CHOICE(pushw a, pushw a, _WTOG push a) -#define PUSH_B(a) CHOICE(push a, pushb a, push a) -#define PUSHA_L CHOICE(pushal, pushal, _LTOG pusha) -#define PUSHA_W CHOICE(pushaw, pushaw, _WTOG pusha) -#define PUSHF_L CHOICE(pushfl, pushfl, _LTOG pushf) -#define PUSHF_W CHOICE(pushfw, pushfw, _WTOG pushf) -#define RCL_L(a, b) CHOICE(rcll ARG2(a,b), rcll ARG2(a,b), _LTOG rcl ARG2(b,a)) -#define RCL_W(a, b) CHOICE(rclw ARG2(a,b), rclw ARG2(a,b), _WTOG rcl ARG2(b,a)) -#define RCL_B(a, b) CHOICE(rclb ARG2(a,b), rclb ARG2(a,b), rclb ARG2(b,a)) -#define RCR_L(a, b) CHOICE(rcrl ARG2(a,b), rcrl ARG2(a,b), _LTOG rcr ARG2(b,a)) -#define RCR_W(a, b) CHOICE(rcrw ARG2(a,b), rcrw ARG2(a,b), _WTOG rcr ARG2(b,a)) -#define RCR_B(a, b) CHOICE(rcrb ARG2(a,b), rcrb ARG2(a,b), rcrb ARG2(b,a)) -#define ROL_L(a, b) CHOICE(roll ARG2(a,b), roll ARG2(a,b), _LTOG rol ARG2(b,a)) -#define ROL_W(a, b) CHOICE(rolw ARG2(a,b), rolw ARG2(a,b), _WTOG rol ARG2(b,a)) -#define ROL_B(a, b) CHOICE(rolb ARG2(a,b), rolb ARG2(a,b), rolb ARG2(b,a)) -#define ROR_L(a, b) CHOICE(rorl ARG2(a,b), rorl ARG2(a,b), _LTOG ror ARG2(b,a)) -#define ROR_W(a, b) CHOICE(rorw ARG2(a,b), rorw ARG2(a,b), _WTOG ror ARG2(b,a)) -#define ROR_B(a, b) CHOICE(rorb ARG2(a,b), rorb ARG2(a,b), rorb ARG2(b,a)) -#define REP CHOICE(rep ;, rep ;, repe) -#define REPE CHOICE(repz ;, repe ;, repe) -#define REPNE CHOICE(repnz ;, repne ;, repne) -#define REPNZ REPNE -#define REPZ REPE -#define RET CHOICE(ret, ret, ret) -#define SAHF CHOICE(sahf, sahf, sahf) -#define SAL_L(a, b) CHOICE(sall ARG2(a,b), sall ARG2(a,b), _LTOG sal ARG2(b,a)) -#define SAL_W(a, b) CHOICE(salw ARG2(a,b), salw ARG2(a,b), _WTOG sal ARG2(b,a)) -#define SAL_B(a, b) CHOICE(salb ARG2(a,b), salb ARG2(a,b), salb ARG2(b,a)) -#define SAR_L(a, b) CHOICE(sarl ARG2(a,b), sarl ARG2(a,b), _LTOG sar ARG2(b,a)) -#define SAR_W(a, b) CHOICE(sarw ARG2(a,b), sarw ARG2(a,b), _WTOG sar ARG2(b,a)) -#define SAR_B(a, b) CHOICE(sarb ARG2(a,b), sarb ARG2(a,b), sarb ARG2(b,a)) -#define SBB_L(a, b) CHOICE(sbbl ARG2(a,b), sbbl ARG2(a,b), _LTOG sbb ARG2(b,a)) -#define SBB_W(a, b) CHOICE(sbbw ARG2(a,b), sbbw ARG2(a,b), _WTOG sbb ARG2(b,a)) -#define SBB_B(a, b) CHOICE(sbbb ARG2(a,b), sbbb ARG2(a,b), sbbb ARG2(b,a)) -#define SCAS_L CHOICE(scasl, scasl, _LTOG scas) -#define SCAS_W CHOICE(scasw, scasw, _WTOG scas) -#define SCAS_B CHOICE(scasb, scasb, scasb) -#define SETA(a) CHOICE(seta a, seta a, seta a) -#define SETAE(a) CHOICE(setae a, setae a, setae a) -#define SETB(a) CHOICE(setb a, setb a, setb a) -#define SETBE(a) CHOICE(setbe a, setbe a, setbe a) -#define SETC(a) CHOICE(setc a, setb a, setb a) -#define SETE(a) CHOICE(sete a, sete a, sete a) -#define SETG(a) CHOICE(setg a, setg a, setg a) -#define SETGE(a) CHOICE(setge a, setge a, setge a) -#define SETL(a) CHOICE(setl a, setl a, setl a) -#define SETLE(a) CHOICE(setle a, setle a, setle a) -#define SETNA(a) CHOICE(setna a, setna a, setna a) -#define SETNAE(a) CHOICE(setnae a, setnae a, setnae a) -#define SETNB(a) CHOICE(setnb a, setnb a, setnb a) -#define SETNBE(a) CHOICE(setnbe a, setnbe a, setnbe a) -#define SETNC(a) CHOICE(setnc a, setnb a, setnb a) -#define SETNE(a) CHOICE(setne a, setne a, setne a) -#define SETNG(a) CHOICE(setng a, setng a, setng a) -#define SETNGE(a) CHOICE(setnge a, setnge a, setnge a) -#define SETNL(a) CHOICE(setnl a, setnl a, setnl a) -#define SETNLE(a) CHOICE(setnle a, setnle a, setnle a) -#define SETNO(a) CHOICE(setno a, setno a, setno a) -#define SETNP(a) CHOICE(setnp a, setnp a, setnp a) -#define SETNS(a) CHOICE(setns a, setns a, setna a) -#define SETNZ(a) CHOICE(setnz a, setnz a, setnz a) -#define SETO(a) CHOICE(seto a, seto a, seto a) -#define SETP(a) CHOICE(setp a, setp a, setp a) -#define SETPE(a) CHOICE(setpe a, setpe a, setpe a) -#define SETPO(a) CHOICE(setpo a, setpo a, setpo a) -#define SETS(a) CHOICE(sets a, sets a, seta a) -#define SETZ(a) CHOICE(setz a, setz a, setz a) -#define SGDT(a) CHOICE(sgdt a, sgdt a, sgdt a) -#define SIDT(a) CHOICE(sidt a, sidt a, sidt a) -#define SHL_L(a, b) CHOICE(shll ARG2(a,b), shll ARG2(a,b), _LTOG shl ARG2(b,a)) -#define SHL_W(a, b) CHOICE(shlw ARG2(a,b), shlw ARG2(a,b), _WTOG shl ARG2(b,a)) -#define SHL_B(a, b) CHOICE(shlb ARG2(a,b), shlb ARG2(a,b), shlb ARG2(b,a)) -#define SHLD_L(a,b,c) CHOICE(shldl ARG3(a,b,c), shldl ARG3(a,b,c), _LTOG shld ARG3(c,b,a)) -#define SHLD2_L(a,b) CHOICE(shldl ARG2(a,b), shldl ARG3(CL,a,b), _LTOG shld ARG3(b,a,CL)) -#define SHLD_W(a,b,c) CHOICE(shldw ARG3(a,b,c), shldw ARG3(a,b,c), _WTOG shld ARG3(c,b,a)) -#define SHLD2_W(a,b) CHOICE(shldw ARG2(a,b), shldw ARG3(CL,a,b), _WTOG shld ARG3(b,a,CL)) -#define SHR_L(a, b) CHOICE(shrl ARG2(a,b), shrl ARG2(a,b), _LTOG shr ARG2(b,a)) -#define SHR_W(a, b) CHOICE(shrw ARG2(a,b), shrw ARG2(a,b), _WTOG shr ARG2(b,a)) -#define SHR_B(a, b) CHOICE(shrb ARG2(a,b), shrb ARG2(a,b), shrb ARG2(b,a)) -#define SHRD_L(a,b,c) CHOICE(shrdl ARG3(a,b,c), shrdl ARG3(a,b,c), _LTOG shrd ARG3(c,b,a)) -#define SHRD2_L(a,b) CHOICE(shrdl ARG2(a,b), shrdl ARG3(CL,a,b), _LTOG shrd ARG3(b,a,CL)) -#define SHRD_W(a,b,c) CHOICE(shrdw ARG3(a,b,c), shrdw ARG3(a,b,c), _WTOG shrd ARG3(c,b,a)) -#define SHRD2_W(a,b) CHOICE(shrdw ARG2(a,b), shrdw ARG3(CL,a,b), _WTOG shrd ARG3(b,a,CL)) -#define SLDT(a) CHOICE(sldt a, sldt a, sldt a) -#define SMSW(a) CHOICE(smsw a, smsw a, smsw a) -#define STC CHOICE(stc, stc, stc) -#define STD CHOICE(std, std, std) -#define STI CHOICE(sti, sti, sti) -#define STOS_L CHOICE(stosl, stosl, _LTOG stos) -#define STOS_W CHOICE(stosw, stosw, _WTOG stos) -#define STOS_B CHOICE(stosb, stosb, stosb) -#define STR(a) CHOICE(str a, str a, str a) -#define SUB_L(a, b) CHOICE(subl ARG2(a,b), subl ARG2(a,b), _LTOG sub ARG2(b,a)) -#define SUB_W(a, b) CHOICE(subw ARG2(a,b), subw ARG2(a,b), _WTOG sub ARG2(b,a)) -#define SUB_B(a, b) CHOICE(subb ARG2(a,b), subb ARG2(a,b), subb ARG2(b,a)) -#define TEST_L(a, b) CHOICE(testl ARG2(a,b), testl ARG2(a,b), _LTOG test ARG2(b,a)) -#define TEST_W(a, b) CHOICE(testw ARG2(a,b), testw ARG2(a,b), _WTOG test ARG2(b,a)) -#define TEST_B(a, b) CHOICE(testb ARG2(a,b), testb ARG2(a,b), testb ARG2(b,a)) -#define VERR(a) CHOICE(verr a, verr a, verr a) -#define VERW(a) CHOICE(verw a, verw a, verw a) -#define WAIT CHOICE(wait, wait, wait) -#define XCHG_L(a, b) CHOICE(xchgl ARG2(a,b), xchgl ARG2(a,b), _LTOG xchg ARG2(b,a)) -#define XCHG_W(a, b) CHOICE(xchgw ARG2(a,b), xchgw ARG2(a,b), _WTOG xchg ARG2(b,a)) -#define XCHG_B(a, b) CHOICE(xchgb ARG2(a,b), xchgb ARG2(a,b), xchgb ARG2(b,a)) -#define XLAT CHOICE(xlat, xlat, xlat) -#define XOR_L(a, b) CHOICE(xorl ARG2(a,b), xorl ARG2(a,b), _LTOG xor ARG2(b,a)) -#define XOR_W(a, b) CHOICE(xorw ARG2(a,b), xorw ARG2(a,b), _WTOG xor ARG2(b,a)) -#define XOR_B(a, b) CHOICE(xorb ARG2(a,b), xorb ARG2(a,b), xorb ARG2(b,a)) - - -/* Floating Point Instructions */ -#define F2XM1 CHOICE(f2xm1, f2xm1, f2xm1) -#define FABS CHOICE(fabs, fabs, fabs) -#define FADD_D(a) CHOICE(faddl a, faddl a, faddd a) -#define FADD_S(a) CHOICE(fadds a, fadds a, fadds a) -#define FADD2(a, b) CHOICE(fadd ARG2(a,b), fadd ARG2(a,b), fadd ARG2(b,a)) -#define FADDP(a, b) CHOICE(faddp ARG2(a,b), faddp ARG2(a,b), faddp ARG2(b,a)) -#define FIADD_L(a) CHOICE(fiaddl a, fiaddl a, fiaddl a) -#define FIADD_W(a) CHOICE(fiadd a, fiadds a, fiadds a) -#define FBLD(a) CHOICE(fbld a, fbld a, fbld a) -#define FBSTP(a) CHOICE(fbstp a, fbstp a, fbstp a) -#define FCHS CHOICE(fchs, fchs, fchs) -#define FCLEX CHOICE(fclex, wait; fnclex, wait; fclex) -#define FNCLEX CHOICE(fnclex, fnclex, fclex) -#define FCOM(a) CHOICE(fcom a, fcom a, fcom a) -#define FCOM_D(a) CHOICE(fcoml a, fcoml a, fcomd a) -#define FCOM_S(a) CHOICE(fcoms a, fcoms a, fcoms a) -#define FCOMP(a) CHOICE(fcomp a, fcomp a, fcomp a) -#define FCOMP_D(a) CHOICE(fcompl a, fcompl a, fcompd a) -#define FCOMP_S(a) CHOICE(fcomps a, fcomps a, fcomps a) -#define FCOMPP CHOICE(fcompp, fcompp, fcompp) -#define FCOS CHOICE(fcos, fcos, fcos) -#define FDECSTP CHOICE(fdecstp, fdecstp, fdecstp) -#define FDIV_D(a) CHOICE(fdivl a, fdivl a, fdivd a) -#define FDIV_S(a) CHOICE(fdivs a, fdivs a, fdivs a) -#define FDIV2(a, b) CHOICE(fdiv ARG2(a,b), fdiv ARG2(a,b), fdiv ARG2(b,a)) -#define FDIVP(a, b) CHOICE(fdivp ARG2(a,b), fdivp ARG2(a,b), fdivp ARG2(b,a)) -#define FIDIV_L(a) CHOICE(fidivl a, fidivl a, fidivl a) -#define FIDIV_W(a) CHOICE(fidiv a, fidivs a, fidivs a) -#define FDIVR_D(a) CHOICE(fdivrl a, fdivrl a, fdivrd a) -#define FDIVR_S(a) CHOICE(fdivrs a, fdivrs a, fdivrs a) -#define FDIVR2(a, b) CHOICE(fdivr ARG2(a,b), fdivr ARG2(a,b), fdivr ARG2(b,a)) -#define FDIVRP(a, b) CHOICE(fdivrp ARG2(a,b), fdivrp ARG2(a,b), fdivrp ARG2(b,a)) -#define FIDIVR_L(a) CHOICE(fidivrl a, fidivrl a, fidivrl a) -#define FIDIVR_W(a) CHOICE(fidivr a, fidivrs a, fidivrs a) -#define FFREE(a) CHOICE(ffree a, ffree a, ffree a) -#define FICOM_L(a) CHOICE(ficoml a, ficoml a, ficoml a) -#define FICOM_W(a) CHOICE(ficom a, ficoms a, ficoms a) -#define FICOMP_L(a) CHOICE(ficompl a, ficompl a, ficompl a) -#define FICOMP_W(a) CHOICE(ficomp a, ficomps a, ficomps a) -#define FILD_Q(a) CHOICE(fildll a, fildq a, fildq a) -#define FILD_L(a) CHOICE(fildl a, fildl a, fildl a) -#define FILD_W(a) CHOICE(fild a, filds a, filds a) -#define FINCSTP CHOICE(fincstp, fincstp, fincstp) -#define FINIT CHOICE(finit, wait; fninit, wait; finit) -#define FNINIT CHOICE(fninit, fninit, finit) -#define FIST_L(a) CHOICE(fistl a, fistl a, fistl a) -#define FIST_W(a) CHOICE(fist a, fists a, fists a) -#define FISTP_Q(a) CHOICE(fistpll a, fistpq a, fistpq a) -#define FISTP_L(a) CHOICE(fistpl a, fistpl a, fistpl a) -#define FISTP_W(a) CHOICE(fistp a, fistps a, fistps a) -#define FLD_X(a) CHOICE(fldt a, fldt a, fldx a) /* 80 bit data type! */ -#define FLD_D(a) CHOICE(fldl a, fldl a, fldd a) -#define FLD_S(a) CHOICE(flds a, flds a, flds a) -#define FLD1 CHOICE(fld1, fld1, fld1) -#define FLDL2T CHOICE(fldl2t, fldl2t, fldl2t) -#define FLDL2E CHOICE(fldl2e, fldl2e, fldl2e) -#define FLDPI CHOICE(fldpi, fldpi, fldpi) -#define FLDLG2 CHOICE(fldlg2, fldlg2, fldlg2) -#define FLDLN2 CHOICE(fldln2, fldln2, fldln2) -#define FLDZ CHOICE(fldz, fldz, fldz) -#define FLDCW(a) CHOICE(fldcw a, fldcw a, fldcw a) -#define FLDENV(a) CHOICE(fldenv a, fldenv a, fldenv a) -#define FMUL_S(a) CHOICE(fmuls a, fmuls a, fmuls a) -#define FMUL_D(a) CHOICE(fmull a, fmull a, fmuld a) -#define FMUL2(a, b) CHOICE(fmul ARG2(a,b), fmul ARG2(a,b), fmul ARG2(b,a)) -#define FMULP(a, b) CHOICE(fmulp ARG2(a,b), fmulp ARG2(a,b), fmulp ARG2(b,a)) -#define FIMUL_L(a) CHOICE(fimull a, fimull a, fimull a) -#define FIMUL_W(a) CHOICE(fimul a, fimuls a, fimuls a) -#define FNOP CHOICE(fnop, fnop, fnop) -#define FPATAN CHOICE(fpatan, fpatan, fpatan) -#define FPREM CHOICE(fprem, fprem, fprem) -#define FPREM1 CHOICE(fprem1, fprem1, fprem1) -#define FPTAN CHOICE(fptan, fptan, fptan) -#define FRNDINT CHOICE(frndint, frndint, frndint) -#define FRSTOR(a) CHOICE(frstor a, frstor a, frstor a) -#define FSAVE(a) CHOICE(fsave a, wait; fnsave a, wait; fsave a) -#define FNSAVE(a) CHOICE(fnsave a, fnsave a, fsave a) -#define FSCALE CHOICE(fscale, fscale, fscale) -#define FSIN CHOICE(fsin, fsin, fsin) -#define FSINCOS CHOICE(fsincos, fsincos, fsincos) -#define FSQRT CHOICE(fsqrt, fsqrt, fsqrt) -#define FST_D(a) CHOICE(fstl a, fstl a, fstd a) -#define FST_S(a) CHOICE(fsts a, fsts a, fsts a) -#define FSTP_X(a) CHOICE(fstpt a, fstpt a, fstpx a) -#define FSTP_D(a) CHOICE(fstpl a, fstpl a, fstpd a) -#define FSTP_S(a) CHOICE(fstps a, fstps a, fstps a) -#define FSTP(a) CHOICE(fstp a, fstp a, fstp a) -#define FSTCW(a) CHOICE(fstcw a, wait; fnstcw a, wait; fstcw a) -#define FNSTCW(a) CHOICE(fnstcw a, fnstcw a, fstcw a) -#define FSTENV(a) CHOICE(fstenv a, wait; fnstenv a, fstenv a) -#define FNSTENV(a) CHOICE(fnstenv a, fnstenv a, fstenv a) -#define FSTSW(a) CHOICE(fstsw a, wait; fnstsw a, wait; fstsw a) -#define FNSTSW(a) CHOICE(fnstsw a, fnstsw a, fstsw a) -#define FSUB_S(a) CHOICE(fsubs a, fsubs a, fsubs a) -#define FSUB_D(a) CHOICE(fsubl a, fsubl a, fsubd a) -#define FSUB2(a, b) CHOICE(fsub ARG2(a,b), fsub ARG2(a,b), fsub ARG2(b,a)) -#define FSUBP(a, b) CHOICE(fsubp ARG2(a,b), fsubp ARG2(a,b), fsubp ARG2(b,a)) -#define FISUB_L(a) CHOICE(fisubl a, fisubl a, fisubl a) -#define FISUB_W(a) CHOICE(fisub a, fisubs a, fisubs a) -#define FSUBR_S(a) CHOICE(fsubrs a, fsubrs a, fsubrs a) -#define FSUBR_D(a) CHOICE(fsubrl a, fsubrl a, fsubrd a) -#define FSUBR2(a, b) CHOICE(fsubr ARG2(a,b), fsubr ARG2(a,b), fsubr ARG2(b,a)) -#define FSUBRP(a, b) CHOICE(fsubrp ARG2(a,b), fsubrp ARG2(a,b), fsubrp ARG2(b,a)) -#define FISUBR_L(a) CHOICE(fisubrl a, fisubrl a, fisubrl a) -#define FISUBR_W(a) CHOICE(fisubr a, fisubrs a, fisubrs a) -#define FTST CHOICE(ftst, ftst, ftst) -#define FUCOM(a) CHOICE(fucom a, fucom a, fucom a) -#define FUCOMP(a) CHOICE(fucomp a, fucomp a, fucomp a) -#define FUCOMPP CHOICE(fucompp, fucompp, fucompp) -#define FWAIT CHOICE(wait, wait, wait) -#define FXAM CHOICE(fxam, fxam, fxam) -#define FXCH(a) CHOICE(fxch a, fxch a, fxch a) -#define FXTRACT CHOICE(fxtract, fxtract, fxtract) -#define FYL2X CHOICE(fyl2x, fyl2x, fyl2x) -#define FYL2XP1 CHOICE(fyl2xp1, fyl2xp1, fyl2xp1) - -/* New instructions */ -#define CPUID CHOICE(D_BYTE ARG2(15, 162), cpuid, D_BYTE ARG2(15, 162)) -#define RDTSC CHOICE(D_BYTE ARG2(15, 49), rdtsc, D_BYTE ARG2(15, 49)) - -#else /* NASM_ASSEMBLER || MASM_ASSEMBLER is defined */ - - /****************************************/ - /* */ - /* Intel style assemblers. */ - /* (NASM and MASM) */ - /* */ - /****************************************/ - -#define P_EAX EAX -#define L_EAX EAX -#define W_AX AX -#define B_AH AH -#define B_AL AL - -#define P_EBX EBX -#define L_EBX EBX -#define W_BX BX -#define B_BH BH -#define B_BL BL - -#define P_ECX ECX -#define L_ECX ECX -#define W_CX CX -#define B_CH CH -#define B_CL CL - -#define P_EDX EDX -#define L_EDX EDX -#define W_DX DX -#define B_DH DH -#define B_DL DL - -#define P_EBP EBP -#define L_EBP EBP -#define W_BP BP - -#define P_ESI ESI -#define L_ESI ESI -#define W_SI SI - -#define P_EDI EDI -#define L_EDI EDI -#define W_DI DI - -#define P_ESP ESP -#define L_ESP ESP -#define W_SP SP - -#define W_CS CS -#define W_SS SS -#define W_DS DS -#define W_ES ES -#define W_FS FS -#define W_GS GS - -#define X_ST ST -#define D_ST ST -#define L_ST ST - -#define P_MM0 mm0 -#define P_MM1 mm1 -#define P_MM2 mm2 -#define P_MM3 mm3 -#define P_MM4 mm4 -#define P_MM5 mm5 -#define P_MM6 mm6 -#define P_MM7 mm7 - -#define P_XMM0 xmm0 -#define P_XMM1 xmm1 -#define P_XMM2 xmm2 -#define P_XMM3 xmm3 -#define P_XMM4 xmm4 -#define P_XMM5 xmm5 -#define P_XMM6 xmm6 -#define P_XMM7 xmm7 - -#define CONCAT(x, y) x ## y -#define CONCAT3(x, y, z) x ## y ## z - -#if defined(NASM_ASSEMBLER) - -#define ST(n) st ## n -#define ST0 st0 - -#define TBYTE_PTR tword -#define QWORD_PTR qword -#define DWORD_PTR dword -#define WORD_PTR word -#define BYTE_PTR byte - -#define OFFSET - -#define GLOBL GLOBAL -#define ALIGNTEXT32 ALIGN 32 -#define ALIGNTEXT16 ALIGN 16 -#define ALIGNTEXT8 ALIGN 8 -#define ALIGNTEXT4 ALIGN 4 -#define ALIGNTEXT2 ALIGN 2 -#define ALIGNTEXT32ifNOP ALIGN 32 -#define ALIGNTEXT16ifNOP ALIGN 16 -#define ALIGNTEXT8ifNOP ALIGN 8 -#define ALIGNTEXT4ifNOP ALIGN 4 -#define ALIGNDATA32 ALIGN 32 -#define ALIGNDATA16 ALIGN 16 -#define ALIGNDATA8 ALIGN 8 -#define ALIGNDATA4 ALIGN 4 -#define ALIGNDATA2 ALIGN 2 -#define FILE(s) -#define STRING(s) db s -#define D_LONG dd -#define D_WORD dw -#define D_BYTE db -/* #define SPACE */ -/* #define COMM */ -#if defined(__WATCOMC__) -SECTION _TEXT public align=16 class=CODE use32 flat -SECTION _DATA public align=16 class=DATA use32 flat -#define SEG_TEXT SECTION _TEXT -#define SEG_DATA SECTION _DATA -#define SEG_BSS SECTION .bss -#else -#define SEG_DATA SECTION .data -#define SEG_TEXT SECTION .text -#define SEG_BSS SECTION .bss -#endif - -#define D_SPACE(n) db n REP 0 - -#define AS_BEGIN - -/* Jcc's should be handled better than this... */ -#define NEAR near - -#else /* MASM */ - -#define TBYTE_PTR tbyte ptr -#define QWORD_PTR qword ptr -#define DWORD_PTR dword ptr -#define WORD_PTR word ptr -#define BYTE_PTR byte ptr - -#define OFFSET offset - -#define GLOBL GLOBAL -#define ALIGNTEXT32 ALIGN 32 -#define ALIGNTEXT16 ALIGN 16 -#define ALIGNTEXT8 ALIGN 8 -#define ALIGNTEXT4 ALIGN 4 -#define ALIGNTEXT2 ALIGN 2 -#define ALIGNTEXT32ifNOP ALIGN 32 -#define ALIGNTEXT16ifNOP ALIGN 16 -#define ALIGNTEXT8ifNOP ALIGN 8 -#define ALIGNTEXT4ifNOP ALIGN 4 -#define ALIGNDATA32 ALIGN 32 -#define ALIGNDATA16 ALIGN 16 -#define ALIGNDATA8 ALIGN 8 -#define ALIGNDATA4 ALIGN 4 -#define ALIGNDATA2 ALIGN 2 -#define FILE(s) -#define STRING(s) db s -#define D_LONG dd -#define D_WORD dw -#define D_BYTE db -/* #define SPACE */ -/* #define COMM */ -#define SEG_DATA .DATA -#define SEG_TEXT .CODE -#define SEG_BSS .DATA - -#define D_SPACE(n) db n REP 0 - -#define AS_BEGIN - -#define NEAR - -#endif - -#if defined(Lynx) || (defined(SYSV) || defined(SVR4)) \ - || (defined(__linux__) || defined(__OS2ELF__)) && defined(__ELF__) \ - || (defined(__FreeBSD__) && __FreeBSD__ >= 3) \ - || (defined(__NetBSD__) && defined(__ELF__)) -#define GLNAME(a) a -#else -#define GLNAME(a) CONCAT(_, a) -#endif - -/* - * Addressing Modes - */ - -/* Immediate Mode */ -#define P_ADDR(a) OFFSET a -#define X_ADDR(a) OFFSET a -#define D_ADDR(a) OFFSET a -#define L_ADDR(a) OFFSET a -#define W_ADDR(a) OFFSET a -#define B_ADDR(a) OFFSET a - -#define P_CONST(a) a -#define X_CONST(a) a -#define D_CONST(a) a -#define L_CONST(a) a -#define W_CONST(a) a -#define B_CONST(a) a - -/* Indirect Mode */ -#ifdef NASM_ASSEMBLER -#define P_CONTENT(a) [a] -#define X_CONTENT(a) TBYTE_PTR [a] -#define D_CONTENT(a) QWORD_PTR [a] -#define L_CONTENT(a) DWORD_PTR [a] -#define W_CONTENT(a) WORD_PTR [a] -#define B_CONTENT(a) BYTE_PTR [a] -#else -#define P_CONTENT(a) a -#define X_CONTENT(a) TBYTE_PTR a -#define D_CONTENT(a) QWORD_PTR a -#define L_CONTENT(a) DWORD_PTR a -#define W_CONTENT(a) WORD_PTR a -#define B_CONTENT(a) BYTE_PTR a -#endif - -/* Register a indirect */ -#define P_REGIND(a) [a] -#define X_REGIND(a) TBYTE_PTR [a] -#define D_REGIND(a) QWORD_PTR [a] -#define L_REGIND(a) DWORD_PTR [a] -#define W_REGIND(a) WORD_PTR [a] -#define B_REGIND(a) BYTE_PTR [a] - -/* Register b indirect plus displacement a */ -#define P_REGOFF(a, b) [b + a] -#define X_REGOFF(a, b) TBYTE_PTR [b + a] -#define D_REGOFF(a, b) QWORD_PTR [b + a] -#define L_REGOFF(a, b) DWORD_PTR [b + a] -#define W_REGOFF(a, b) WORD_PTR [b + a] -#define B_REGOFF(a, b) BYTE_PTR [b + a] - -/* Reg indirect Base + Index + Displacement - this is mainly for 16-bit mode - * which has no scaling - */ -#define P_REGBID(b, i, d) [b + i + d] -#define X_REGBID(b, i, d) TBYTE_PTR [b + i + d] -#define D_REGBID(b, i, d) QWORD_PTR [b + i + d] -#define L_REGBID(b, i, d) DWORD_PTR [b + i + d] -#define W_REGBID(b, i, d) WORD_PTR [b + i + d] -#define B_REGBID(b, i, d) BYTE_PTR [b + i + d] - -/* Reg indirect Base + (Index * Scale) */ -#define P_REGBIS(b, i, s) [b + i * s] -#define X_REGBIS(b, i, s) TBYTE_PTR [b + i * s] -#define D_REGBIS(b, i, s) QWORD_PTR [b + i * s] -#define L_REGBIS(b, i, s) DWORD_PTR [b + i * s] -#define W_REGBIS(b, i, s) WORD_PTR [b + i * s] -#define B_REGBIS(b, i, s) BYTE_PTR [b + i * s] - -/* Reg indirect Base + (Index * Scale) + Displacement */ -#define P_REGBISD(b, i, s, d) [b + i * s + d] -#define X_REGBISD(b, i, s, d) TBYTE_PTR [b + i * s + d] -#define D_REGBISD(b, i, s, d) QWORD_PTR [b + i * s + d] -#define L_REGBISD(b, i, s, d) DWORD_PTR [b + i * s + d] -#define W_REGBISD(b, i, s, d) WORD_PTR [b + i * s + d] -#define B_REGBISD(b, i, s, d) BYTE_PTR [b + i * s + d] - -/* Displaced Scaled Index: */ -#define P_REGDIS(d, i, s) [i * s + d] -#define X_REGDIS(d, i, s) TBYTE_PTR [i * s + d] -#define D_REGDIS(d, i, s) QWORD_PTR [i * s + d] -#define L_REGDIS(d, i, s) DWORD_PTR [i * s + d] -#define W_REGDIS(d, i, s) WORD_PTR [i * s + d] -#define B_REGDIS(d, i, s) BYTE_PTR [i * s + d] - -/* Indexed Base: */ -#define P_REGBI(b, i) [b + i] -#define X_REGBI(b, i) TBYTE_PTR [b + i] -#define D_REGBI(b, i) QWORD_PTR [b + i] -#define L_REGBI(b, i) DWORD_PTR [b + i] -#define W_REGBI(b, i) WORD_PTR [b + i] -#define B_REGBI(b, i) BYTE_PTR [b + i] - -/* Displaced Base: */ -#define P_REGDB(d, b) [b + d] -#define X_REGDB(d, b) TBYTE_PTR [b + d] -#define D_REGDB(d, b) QWORD_PTR [b + d] -#define L_REGDB(d, b) DWORD_PTR [b + d] -#define W_REGDB(d, b) WORD_PTR [b + d] -#define B_REGDB(d, b) BYTE_PTR [b + d] - -/* Variable indirect: */ -#define VARINDIRECT(var) [var] - -/* Use register contents as jump/call target: */ -#define CODEPTR(reg) P_(reg) - -/* - * Redefine assembler commands - */ - -#define P_(a) P_ ## a -#define X_(a) X_ ## a -#define D_(a) D_ ## a -#define SR_(a) W_ ## a -#define S_(a) L_ ## a -#define L_(a) L_ ## a -#define W_(a) W_ ## a -#define B_(a) B_ ## a - -#define AAA aaa -#define AAD aad -#define AAM aam -#define AAS aas -#define ADC_L(a, b) adc L_(b), L_(a) -#define ADC_W(a, b) adc W_(b), W_(a) -#define ADC_B(a, b) adc B_(b), B_(a) -#define ADD_L(a, b) add L_(b), L_(a) -#define ADD_W(a, b) add W_(b), W_(a) -#define ADD_B(a, b) add B_(b), B_(a) -#define AND_L(a, b) and L_(b), L_(a) -#define AND_W(a, b) and W_(b), W_(a) -#define AND_B(a, b) and B_(b), B_(a) -#define ARPL(a,b) arpl W_(b), a -#define BOUND_L(a, b) bound L_(b), L_(a) -#define BOUND_W(a, b) bound W_(b), W_(a) -#define BSF_L(a, b) bsf L_(b), L_(a) -#define BSF_W(a, b) bsf W_(b), W_(a) -#define BSR_L(a, b) bsr L_(b), L_(a) -#define BSR_W(a, b) bsr W_(b), W_(a) -#define BT_L(a, b) bt L_(b), L_(a) -#define BT_W(a, b) bt W_(b), W_(a) -#define BTC_L(a, b) btc L_(b), L_(a) -#define BTC_W(a, b) btc W_(b), W_(a) -#define BTR_L(a, b) btr L_(b), L_(a) -#define BTR_W(a, b) btr W_(b), W_(a) -#define BTS_L(a, b) bts L_(b), L_(a) -#define BTS_W(a, b) bts W_(b), W_(a) -#define CALL(a) call a -#define CALLF(s,a) call far s:a -#define CBW cbw -#define CWDE cwde -#define CLC clc -#define CLD cld -#define CLI cli -#define CLTS clts -#define CMC cmc -#define CMP_L(a, b) cmp L_(b), L_(a) -#define CMP_W(a, b) cmp W_(b), W_(a) -#define CMP_B(a, b) cmp B_(b), B_(a) -#define CMPS_L cmpsd -#define CMPS_W cmpsw -#define CMPS_B cmpsb -#define CPUID cpuid -#define CWD cwd -#define CDQ cdq -#define DAA daa -#define DAS das -#define DEC_L(a) dec L_(a) -#define DEC_W(a) dec W_(a) -#define DEC_B(a) dec B_(a) -#define DIV_L(a) div L_(a) -#define DIV_W(a) div W_(a) -#define DIV_B(a) div B_(a) -#define ENTER(a,b) enter b, a -#define HLT hlt -#define IDIV_L(a) idiv L_(a) -#define IDIV_W(a) idiv W_(a) -#define IDIV_B(a) idiv B_(a) -#define IMUL_L(a, b) imul L_(b), L_(a) -#define IMUL_W(a, b) imul W_(b), W_(a) -#define IMUL_B(a) imul B_(a) -#define IN_L in EAX, DX -#define IN_W in AX, DX -#define IN_B in AL, DX -#define IN1_L(a) in1 L_(a) -#define IN1_W(a) in1 W_(a) -#define IN1_B(a) in1 B_(a) -#define INC_L(a) inc L_(a) -#define INC_W(a) inc W_(a) -#define INC_B(a) inc B_(a) -#define INS_L ins -#define INS_W ins -#define INS_B ins -#define INT(a) int B_(a) -#define INT3 int3 -#define INTO into -#define IRET iret -#define IRETD iretd -#define JA(a) ja NEAR a -#define JAE(a) jae NEAR a -#define JB(a) jb NEAR a -#define JBE(a) jbe NEAR a -#define JC(a) jc NEAR a -#define JE(a) je NEAR a -#define JG(a) jg NEAR a -#define JGE(a) jge NEAR a -#define JL(a) jl NEAR a -#define JLE(a) jle NEAR a -#define JNA(a) jna NEAR a -#define JNAE(a) jnae NEAR a -#define JNB(a) jnb NEAR a -#define JNBE(a) jnbe NEAR a -#define JNC(a) jnc NEAR a -#define JNE(a) jne NEAR a -#define JNG(a) jng NEAR a -#define JNGE(a) jnge NEAR a -#define JNL(a) jnl NEAR a -#define JNLE(a) jnle NEAR a -#define JNO(a) jno NEAR a -#define JNP(a) jnp NEAR a -#define JNS(a) jns NEAR a -#define JNZ(a) jnz NEAR a -#define JO(a) jo NEAR a -#define JP(a) jp NEAR a -#define JPE(a) jpe NEAR a -#define JPO(a) jpo NEAR a -#define JS(a) js NEAR a -#define JZ(a) jz NEAR a -#define JMP(a) jmp a -#define JMPF(s,a) jmp far s:a -#define LAHF lahf -#define LAR(a, b) lar b, a -#define LEA_L(a, b) lea P_(b), P_(a) -#define LEA_W(a, b) lea P_(b), P_(a) -#define LEAVE leave -#define LGDT(a) lgdt a -#define LIDT(a) lidt a -#define LDS(a, b) lds b, P_(a) -#define LES(a, b) les b, P_(a) -#define LFS(a, b) lfs b, P_(a) -#define LGS(a, b) lgs b, P_(a) -#define LSS(a, b) lss b, P_(a) -#define LLDT(a) lldt a -#define LMSW(a) lmsw a -#define LOCK lock -#define LODS_L lodsd -#define LODS_W lodsw -#define LODS_B lodsb -#define LOOP(a) loop a -#define LOOPE(a) loope a -#define LOOPZ(a) loopz a -#define LOOPNE(a) loopne a -#define LOOPNZ(a) loopnz a -#define LSL(a, b) lsl b, a -#define LTR(a) ltr a -#define MOV_SR(a, b) mov SR_(b), SR_(a) -#define MOV_L(a, b) mov L_(b), L_(a) -#define MOV_W(a, b) mov W_(b), W_(a) -#define MOV_B(a, b) mov B_(b), B_(a) -#define MOVS_L movsd -#define MOVS_W movsw -#define MOVS_B movsb -#define MOVSX_BL(a, b) movsx B_(b), B_(a) -#define MOVSX_BW(a, b) movsx B_(b), B_(a) -#define MOVSX_WL(a, b) movsx W_(b), W_(a) -#define MOVZX_BL(a, b) movzx B_(b), B_(a) -#define MOVZX_BW(a, b) movzx B_(b), B_(a) -#define MOVZX_WL(a, b) movzx W_(b), W_(a) -#define MUL_L(a) mul L_(a) -#define MUL_W(a) mul W_(a) -#define MUL_B(a) mul B_(a) -#define NEG_L(a) neg L_(a) -#define NEG_W(a) neg W_(a) -#define NEG_B(a) neg B_(a) -#define NOP nop -#define NOT_L(a) not L_(a) -#define NOT_W(a) not W_(a) -#define NOT_B(a) not B_(a) -#define OR_L(a,b) or L_(b), L_(a) -#define OR_W(a,b) or W_(b), W_(a) -#define OR_B(a,b) or B_(b), B_(a) -#define OUT_L out DX, EAX -#define OUT_W out DX, AX -#define OUT_B out DX, AL -#define OUT1_L(a) out1 L_(a) -#define OUT1_W(a) out1 W_(a) -#define OUT1_B(a) out1 B_(a) -#define OUTS_L outsd -#define OUTS_W outsw -#define OUTS_B outsb -#define POP_SR(a) pop SR_(a) -#define POP_L(a) pop L_(a) -#define POP_W(a) pop W_(a) -#define POPA_L popad -#define POPA_W popa -#define POPF_L popfd -#define POPF_W popf -#define PUSH_SR(a) push SR_(a) -#define PUSH_L(a) push L_(a) -#define PUSH_W(a) push W_(a) -#define PUSH_B(a) push B_(a) -#define PUSHA_L pushad -#define PUSHA_W pusha -#define PUSHF_L pushfd -#define PUSHF_W pushf -#define RCL_L(a, b) rcl L_(b), L_(a) -#define RCL_W(a, b) rcl W_(b), W_(a) -#define RCL_B(a, b) rcl B_(b), B_(a) -#define RCR_L(a, b) rcr L_(b), L_(a) -#define RCR_W(a, b) rcr W_(b), W_(a) -#define RCR_B(a, b) rcr B_(b), B_(a) -#define RDTSC rdtsc -#define ROL_L(a, b) rol L_(b), L_(a) -#define ROL_W(a, b) rol W_(b), W_(a) -#define ROL_B(a, b) rol B_(b), B_(a) -#define ROR_L(a, b) ror L_(b), L_(a) -#define ROR_W(a, b) ror W_(b), W_(a) -#define ROR_B(a, b) ror B_(b), B_(a) -#define REP rep -#define REPE repe -#define REPNE repne -#define REPNZ REPNE -#define REPZ REPE -#define RET ret -#define SAHF sahf -#define SAL_L(a, b) sal L_(b), B_(a) -#define SAL_W(a, b) sal W_(b), B_(a) -#define SAL_B(a, b) sal B_(b), B_(a) -#define SAR_L(a, b) sar L_(b), B_(a) -#define SAR_W(a, b) sar W_(b), B_(a) -#define SAR_B(a, b) sar B_(b), B_(a) -#define SBB_L(a, b) sbb L_(b), L_(a) -#define SBB_W(a, b) sbb W_(b), W_(a) -#define SBB_B(a, b) sbb B_(b), B_(a) -#define SCAS_L scas -#define SCAS_W scas -#define SCAS_B scas -#define SETA(a) seta a -#define SETAE(a) setae a -#define SETB(a) setb a -#define SETBE(a) setbe a -#define SETC(a) setc a -#define SETE(a) sete a -#define SETG(a) setg a -#define SETGE(a) setge a -#define SETL(a) setl a -#define SETLE(a) setle a -#define SETNA(a) setna a -#define SETNAE(a) setnae a -#define SETNB(a) setnb a -#define SETNBE(a) setnbe a -#define SETNC(a) setnc a -#define SETNE(a) setne a -#define SETNG(a) setng a -#define SETNGE(a) setnge a -#define SETNL(a) setnl a -#define SETNLE(a) setnle a -#define SETNO(a) setno a -#define SETNP(a) setnp a -#define SETNS(a) setns a -#define SETNZ(a) setnz a -#define SETO(a) seto a -#define SETP(a) setp a -#define SETPE(a) setpe a -#define SETPO(a) setpo a -#define SETS(a) sets a -#define SETZ(a) setz a -#define SGDT(a) sgdt a -#define SIDT(a) sidt a -#define SHL_L(a, b) shl L_(b), B_(a) -#define SHL_W(a, b) shl W_(b), B_(a) -#define SHL_B(a, b) shl B_(b), B_(a) -#define SHLD_L(a,b,c) shld -#define SHLD2_L(a,b) shld L_(b), L_(a) -#define SHLD_W(a,b,c) shld -#define SHLD2_W(a,b) shld W_(b), W_(a) -#define SHR_L(a, b) shr L_(b), B_(a) -#define SHR_W(a, b) shr W_(b), B_(a) -#define SHR_B(a, b) shr B_(b), B_(a) -#define SHRD_L(a,b,c) shrd -#define SHRD2_L(a,b) shrd L_(b), L_(a) -#define SHRD_W(a,b,c) shrd -#define SHRD2_W(a,b) shrd W_(b), W_(a) -#define SLDT(a) sldt a -#define SMSW(a) smsw a -#define STC stc -#define STD std -#define STI sti -#define STOS_L stosd -#define STOS_W stosw -#define STOS_B stosb -#define STR(a) str a -#define SUB_L(a, b) sub L_(b), L_(a) -#define SUB_W(a, b) sub W_(b), W_(a) -#define SUB_B(a, b) sub B_(b), B_(a) -#define TEST_L(a, b) test L_(b), L_(a) -#define TEST_W(a, b) test W_(b), W_(a) -#define TEST_B(a, b) test B_(b), B_(a) -#define VERR(a) verr a -#define VERW(a) verw a -#define WAIT wait -#define XCHG_L(a, b) xchg L_(b), L_(a) -#define XCHG_W(a, b) xchg W_(b), W_(a) -#define XCHG_B(a, b) xchg B_(b), B_(a) -#define XLAT xlat -#define XOR_L(a, b) xor L_(b), L_(a) -#define XOR_W(a, b) xor W_(b), W_(a) -#define XOR_B(a, b) xor B_(b), B_(a) - - -/* Floating Point Instructions */ -#define F2XM1 f2xm1 -#define FABS fabs -#define FADD_D(a) fadd D_(a) -#define FADD_S(a) fadd S_(a) -#define FADD2(a, b) fadd b, a -#define FADDP(a, b) faddp b, a -#define FIADD_L(a) fiadd L_(a) -#define FIADD_W(a) fiadd W_(a) -#define FBLD(a) fbld a -#define FBSTP(a) fbstp a -#define FCHS fchs -#define FCLEX fclex -#define FNCLEX fnclex -#define FCOM(a) fcom a -#define FCOM_D(a) fcom D_(a) -#define FCOM_S(a) fcom S_(a) -#define FCOMP(a) fcomp a -#define FCOMP_D(a) fcomp D_(a) -#define FCOMP_S(a) fcomp S_(a) -#define FCOMPP fcompp -#define FCOS fcos -#define FDECSTP fdecstp -#define FDIV_D(a) fdiv D_(a) -#define FDIV_S(a) fdiv S_(a) -#define FDIV2(a, b) fdiv b, a -#define FDIVP(a, b) fdivp b, a -#define FIDIV_L(a) fidiv L_(a) -#define FIDIV_W(a) fidiv W_(a) -#define FDIVR_D(a) fdivr D_(a) -#define FDIVR_S(a) fdivr S_(a) -#define FDIVR2(a, b) fdivr b, a -#define FDIVRP(a, b) fdivrp b, a -#define FIDIVR_L(a) fidivr L_(a) -#define FIDIVR_W(a) fidivr W_(a) -#define FFREE(a) ffree a -#define FICOM_L(a) ficom L_(a) -#define FICOM_W(a) ficom W_(a) -#define FICOMP_L(a) ficomp L_(a) -#define FICOMP_W(a) ficomp W_(a) -#define FILD_Q(a) fild D_(a) -#define FILD_L(a) fild L_(a) -#define FILD_W(a) fild W_(a) -#define FINCSTP fincstp -#define FINIT finit -#define FNINIT fninit -#define FIST_L(a) fist L_(a) -#define FIST_W(a) fist W_(a) -#define FISTP_Q(a) fistp D_(a) -#define FISTP_L(a) fistp L_(a) -#define FISTP_W(a) fistp W_(a) -#define FLD_X(a) fld X_(a) -#define FLD_D(a) fld D_(a) -#define FLD_S(a) fld S_(a) -#define FLD1 fld1 -#define FLDL2T fldl2t -#define FLDL2E fldl2e -#define FLDPI fldpi -#define FLDLG2 fldlg2 -#define FLDLN2 fldln2 -#define FLDZ fldz -#define FLDCW(a) fldcw a -#define FLDENV(a) fldenv a -#define FMUL_S(a) fmul S_(a) -#define FMUL_D(a) fmul D_(a) -#define FMUL2(a, b) fmul b, a -#define FMULP(a, b) fmulp b, a -#define FIMUL_L(a) fimul L_(a) -#define FIMUL_W(a) fimul W_(a) -#define FNOP fnop -#define FPATAN fpatan -#define FPREM fprem -#define FPREM1 fprem1 -#define FPTAN fptan -#define FRNDINT frndint -#define FRSTOR(a) frstor a -#define FSAVE(a) fsave a -#define FNSAVE(a) fnsave a -#define FSCALE fscale -#define FSIN fsin -#define FSINCOS fsincos -#define FSQRT fsqrt -#define FST_D(a) fst D_(a) -#define FST_S(a) fst S_(a) -#define FSTP_X(a) fstp X_(a) -#define FSTP_D(a) fstp D_(a) -#define FSTP_S(a) fstp S_(a) -#define FSTP(a) fstp a -#define FSTCW(a) fstcw a -#define FNSTCW(a) fnstcw a -#define FSTENV(a) fstenv a -#define FNSTENV(a) fnstenv a -#define FSTSW(a) fstsw a -#define FNSTSW(a) fnstsw a -#define FSUB_S(a) fsub S_(a) -#define FSUB_D(a) fsub D_(a) -#define FSUB2(a, b) fsub b, a -#define FSUBP(a, b) fsubp b, a -#define FISUB_L(a) fisub L_(a) -#define FISUB_W(a) fisub W_(a) -#define FSUBR_S(a) fsubr S_(a) -#define FSUBR_D(a) fsubr D_(a) -#define FSUBR2(a, b) fsubr b, a -#define FSUBRP(a, b) fsubrp b, a -#define FISUBR_L(a) fisubr L_(a) -#define FISUBR_W(a) fisubr W_(a) -#define FTST ftst -#define FUCOM(a) fucom a -#define FUCOMP(a) fucomp a -#define FUCOMPP fucompp -#define FWAIT fwait -#define FXAM fxam -#define FXCH(a) fxch a -#define FXTRACT fxtract -#define FYL2X fyl2x -#define FYL2XP1 fyl2xp1 - -#endif /* NASM_ASSEMBLER, MASM_ASSEMBLER */ - - /****************************************/ - /* */ - /* Extensions to x86 insn set - */ - /* MMX, 3DNow! */ - /* */ - /****************************************/ - -#if defined(NASM_ASSEMBLER) || defined(MASM_ASSEMBLER) -#define P_ARG1(a) P_ ## a -#define P_ARG2(a, b) P_ ## b, P_ ## a -#define P_ARG3(a, b, c) P_ ## c, P_ ## b, P_ ## a -#else -#define P_ARG1(a) a -#define P_ARG2(a, b) a, b -#define P_ARG3(a, b, c) a, b, c -#endif - -/* MMX */ -#define MOVD(a, b) movd P_ARG2(a, b) -#define MOVQ(a, b) movq P_ARG2(a, b) - -#define PADDB(a, b) paddb P_ARG2(a, b) -#define PADDW(a, b) paddw P_ARG2(a, b) -#define PADDD(a, b) paddd P_ARG2(a, b) - -#define PADDSB(a, b) paddsb P_ARG2(a, b) -#define PADDSW(a, b) paddsw P_ARG2(a, b) - -#define PADDUSB(a, b) paddusb P_ARG2(a, b) -#define PADDUSW(a, b) paddusw P_ARG2(a, b) - -#define PSUBB(a, b) psubb P_ARG2(a, b) -#define PSUBW(a, b) psubw P_ARG2(a, b) -#define PSUBD(a, b) psubd P_ARG2(a, b) - -#define PSUBSB(a, b) psubsb P_ARG2(a, b) -#define PSUBSW(a, b) psubsw P_ARG2(a, b) - -#define PSUBUSB(a, b) psubusb P_ARG2(a, b) -#define PSUBUSW(a, b) psubusw P_ARG2(a, b) - -#define PCMPEQB(a, b) pcmpeqb P_ARG2(a, b) -#define PCMPEQW(a, b) pcmpeqw P_ARG2(a, b) -#define PCMPEQD(a, b) pcmpeqd P_ARG2(a, b) - -#define PCMPGTB(a, b) pcmpgtb P_ARG2(a, b) -#define PCMPGTW(a, b) pcmpgtw P_ARG2(a, b) -#define PCMPGTD(a, b) pcmpgtd P_ARG2(a, b) - -#define PMULHW(a, b) pmulhw P_ARG2(a, b) -#define PMULLW(a, b) pmullw P_ARG2(a, b) - -#define PMADDWD(a, b) pmaddwd P_ARG2(a, b) - -#define PAND(a, b) pand P_ARG2(a, b) - -#define PANDN(a, b) pandn P_ARG2(a, b) - -#define POR(a, b) por P_ARG2(a, b) - -#define PXOR(a, b) pxor P_ARG2(a, b) - -#define PSRAW(a, b) psraw P_ARG2(a, b) -#define PSRAD(a, b) psrad P_ARG2(a, b) - -#define PSRLW(a, b) psrlw P_ARG2(a, b) -#define PSRLD(a, b) psrld P_ARG2(a, b) -#define PSRLQ(a, b) psrlq P_ARG2(a, b) - -#define PSLLW(a, b) psllw P_ARG2(a, b) -#define PSLLD(a, b) pslld P_ARG2(a, b) -#define PSLLQ(a, b) psllq P_ARG2(a, b) - -#define PACKSSWB(a, b) packsswb P_ARG2(a, b) -#define PACKSSDW(a, b) packssdw P_ARG2(a, b) -#define PACKUSWB(a, b) packuswb P_ARG2(a, b) - -#define PUNPCKHBW(a, b) punpckhbw P_ARG2(a, b) -#define PUNPCKHWD(a, b) punpckhwd P_ARG2(a, b) -#define PUNPCKHDQ(a, b) punpckhdq P_ARG2(a, b) -#define PUNPCKLBW(a, b) punpcklbw P_ARG2(a, b) -#define PUNPCKLWD(a, b) punpcklwd P_ARG2(a, b) -#define PUNPCKLDQ(a, b) punpckldq P_ARG2(a, b) - -#define EMMS emms - -/* AMD 3DNow! */ -#define PAVGUSB(a, b) pavgusb P_ARG2(a, b) -#define PFADD(a, b) pfadd P_ARG2(a, b) -#define PFSUB(a, b) pfsub P_ARG2(a, b) -#define PFSUBR(a, b) pfsubr P_ARG2(a, b) -#define PFACC(a, b) pfacc P_ARG2(a, b) -#define PFCMPGE(a, b) pfcmpge P_ARG2(a, b) -#define PFCMPGT(a, b) pfcmpgt P_ARG2(a, b) -#define PFCMPEQ(a, b) pfcmpeq P_ARG2(a, b) -#define PFMIN(a, b) pfmin P_ARG2(a, b) -#define PFMAX(a, b) pfmax P_ARG2(a, b) -#define PI2FD(a, b) pi2fd P_ARG2(a, b) -#define PF2ID(a, b) pf2id P_ARG2(a, b) -#define PFRCP(a, b) pfrcp P_ARG2(a, b) -#define PFRSQRT(a, b) pfrsqrt P_ARG2(a, b) -#define PFMUL(a, b) pfmul P_ARG2(a, b) -#define PFRCPIT1(a, b) pfrcpit1 P_ARG2(a, b) -#define PFRSQIT1(a, b) pfrsqit1 P_ARG2(a, b) -#define PFRCPIT2(a, b) pfrcpit2 P_ARG2(a, b) -#define PMULHRW(a, b) pmulhrw P_ARG2(a, b) - -#define FEMMS femms -#define PREFETCH(a) prefetch P_ARG1(a) -#define PREFETCHW(a) prefetchw P_ARG1(a) - -/* Intel SSE */ -#define ADDPS(a, b) addps P_ARG2(a, b) -#define ADDSS(a, b) addss P_ARG2(a, b) -#define ANDNPS(a, b) andnps P_ARG2(a, b) -#define ANDPS(a, b) andps P_ARG2(a, b) -/* NASM only knows the pseudo ops for these. -#define CMPPS(a, b, c) cmpps P_ARG3(a, b, c) -#define CMPSS(a, b, c) cmpss P_ARG3(a, b, c) -*/ -#define CMPEQPS(a, b) cmpeqps P_ARG2(a, b) -#define CMPLTPS(a, b) cmpltps P_ARG2(a, b) -#define CMPLEPS(a, b) cmpleps P_ARG2(a, b) -#define CMPUNORDPS(a, b) cmpunordps P_ARG2(a, b) -#define CMPNEQPS(a, b) cmpneqps P_ARG2(a, b) -#define CMPNLTPS(a, b) cmpnltps P_ARG2(a, b) -#define CMPNLEPS(a, b) cmpnleps P_ARG2(a, b) -#define CMPORDPS(a, b) cmpordps P_ARG2(a, b) -#define CMPEQSS(a, b) cmpeqss P_ARG2(a, b) -#define CMPLTSS(a, b) cmpltss P_ARG2(a, b) -#define CMPLESS(a, b) cmpless P_ARG2(a, b) -#define CMPUNORDSS(a, b) cmpunordss P_ARG2(a, b) -#define CMPNEQSS(a, b) cmpneqss P_ARG2(a, b) -#define CMPNLTSS(a, b) cmpnltss P_ARG2(a, b) -#define CMPNLESS(a, b) cmpnless P_ARG2(a, b) -#define CMPORDSS(a, b) cmpordss P_ARG2(a, b) -#define COMISS(a, b) comiss P_ARG2(a, b) -#define CVTPI2PS(a, b) cvtpi2ps P_ARG2(a, b) -#define CVTPS2PI(a, b) cvtps2pi P_ARG2(a, b) -#define CVTSI2SS(a, b) cvtsi2ss P_ARG2(a, b) -#define CVTSS2SI(a, b) cvtss2si P_ARG2(a, b) -#define CVTTPS2PI(a, b) cvttps2pi P_ARG2(a, b) -#define CVTTSS2SI(a, b) cvttss2si P_ARG2(a, b) -#define DIVPS(a, b) divps P_ARG2(a, b) -#define DIVSS(a, b) divss P_ARG2(a, b) -#define FXRSTOR(a) fxrstor P_ARG1(a) -#define FXSAVE(a) fxsave P_ARG1(a) -#define LDMXCSR(a) ldmxcsr P_ARG1(a) -#define MAXPS(a, b) maxps P_ARG2(a, b) -#define MAXSS(a, b) maxss P_ARG2(a, b) -#define MINPS(a, b) minps P_ARG2(a, b) -#define MINSS(a, b) minss P_ARG2(a, b) -#define MOVAPS(a, b) movaps P_ARG2(a, b) -#define MOVHLPS(a, b) movhlps P_ARG2(a, b) -#define MOVHPS(a, b) movhps P_ARG2(a, b) -#define MOVLHPS(a, b) movlhps P_ARG2(a, b) -#define MOVLPS(a, b) movlps P_ARG2(a, b) -#define MOVMSKPS(a, b) movmskps P_ARG2(a, b) -#define MOVNTPS(a, b) movntps P_ARG2(a, b) -#define MOVNTQ(a, b) movntq P_ARG2(a, b) -#define MOVSS(a, b) movss P_ARG2(a, b) -#define MOVUPS(a, b) movups P_ARG2(a, b) -#define MULPS(a, b) mulps P_ARG2(a, b) -#define MULSS(a, b) mulss P_ARG2(a, b) -#define ORPS(a, b) orps P_ARG2(a, b) -#define RCPPS(a, b) rcpps P_ARG2(a, b) -#define RCPSS(a, b) rcpss P_ARG2(a, b) -#define RSQRTPS(a, b) rsqrtps P_ARG2(a, b) -#define RSQRTSS(a, b) rsqrtss P_ARG2(a, b) -#define SHUFPS(a, b, c) shufps P_ARG3(a, b, c) -#define SQRTPS(a, b) sqrtps P_ARG2(a, b) -#define SQRTSS(a, b) sqrtss P_ARG2(a, b) -#define STMXCSR(a) stmxcsr P_ARG1(a) -#define SUBPS(a, b) subps P_ARG2(a, b) -#define UCOMISS(a, b) ucomiss P_ARG2(a, b) -#define UNPCKHPS(a, b) unpckhps P_ARG2(a, b) -#define UNPCKLPS(a, b) unpcklps P_ARG2(a, b) -#define XORPS(a, b) xorps P_ARG2(a, b) - -#define PREFETCHNTA(a) prefetchnta P_ARG1(a) -#define PREFETCHT0(a) prefetcht0 P_ARG1(a) -#define PREFETCHT1(a) prefetcht1 P_ARG1(a) -#define PREFETCHT2(a) prefetcht2 P_ARG1(a) -#define SFENCE sfence - -/* Added by BrianP for FreeBSD (per David Dawes) */ -#if !defined(NASM_ASSEMBLER) && !defined(MASM_ASSEMBLER) && !defined(__bsdi__) -#define LLBL(a) CONCAT(.L,a) -#define LLBL2(a,b) CONCAT3(.L,a,b) -#else -#define LLBL(a) a -#define LLBL2(a,b) CONCAT(a,b) -#endif - -/* Segment overrides */ -#define SEGCS D_BYTE 46 -#define SEGDS D_BYTE 62 -#define SEGES D_BYTE 38 -#define SEGFS D_BYTE 100 -#define SEGGS D_BYTE 101 - -/* Temporary labels: valid until next non-local label */ -#ifdef NASM_ASSEMBLER -#define TLBL(a) CONCAT(.,a) -#else -#define TLBL(a) CONCAT(a,$) -#endif - -/* Hidden symbol visibility support. - * If we build with gcc's -fvisibility=hidden flag, we'll need to change - * the symbol visibility mode to 'default'. - */ -#if defined(GNU_ASSEMBLER) && !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) -# define HIDDEN(x) .hidden x -#elif defined(__GNUC__) && !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) -# pragma GCC visibility push(default) -# define HIDDEN(x) .hidden x -#else -# define HIDDEN(x) -#endif - -#endif /* __ASSYNTAX_H__ */ + +#ifndef __ASSYNTAX_H__ +#define __ASSYNTAX_H__ + +/* + * Copyright 1992 Vrije Universiteit, The Netherlands + * + * Permission to use, copy, modify, and distribute this software and its + * documentation for any purpose and without fee is hereby granted, provided + * that the above copyright notice appear in all copies and that both that + * copyright notice and this permission notice appear in supporting + * documentation, and that the name of the Vrije Universiteit not be used in + * advertising or publicity pertaining to distribution of the software without + * specific, written prior permission. The Vrije Universiteit makes no + * representations about the suitability of this software for any purpose. + * It is provided "as is" without express or implied warranty. + * + * The Vrije Universiteit DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS + * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, + * IN NO EVENT SHALL The Vrije Universiteit BE LIABLE FOR ANY SPECIAL, + * INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM + * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE + * OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR + * PERFORMANCE OF THIS SOFTWARE. + */ + +/* + * assyntax.h + * + * Select the syntax appropriate to the 386 assembler being used + * To add support for more assemblers add more columns to the CHOICE + * macro. Note that register names must also have uppercase names + * to avoid macro recursion. e.g., #define ah %ah recurses! + * + * NB 1. Some of the macros for certain assemblers imply that the code is to + * run in protected mode!! Caveat emptor. + * + * NB 2. 486 specific instructions are not included. This is to discourage + * their accidental use in code that is intended to run on 386 and 486 + * systems. + * + * Supported assemblers: + * + * (a) AT&T SysVr4 as(1): define ATT_ASSEMBLER + * (b) GNU Assembler gas: define GNU_ASSEMBLER (default) + * (c) Amsterdam Compiler kit: define ACK_ASSEMBLER + * (d) The Netwide Assembler: define NASM_ASSEMBLER + * (e) Microsoft Assembler: define MASM_ASSEMBLER (UNTESTED!) + * + * The following naming conventions have been used to identify the various + * data types: + * _SR = segment register version + * Integer: + * _Q = quadword = 64 bits + * _L = long = 32 bits + * _W = short = 16 bits + * _B = byte = 8 bits + * Floating-point: + * _X = m80real = 80 bits + * _D = double = 64 bits + * _S = single = 32 bits + * + * Author: Gregory J. Sharp, Sept 1992 + * Vrije Universiteit, Amsterdam, The Netherlands + * + * [support for Intel syntax added by Josh Vanderhoof, 1999] + */ + +#if !(defined(NASM_ASSEMBLER) || defined(MASM_ASSEMBLER)) + +/* Default to ATT_ASSEMBLER when SVR4 or SYSV are defined */ +#if (defined(SVR4) || defined(SYSV)) && !defined(GNU_ASSEMBLER) +#define ATT_ASSEMBLER +#endif + +#if !defined(ATT_ASSEMBLER) && !defined(GNU_ASSEMBLER) && !defined(ACK_ASSEMBLER) +#define GNU_ASSEMBLER +#endif + +#if (defined(__STDC__) && !defined(UNIXCPP)) || (defined (sun) && defined (i386) && defined (SVR4) && defined (__STDC__) && !defined (__GNUC__)) +#define CONCAT(x, y) x ## y +#define CONCAT3(x, y, z) x ## y ## z +#else +#define CONCAT(x, y) x/**/y +#define CONCAT3(x, y, z) x/**/y/**/z +#endif + +#ifdef ACK_ASSEMBLER + +/* Assume we write code for 32-bit protected mode! */ + +/* Redefine register names for GAS & AT&T assemblers */ +#define AL al +#define AH ah +#define AX ax +#define EAX ax +#define BL bl +#define BH bh +#define BX bx +#define EBX bx +#define CL cl +#define CH ch +#define CX cx +#define ECX cx +#define DL dl +#define DH dh +#define DX dx +#define EDX dx +#define BP bp +#define EBP bp +#define SI si +#define ESI si +#define DI di +#define EDI di +#define SP sp +#define ESP sp +#define CS cs +#define SS ss +#define DS ds +#define ES es +#define FS fs +#define GS gs +/* Control Registers */ +#define CR0 cr0 +#define CR1 cr1 +#define CR2 cr2 +#define CR3 cr3 +/* Debug Registers */ +#define DR0 dr0 +#define DR1 dr1 +#define DR2 dr2 +#define DR3 dr3 +#define DR4 dr4 +#define DR5 dr5 +#define DR6 dr6 +#define DR7 dr7 +/* Floating-point Stack */ +#define ST st + +#define AS_BEGIN .sect .text; .sect .rom; .sect .data; .sect .bss; .sect .text + + +#define _WTOG o16 /* word toggle for _W instructions */ +#define _LTOG /* long toggle for _L instructions */ +#define ADDR_TOGGLE a16 +#define OPSZ_TOGGLE o16 +#define USE16 .use16 +#define USE32 .use32 + +#define CHOICE(a,b,c) c + +#else /* AT&T or GAS */ + +/* Redefine register names for GAS & AT&T assemblers */ +#define AL %al +#define AH %ah +#define AX %ax +#define EAX %eax +#define BL %bl +#define BH %bh +#define BX %bx +#define EBX %ebx +#define CL %cl +#define CH %ch +#define CX %cx +#define ECX %ecx +#define DL %dl +#define DH %dh +#define DX %dx +#define EDX %edx +#define BP %bp +#define EBP %ebp +#define SI %si +#define ESI %esi +#define DI %di +#define EDI %edi +#define SP %sp +#define ESP %esp +#define CS %cs +#define SS %ss +#define DS %ds +#define ES %es +#define FS %fs +#define GS %gs +/* Control Registers */ +#define CR0 %cr0 +#define CR1 %cr1 +#define CR2 %cr2 +#define CR3 %cr3 +/* Debug Registers */ +#define DR0 %db0 +#define DR1 %db1 +#define DR2 %db2 +#define DR3 %db3 +#define DR4 %db4 +#define DR5 %db5 +#define DR6 %db6 +#define DR7 %db7 +/* Floating-point Stack */ +#define _STX0 %st(0) +#define _STX1 %st(1) +#define _STX2 %st(2) +#define _STX3 %st(3) +#define _STX4 %st(4) +#define _STX5 %st(5) +#define _STX6 %st(6) +#define _STX7 %st(7) +#define ST(x) CONCAT(_STX,x) +#ifdef GNU_ASSEMBLER +#define ST0 %st(0) +#else +#define ST0 %st +#endif +/* MMX Registers */ +#define MM0 %mm0 +#define MM1 %mm1 +#define MM2 %mm2 +#define MM3 %mm3 +#define MM4 %mm4 +#define MM5 %mm5 +#define MM6 %mm6 +#define MM7 %mm7 +/* SSE Registers */ +#define XMM0 %xmm0 +#define XMM1 %xmm1 +#define XMM2 %xmm2 +#define XMM3 %xmm3 +#define XMM4 %xmm4 +#define XMM5 %xmm5 +#define XMM6 %xmm6 +#define XMM7 %xmm7 + +#define AS_BEGIN +#define USE16 +#define USE32 + +#ifdef GNU_ASSEMBLER + +#define ADDR_TOGGLE aword +#define OPSZ_TOGGLE word + +#define CHOICE(a,b,c) b + +#else +/* + * AT&T ASSEMBLER SYNTAX + * ********************* + */ +#define CHOICE(a,b,c) a + +#define ADDR_TOGGLE addr16 +#define OPSZ_TOGGLE data16 + +#endif /* GNU_ASSEMBLER */ +#endif /* ACK_ASSEMBLER */ + + +#if defined(__QNX__) || defined(Lynx) || (defined(SYSV) || defined(SVR4)) && !defined(ACK_ASSEMBLER) || defined(__ELF__) || defined(__GNU__) || defined(__GNUC__) && !defined(__DJGPP__) && !defined(__MINGW32__) +#define GLNAME(a) a +#else +#define GLNAME(a) CONCAT(_,a) +#endif + + + /****************************************/ + /* */ + /* Select the various choices */ + /* */ + /****************************************/ + + +/* Redefine assembler directives */ +/*********************************/ +#define GLOBL CHOICE(.globl, .globl, .extern) +#define GLOBAL GLOBL +#define EXTERN GLOBL +#ifndef __AOUT__ +#define ALIGNTEXT32 CHOICE(.align 32, .balign 32, .align 32) +#define ALIGNTEXT16 CHOICE(.align 16, .balign 16, .align 16) +#define ALIGNTEXT8 CHOICE(.align 8, .balign 8, .align 8) +#define ALIGNTEXT4 CHOICE(.align 4, .balign 4, .align 4) +#define ALIGNTEXT2 CHOICE(.align 2, .balign 2, .align 2) +/* ALIGNTEXT4ifNOP is the same as ALIGNTEXT4, but only if the space is + * guaranteed to be filled with NOPs. Otherwise it does nothing. + */ +#define ALIGNTEXT32ifNOP CHOICE(.align 32, .balign ARG2(32,0x90), /*can't do it*/) +#define ALIGNTEXT16ifNOP CHOICE(.align 16, .balign ARG2(16,0x90), /*can't do it*/) +#define ALIGNTEXT8ifNOP CHOICE(.align 8, .balign ARG2(8,0x90), /*can't do it*/) +#define ALIGNTEXT4ifNOP CHOICE(.align 4, .balign ARG2(4,0x90), /*can't do it*/) +#define ALIGNDATA32 CHOICE(.align 32, .balign ARG2(32,0x0), .align 32) +#define ALIGNDATA16 CHOICE(.align 16, .balign ARG2(16,0x0), .align 16) +#define ALIGNDATA8 CHOICE(.align 8, .balign ARG2(8,0x0), .align 8) +#define ALIGNDATA4 CHOICE(.align 4, .balign ARG2(4,0x0), .align 4) +#define ALIGNDATA2 CHOICE(.align 2, .balign ARG2(2,0x0), .align 2) +#else +/* 'as -aout' on FreeBSD doesn't have .balign */ +#define ALIGNTEXT32 CHOICE(.align 32, .align ARG2(5,0x90), .align 32) +#define ALIGNTEXT16 CHOICE(.align 16, .align ARG2(4,0x90), .align 16) +#define ALIGNTEXT8 CHOICE(.align 8, .align ARG2(3,0x90), .align 8) +#define ALIGNTEXT4 CHOICE(.align 4, .align ARG2(2,0x90), .align 4) +#define ALIGNTEXT2 CHOICE(.align 2, .align ARG2(1,0x90), .align 2) +/* ALIGNTEXT4ifNOP is the same as ALIGNTEXT4, but only if the space is + * guaranteed to be filled with NOPs. Otherwise it does nothing. + */ +#define ALIGNTEXT32ifNOP CHOICE(.align 32, .align ARG2(5,0x90), /*can't do it*/) +#define ALIGNTEXT16ifNOP CHOICE(.align 16, .align ARG2(4,0x90), /*can't do it*/) +#define ALIGNTEXT8ifNOP CHOICE(.align 8, .align ARG2(3,0x90), /*can't do it*/) +#define ALIGNTEXT4ifNOP CHOICE(.align 4, .align ARG2(2,0x90), /*can't do it*/) +#define ALIGNDATA32 CHOICE(.align 32, .align ARG2(5,0x0), .align 32) +#define ALIGNDATA16 CHOICE(.align 16, .align ARG2(4,0x0), .align 16) +#define ALIGNDATA8 CHOICE(.align 8, .align ARG2(3,0x0), .align 8) +#define ALIGNDATA4 CHOICE(.align 4, .align ARG2(2,0x0), .align 4) +#define ALIGNDATA2 CHOICE(.align 2, .align ARG2(1,0x0), .align 2) +#endif /* __AOUT__ */ +#define FILE(s) CHOICE(.file s, .file s, .file s) +#define STRING(s) CHOICE(.string s, .asciz s, .asciz s) +#define D_LONG CHOICE(.long, .long, .data4) +#define D_WORD CHOICE(.value, .short, .data2) +#define D_BYTE CHOICE(.byte, .byte, .data1) +#define SPACE CHOICE(.comm, .space, .space) +#define COMM CHOICE(.comm, .comm, .comm) +#define SEG_DATA CHOICE(.data, .data, .sect .data) +#define SEG_TEXT CHOICE(.text, .text, .sect .text) +#define SEG_BSS CHOICE(.bss, .bss, .sect .bss) + +#ifdef GNU_ASSEMBLER +#define D_SPACE(n) . = . + n +#else +#define D_SPACE(n) .space n +#endif + +/* Addressing Modes */ +/* Immediate Mode */ +#define ADDR(a) CHOICE(CONCAT($,a), $a, a) +#define CONST(a) CHOICE(CONCAT($,a), $a, a) + +/* Indirect Mode */ +#define CONTENT(a) CHOICE(a, a, (a)) /* take contents of variable */ +#define REGIND(a) CHOICE((a), (a), (a)) /* Register a indirect */ +/* Register b indirect plus displacement a */ +#define REGOFF(a, b) CHOICE(a(b), a(b), a(b)) +/* Reg indirect Base + Index + Displacement - this is mainly for 16-bit mode + * which has no scaling + */ +#define REGBID(b,i,d) CHOICE(d(b,i), d(b,i), d(b)(i)) +/* Reg indirect Base + (Index * Scale) */ +#define REGBIS(b,i,s) CHOICE((b,i,s), (b,i,s), (b)(i*s)) +/* Reg indirect Base + (Index * Scale) + Displacement */ +#define REGBISD(b,i,s,d) CHOICE(d(b,i,s), d(b,i,s), d(b)(i*s)) +/* Displaced Scaled Index: */ +#define REGDIS(d,i,s) CHOICE(d(,i,s), d(,i,s), d(i * s)) +/* Indexed Base: */ +#define REGBI(b,i) CHOICE((b,i), (b,i), (b)(i)) +/* Displaced Base: */ +#define REGDB(d,b) CHOICE(d(b), d(b), d(b)) +/* Variable indirect: */ +#define VARINDIRECT(var) CHOICE(*var, *var, (var)) +/* Use register contents as jump/call target: */ +#define CODEPTR(reg) CHOICE(*reg, *reg, reg) + +/* For expressions requiring bracketing + * eg. (CRT0_PM | CRT_EM) + */ + +#define EXPR(a) CHOICE([a], (a), [a]) +#define ENOT(a) CHOICE(0!a, ~a, ~a) +#define EMUL(a,b) CHOICE(a\*b, a*b, a*b) +#define EDIV(a,b) CHOICE(a\/b, a/b, a/b) + +/* + * We have to beat the problem of commas within arguments to choice. + * eg. choice (add a,b, add b,a) will get argument mismatch. Luckily ANSI + * and other known cpp definitions evaluate arguments before substitution + * so the following works. + */ +#define ARG2(a, b) a,b +#define ARG3(a,b,c) a,b,c + +/* Redefine assembler commands */ +#define AAA CHOICE(aaa, aaa, aaa) +#define AAD CHOICE(aad, aad, aad) +#define AAM CHOICE(aam, aam, aam) +#define AAS CHOICE(aas, aas, aas) +#define ADC_L(a, b) CHOICE(adcl ARG2(a,b), adcl ARG2(a,b), _LTOG adc ARG2(b,a)) +#define ADC_W(a, b) CHOICE(adcw ARG2(a,b), adcw ARG2(a,b), _WTOG adc ARG2(b,a)) +#define ADC_B(a, b) CHOICE(adcb ARG2(a,b), adcb ARG2(a,b), adcb ARG2(b,a)) +#define ADD_L(a, b) CHOICE(addl ARG2(a,b), addl ARG2(a,b), _LTOG add ARG2(b,a)) +#define ADD_W(a, b) CHOICE(addw ARG2(a,b), addw ARG2(a,b), _WTOG add ARG2(b,a)) +#define ADD_B(a, b) CHOICE(addb ARG2(a,b), addb ARG2(a,b), addb ARG2(b,a)) +#define AND_L(a, b) CHOICE(andl ARG2(a,b), andl ARG2(a,b), _LTOG and ARG2(b,a)) +#define AND_W(a, b) CHOICE(andw ARG2(a,b), andw ARG2(a,b), _WTOG and ARG2(b,a)) +#define AND_B(a, b) CHOICE(andb ARG2(a,b), andb ARG2(a,b), andb ARG2(b,a)) +#define ARPL(a,b) CHOICE(arpl ARG2(a,b), arpl ARG2(a,b), arpl ARG2(b,a)) +#define BOUND_L(a, b) CHOICE(boundl ARG2(a,b), boundl ARG2(b,a), _LTOG bound ARG2(b,a)) +#define BOUND_W(a, b) CHOICE(boundw ARG2(a,b), boundw ARG2(b,a), _WTOG bound ARG2(b,a)) +#define BSF_L(a, b) CHOICE(bsfl ARG2(a,b), bsfl ARG2(a,b), _LTOG bsf ARG2(b,a)) +#define BSF_W(a, b) CHOICE(bsfw ARG2(a,b), bsfw ARG2(a,b), _WTOG bsf ARG2(b,a)) +#define BSR_L(a, b) CHOICE(bsrl ARG2(a,b), bsrl ARG2(a,b), _LTOG bsr ARG2(b,a)) +#define BSR_W(a, b) CHOICE(bsrw ARG2(a,b), bsrw ARG2(a,b), _WTOG bsr ARG2(b,a)) +#define BT_L(a, b) CHOICE(btl ARG2(a,b), btl ARG2(a,b), _LTOG bt ARG2(b,a)) +#define BT_W(a, b) CHOICE(btw ARG2(a,b), btw ARG2(a,b), _WTOG bt ARG2(b,a)) +#define BTC_L(a, b) CHOICE(btcl ARG2(a,b), btcl ARG2(a,b), _LTOG btc ARG2(b,a)) +#define BTC_W(a, b) CHOICE(btcw ARG2(a,b), btcw ARG2(a,b), _WTOG btc ARG2(b,a)) +#define BTR_L(a, b) CHOICE(btrl ARG2(a,b), btrl ARG2(a,b), _LTOG btr ARG2(b,a)) +#define BTR_W(a, b) CHOICE(btrw ARG2(a,b), btrw ARG2(a,b), _WTOG btr ARG2(b,a)) +#define BTS_L(a, b) CHOICE(btsl ARG2(a,b), btsl ARG2(a,b), _LTOG bts ARG2(b,a)) +#define BTS_W(a, b) CHOICE(btsw ARG2(a,b), btsw ARG2(a,b), _WTOG bts ARG2(b,a)) +#define CALL(a) CHOICE(call a, call a, call a) +#define CALLF(s,a) CHOICE(lcall ARG2(s,a), lcall ARG2(s,a), callf s:a) +#define CBW CHOICE(cbtw, cbw, cbw) +#define CWDE CHOICE(cwtd, cwde, cwde) +#define CLC CHOICE(clc, clc, clc) +#define CLD CHOICE(cld, cld, cld) +#define CLI CHOICE(cli, cli, cli) +#define CLTS CHOICE(clts, clts, clts) +#define CMC CHOICE(cmc, cmc, cmc) +#define CMP_L(a, b) CHOICE(cmpl ARG2(a,b), cmpl ARG2(a,b), _LTOG cmp ARG2(b,a)) +#define CMP_W(a, b) CHOICE(cmpw ARG2(a,b), cmpw ARG2(a,b), _WTOG cmp ARG2(b,a)) +#define CMP_B(a, b) CHOICE(cmpb ARG2(a,b), cmpb ARG2(a,b), cmpb ARG2(b,a)) +#define CMPS_L CHOICE(cmpsl, cmpsl, _LTOG cmps) +#define CMPS_W CHOICE(cmpsw, cmpsw, _WTOG cmps) +#define CMPS_B CHOICE(cmpsb, cmpsb, cmpsb) +#define CWD CHOICE(cwtl, cwd, cwd) +#define CDQ CHOICE(cltd, cdq, cdq) +#define DAA CHOICE(daa, daa, daa) +#define DAS CHOICE(das, das, das) +#define DEC_L(a) CHOICE(decl a, decl a, _LTOG dec a) +#define DEC_W(a) CHOICE(decw a, decw a, _WTOG dec a) +#define DEC_B(a) CHOICE(decb a, decb a, decb a) +#define DIV_L(a) CHOICE(divl a, divl a, div a) +#define DIV_W(a) CHOICE(divw a, divw a, div a) +#define DIV_B(a) CHOICE(divb a, divb a, divb a) +#define ENTER(a,b) CHOICE(enter ARG2(a,b), enter ARG2(a,b), enter ARG2(b,a)) +#define HLT CHOICE(hlt, hlt, hlt) +#define IDIV_L(a) CHOICE(idivl a, idivl a, _LTOG idiv a) +#define IDIV_W(a) CHOICE(idivw a, idivw a, _WTOG idiv a) +#define IDIV_B(a) CHOICE(idivb a, idivb a, idivb a) +/* More forms than this for imul!! */ +#define IMUL_L(a, b) CHOICE(imull ARG2(a,b), imull ARG2(a,b), _LTOG imul ARG2(b,a)) +#define IMUL_W(a, b) CHOICE(imulw ARG2(a,b), imulw ARG2(a,b), _WTOG imul ARG2(b,a)) +#define IMUL_B(a) CHOICE(imulb a, imulb a, imulb a) +#define IN_L CHOICE(inl (DX), inl ARG2(DX,EAX), _LTOG in DX) +#define IN_W CHOICE(inw (DX), inw ARG2(DX,AX), _WTOG in DX) +#define IN_B CHOICE(inb (DX), inb ARG2(DX,AL), inb DX) +/* Please AS code writer: use the following ONLY, if you refer to ports<256 + * directly, but not in IN1_W(DX), for instance, even if IN1_ looks nicer + */ +#if defined (sun) +#define IN1_L(a) CHOICE(inl (a), inl ARG2(a,EAX), _LTOG in a) +#define IN1_W(a) CHOICE(inw (a), inw ARG2(a,AX), _WTOG in a) +#define IN1_B(a) CHOICE(inb (a), inb ARG2(a,AL), inb a) +#else +#define IN1_L(a) CHOICE(inl a, inl ARG2(a,EAX), _LTOG in a) +#define IN1_W(a) CHOICE(inw a, inw ARG2(a,AX), _WTOG in a) +#define IN1_B(a) CHOICE(inb a, inb ARG2(a,AL), inb a) +#endif +#define INC_L(a) CHOICE(incl a, incl a, _LTOG inc a) +#define INC_W(a) CHOICE(incw a, incw a, _WTOG inc a) +#define INC_B(a) CHOICE(incb a, incb a, incb a) +#define INS_L CHOICE(insl, insl, _LTOG ins) +#define INS_W CHOICE(insw, insw, _WTOG ins) +#define INS_B CHOICE(insb, insb, insb) +#define INT(a) CHOICE(int a, int a, int a) +#define INT3 CHOICE(int CONST(3), int3, int CONST(3)) +#define INTO CHOICE(into, into, into) +#define IRET CHOICE(iret, iret, iret) +#define IRETD CHOICE(iret, iret, iretd) +#define JA(a) CHOICE(ja a, ja a, ja a) +#define JAE(a) CHOICE(jae a, jae a, jae a) +#define JB(a) CHOICE(jb a, jb a, jb a) +#define JBE(a) CHOICE(jbe a, jbe a, jbe a) +#define JC(a) CHOICE(jc a, jc a, jc a) +#define JE(a) CHOICE(je a, je a, je a) +#define JG(a) CHOICE(jg a, jg a, jg a) +#define JGE(a) CHOICE(jge a, jge a, jge a) +#define JL(a) CHOICE(jl a, jl a, jl a) +#define JLE(a) CHOICE(jle a, jle a, jle a) +#define JNA(a) CHOICE(jna a, jna a, jna a) +#define JNAE(a) CHOICE(jnae a, jnae a, jnae a) +#define JNB(a) CHOICE(jnb a, jnb a, jnb a) +#define JNBE(a) CHOICE(jnbe a, jnbe a, jnbe a) +#define JNC(a) CHOICE(jnc a, jnc a, jnc a) +#define JNE(a) CHOICE(jne a, jne a, jne a) +#define JNG(a) CHOICE(jng a, jng a, jng a) +#define JNGE(a) CHOICE(jnge a, jnge a, jnge a) +#define JNL(a) CHOICE(jnl a, jnl a, jnl a) +#define JNLE(a) CHOICE(jnle a, jnle a, jnle a) +#define JNO(a) CHOICE(jno a, jno a, jno a) +#define JNP(a) CHOICE(jnp a, jnp a, jnp a) +#define JNS(a) CHOICE(jns a, jns a, jns a) +#define JNZ(a) CHOICE(jnz a, jnz a, jnz a) +#define JO(a) CHOICE(jo a, jo a, jo a) +#define JP(a) CHOICE(jp a, jp a, jp a) +#define JPE(a) CHOICE(jpe a, jpe a, jpe a) +#define JPO(a) CHOICE(jpo a, jpo a, jpo a) +#define JS(a) CHOICE(js a, js a, js a) +#define JZ(a) CHOICE(jz a, jz a, jz a) +#define JMP(a) CHOICE(jmp a, jmp a, jmp a) +#define JMPF(s,a) CHOICE(ljmp ARG2(s,a), ljmp ARG2(s,a), jmpf s:a) +#define LAHF CHOICE(lahf, lahf, lahf) +#if !defined(_REAL_MODE) && !defined(_V86_MODE) +#define LAR(a, b) CHOICE(lar ARG2(a, b), lar ARG2(a, b), lar ARG2(b, a)) +#endif +#define LEA_L(a, b) CHOICE(leal ARG2(a,b), leal ARG2(a,b), _LTOG lea ARG2(b,a)) +#define LEA_W(a, b) CHOICE(leaw ARG2(a,b), leaw ARG2(a,b), _WTOG lea ARG2(b,a)) +#define LEAVE CHOICE(leave, leave, leave) +#define LGDT(a) CHOICE(lgdt a, lgdt a, lgdt a) +#define LIDT(a) CHOICE(lidt a, lidt a, lidt a) +#define LDS(a, b) CHOICE(ldsl ARG2(a,b), lds ARG2(a,b), lds ARG2(b,a)) +#define LES(a, b) CHOICE(lesl ARG2(a,b), les ARG2(a,b), les ARG2(b,a)) +#define LFS(a, b) CHOICE(lfsl ARG2(a,b), lfs ARG2(a,b), lfs ARG2(b,a)) +#define LGS(a, b) CHOICE(lgsl ARG2(a,b), lgs ARG2(a,b), lgs ARG2(b,a)) +#define LSS(a, b) CHOICE(lssl ARG2(a,b), lss ARG2(a,b), lss ARG2(b,a)) +#define LLDT(a) CHOICE(lldt a, lldt a, lldt a) +#define LMSW(a) CHOICE(lmsw a, lmsw a, lmsw a) +#define LOCK CHOICE(lock, lock, lock) +#define LODS_L CHOICE(lodsl, lodsl, _LTOG lods) +#define LODS_W CHOICE(lodsw, lodsw, _WTOG lods) +#define LODS_B CHOICE(lodsb, lodsb, lodsb) +#define LOOP(a) CHOICE(loop a, loop a, loop a) +#define LOOPE(a) CHOICE(loope a, loope a, loope a) +#define LOOPZ(a) CHOICE(loopz a, loopz a, loopz a) +#define LOOPNE(a) CHOICE(loopne a, loopne a, loopne a) +#define LOOPNZ(a) CHOICE(loopnz a, loopnz a, loopnz a) +#if !defined(_REAL_MODE) && !defined(_V86_MODE) +#define LSL(a, b) CHOICE(lsl ARG2(a,b), lsl ARG2(a,b), lsl ARG2(b,a)) +#endif +#define LTR(a) CHOICE(ltr a, ltr a, ltr a) +#define MOV_SR(a, b) CHOICE(movw ARG2(a,b), mov ARG2(a,b), mov ARG2(b,a)) +#define MOV_L(a, b) CHOICE(movl ARG2(a,b), movl ARG2(a,b), _LTOG mov ARG2(b,a)) +#define MOV_W(a, b) CHOICE(movw ARG2(a,b), movw ARG2(a,b), _WTOG mov ARG2(b,a)) +#define MOV_B(a, b) CHOICE(movb ARG2(a,b), movb ARG2(a,b), movb ARG2(b,a)) +#define MOVS_L CHOICE(movsl, movsl, _LTOG movs) +#define MOVS_W CHOICE(movsw, movsw, _WTOG movs) +#define MOVS_B CHOICE(movsb, movsb, movsb) +#define MOVSX_BL(a, b) CHOICE(movsbl ARG2(a,b), movsbl ARG2(a,b), movsx ARG2(b,a)) +#define MOVSX_BW(a, b) CHOICE(movsbw ARG2(a,b), movsbw ARG2(a,b), movsx ARG2(b,a)) +#define MOVSX_WL(a, b) CHOICE(movswl ARG2(a,b), movswl ARG2(a,b), movsx ARG2(b,a)) +#define MOVZX_BL(a, b) CHOICE(movzbl ARG2(a,b), movzbl ARG2(a,b), movzx ARG2(b,a)) +#define MOVZX_BW(a, b) CHOICE(movzbw ARG2(a,b), movzbw ARG2(a,b), movzx ARG2(b,a)) +#define MOVZX_WL(a, b) CHOICE(movzwl ARG2(a,b), movzwl ARG2(a,b), movzx ARG2(b,a)) +#define MUL_L(a) CHOICE(mull a, mull a, _LTOG mul a) +#define MUL_W(a) CHOICE(mulw a, mulw a, _WTOG mul a) +#define MUL_B(a) CHOICE(mulb a, mulb a, mulb a) +#define NEG_L(a) CHOICE(negl a, negl a, _LTOG neg a) +#define NEG_W(a) CHOICE(negw a, negw a, _WTOG neg a) +#define NEG_B(a) CHOICE(negb a, negb a, negb a) +#define NOP CHOICE(nop, nop, nop) +#define NOT_L(a) CHOICE(notl a, notl a, _LTOG not a) +#define NOT_W(a) CHOICE(notw a, notw a, _WTOG not a) +#define NOT_B(a) CHOICE(notb a, notb a, notb a) +#define OR_L(a,b) CHOICE(orl ARG2(a,b), orl ARG2(a,b), _LTOG or ARG2(b,a)) +#define OR_W(a,b) CHOICE(orw ARG2(a,b), orw ARG2(a,b), _WTOG or ARG2(b,a)) +#define OR_B(a,b) CHOICE(orb ARG2(a,b), orb ARG2(a,b), orb ARG2(b,a)) +#define OUT_L CHOICE(outl (DX), outl ARG2(EAX,DX), _LTOG out DX) +#define OUT_W CHOICE(outw (DX), outw ARG2(AX,DX), _WTOG out DX) +#define OUT_B CHOICE(outb (DX), outb ARG2(AL,DX), outb DX) +/* Please AS code writer: use the following ONLY, if you refer to ports<256 + * directly, but not in OUT1_W(DX), for instance, even if OUT1_ looks nicer + */ +#define OUT1_L(a) CHOICE(outl (a), outl ARG2(EAX,a), _LTOG out a) +#define OUT1_W(a) CHOICE(outw (a), outw ARG2(AX,a), _WTOG out a) +#define OUT1_B(a) CHOICE(outb (a), outb ARG2(AL,a), outb a) +#define OUTS_L CHOICE(outsl, outsl, _LTOG outs) +#define OUTS_W CHOICE(outsw, outsw, _WTOG outs) +#define OUTS_B CHOICE(outsb, outsb, outsb) +#define POP_SR(a) CHOICE(pop a, pop a, pop a) +#define POP_L(a) CHOICE(popl a, popl a, _LTOG pop a) +#define POP_W(a) CHOICE(popw a, popw a, _WTOG pop a) +#define POPA_L CHOICE(popal, popal, _LTOG popa) +#define POPA_W CHOICE(popaw, popaw, _WTOG popa) +#define POPF_L CHOICE(popfl, popfl, _LTOG popf) +#define POPF_W CHOICE(popfw, popfw, _WTOG popf) +#define PUSH_SR(a) CHOICE(push a, push a, push a) +#define PUSH_L(a) CHOICE(pushl a, pushl a, _LTOG push a) +#define PUSH_W(a) CHOICE(pushw a, pushw a, _WTOG push a) +#define PUSH_B(a) CHOICE(push a, pushb a, push a) +#define PUSHA_L CHOICE(pushal, pushal, _LTOG pusha) +#define PUSHA_W CHOICE(pushaw, pushaw, _WTOG pusha) +#define PUSHF_L CHOICE(pushfl, pushfl, _LTOG pushf) +#define PUSHF_W CHOICE(pushfw, pushfw, _WTOG pushf) +#define RCL_L(a, b) CHOICE(rcll ARG2(a,b), rcll ARG2(a,b), _LTOG rcl ARG2(b,a)) +#define RCL_W(a, b) CHOICE(rclw ARG2(a,b), rclw ARG2(a,b), _WTOG rcl ARG2(b,a)) +#define RCL_B(a, b) CHOICE(rclb ARG2(a,b), rclb ARG2(a,b), rclb ARG2(b,a)) +#define RCR_L(a, b) CHOICE(rcrl ARG2(a,b), rcrl ARG2(a,b), _LTOG rcr ARG2(b,a)) +#define RCR_W(a, b) CHOICE(rcrw ARG2(a,b), rcrw ARG2(a,b), _WTOG rcr ARG2(b,a)) +#define RCR_B(a, b) CHOICE(rcrb ARG2(a,b), rcrb ARG2(a,b), rcrb ARG2(b,a)) +#define ROL_L(a, b) CHOICE(roll ARG2(a,b), roll ARG2(a,b), _LTOG rol ARG2(b,a)) +#define ROL_W(a, b) CHOICE(rolw ARG2(a,b), rolw ARG2(a,b), _WTOG rol ARG2(b,a)) +#define ROL_B(a, b) CHOICE(rolb ARG2(a,b), rolb ARG2(a,b), rolb ARG2(b,a)) +#define ROR_L(a, b) CHOICE(rorl ARG2(a,b), rorl ARG2(a,b), _LTOG ror ARG2(b,a)) +#define ROR_W(a, b) CHOICE(rorw ARG2(a,b), rorw ARG2(a,b), _WTOG ror ARG2(b,a)) +#define ROR_B(a, b) CHOICE(rorb ARG2(a,b), rorb ARG2(a,b), rorb ARG2(b,a)) +#define REP CHOICE(rep ;, rep ;, repe) +#define REPE CHOICE(repz ;, repe ;, repe) +#define REPNE CHOICE(repnz ;, repne ;, repne) +#define REPNZ REPNE +#define REPZ REPE +#define RET CHOICE(ret, ret, ret) +#define SAHF CHOICE(sahf, sahf, sahf) +#define SAL_L(a, b) CHOICE(sall ARG2(a,b), sall ARG2(a,b), _LTOG sal ARG2(b,a)) +#define SAL_W(a, b) CHOICE(salw ARG2(a,b), salw ARG2(a,b), _WTOG sal ARG2(b,a)) +#define SAL_B(a, b) CHOICE(salb ARG2(a,b), salb ARG2(a,b), salb ARG2(b,a)) +#define SAR_L(a, b) CHOICE(sarl ARG2(a,b), sarl ARG2(a,b), _LTOG sar ARG2(b,a)) +#define SAR_W(a, b) CHOICE(sarw ARG2(a,b), sarw ARG2(a,b), _WTOG sar ARG2(b,a)) +#define SAR_B(a, b) CHOICE(sarb ARG2(a,b), sarb ARG2(a,b), sarb ARG2(b,a)) +#define SBB_L(a, b) CHOICE(sbbl ARG2(a,b), sbbl ARG2(a,b), _LTOG sbb ARG2(b,a)) +#define SBB_W(a, b) CHOICE(sbbw ARG2(a,b), sbbw ARG2(a,b), _WTOG sbb ARG2(b,a)) +#define SBB_B(a, b) CHOICE(sbbb ARG2(a,b), sbbb ARG2(a,b), sbbb ARG2(b,a)) +#define SCAS_L CHOICE(scasl, scasl, _LTOG scas) +#define SCAS_W CHOICE(scasw, scasw, _WTOG scas) +#define SCAS_B CHOICE(scasb, scasb, scasb) +#define SETA(a) CHOICE(seta a, seta a, seta a) +#define SETAE(a) CHOICE(setae a, setae a, setae a) +#define SETB(a) CHOICE(setb a, setb a, setb a) +#define SETBE(a) CHOICE(setbe a, setbe a, setbe a) +#define SETC(a) CHOICE(setc a, setb a, setb a) +#define SETE(a) CHOICE(sete a, sete a, sete a) +#define SETG(a) CHOICE(setg a, setg a, setg a) +#define SETGE(a) CHOICE(setge a, setge a, setge a) +#define SETL(a) CHOICE(setl a, setl a, setl a) +#define SETLE(a) CHOICE(setle a, setle a, setle a) +#define SETNA(a) CHOICE(setna a, setna a, setna a) +#define SETNAE(a) CHOICE(setnae a, setnae a, setnae a) +#define SETNB(a) CHOICE(setnb a, setnb a, setnb a) +#define SETNBE(a) CHOICE(setnbe a, setnbe a, setnbe a) +#define SETNC(a) CHOICE(setnc a, setnb a, setnb a) +#define SETNE(a) CHOICE(setne a, setne a, setne a) +#define SETNG(a) CHOICE(setng a, setng a, setng a) +#define SETNGE(a) CHOICE(setnge a, setnge a, setnge a) +#define SETNL(a) CHOICE(setnl a, setnl a, setnl a) +#define SETNLE(a) CHOICE(setnle a, setnle a, setnle a) +#define SETNO(a) CHOICE(setno a, setno a, setno a) +#define SETNP(a) CHOICE(setnp a, setnp a, setnp a) +#define SETNS(a) CHOICE(setns a, setns a, setna a) +#define SETNZ(a) CHOICE(setnz a, setnz a, setnz a) +#define SETO(a) CHOICE(seto a, seto a, seto a) +#define SETP(a) CHOICE(setp a, setp a, setp a) +#define SETPE(a) CHOICE(setpe a, setpe a, setpe a) +#define SETPO(a) CHOICE(setpo a, setpo a, setpo a) +#define SETS(a) CHOICE(sets a, sets a, seta a) +#define SETZ(a) CHOICE(setz a, setz a, setz a) +#define SGDT(a) CHOICE(sgdt a, sgdt a, sgdt a) +#define SIDT(a) CHOICE(sidt a, sidt a, sidt a) +#define SHL_L(a, b) CHOICE(shll ARG2(a,b), shll ARG2(a,b), _LTOG shl ARG2(b,a)) +#define SHL_W(a, b) CHOICE(shlw ARG2(a,b), shlw ARG2(a,b), _WTOG shl ARG2(b,a)) +#define SHL_B(a, b) CHOICE(shlb ARG2(a,b), shlb ARG2(a,b), shlb ARG2(b,a)) +#define SHLD_L(a,b,c) CHOICE(shldl ARG3(a,b,c), shldl ARG3(a,b,c), _LTOG shld ARG3(c,b,a)) +#define SHLD2_L(a,b) CHOICE(shldl ARG2(a,b), shldl ARG3(CL,a,b), _LTOG shld ARG3(b,a,CL)) +#define SHLD_W(a,b,c) CHOICE(shldw ARG3(a,b,c), shldw ARG3(a,b,c), _WTOG shld ARG3(c,b,a)) +#define SHLD2_W(a,b) CHOICE(shldw ARG2(a,b), shldw ARG3(CL,a,b), _WTOG shld ARG3(b,a,CL)) +#define SHR_L(a, b) CHOICE(shrl ARG2(a,b), shrl ARG2(a,b), _LTOG shr ARG2(b,a)) +#define SHR_W(a, b) CHOICE(shrw ARG2(a,b), shrw ARG2(a,b), _WTOG shr ARG2(b,a)) +#define SHR_B(a, b) CHOICE(shrb ARG2(a,b), shrb ARG2(a,b), shrb ARG2(b,a)) +#define SHRD_L(a,b,c) CHOICE(shrdl ARG3(a,b,c), shrdl ARG3(a,b,c), _LTOG shrd ARG3(c,b,a)) +#define SHRD2_L(a,b) CHOICE(shrdl ARG2(a,b), shrdl ARG3(CL,a,b), _LTOG shrd ARG3(b,a,CL)) +#define SHRD_W(a,b,c) CHOICE(shrdw ARG3(a,b,c), shrdw ARG3(a,b,c), _WTOG shrd ARG3(c,b,a)) +#define SHRD2_W(a,b) CHOICE(shrdw ARG2(a,b), shrdw ARG3(CL,a,b), _WTOG shrd ARG3(b,a,CL)) +#define SLDT(a) CHOICE(sldt a, sldt a, sldt a) +#define SMSW(a) CHOICE(smsw a, smsw a, smsw a) +#define STC CHOICE(stc, stc, stc) +#define STD CHOICE(std, std, std) +#define STI CHOICE(sti, sti, sti) +#define STOS_L CHOICE(stosl, stosl, _LTOG stos) +#define STOS_W CHOICE(stosw, stosw, _WTOG stos) +#define STOS_B CHOICE(stosb, stosb, stosb) +#define STR(a) CHOICE(str a, str a, str a) +#define SUB_L(a, b) CHOICE(subl ARG2(a,b), subl ARG2(a,b), _LTOG sub ARG2(b,a)) +#define SUB_W(a, b) CHOICE(subw ARG2(a,b), subw ARG2(a,b), _WTOG sub ARG2(b,a)) +#define SUB_B(a, b) CHOICE(subb ARG2(a,b), subb ARG2(a,b), subb ARG2(b,a)) +#define TEST_L(a, b) CHOICE(testl ARG2(a,b), testl ARG2(a,b), _LTOG test ARG2(b,a)) +#define TEST_W(a, b) CHOICE(testw ARG2(a,b), testw ARG2(a,b), _WTOG test ARG2(b,a)) +#define TEST_B(a, b) CHOICE(testb ARG2(a,b), testb ARG2(a,b), testb ARG2(b,a)) +#define VERR(a) CHOICE(verr a, verr a, verr a) +#define VERW(a) CHOICE(verw a, verw a, verw a) +#define WAIT CHOICE(wait, wait, wait) +#define XCHG_L(a, b) CHOICE(xchgl ARG2(a,b), xchgl ARG2(a,b), _LTOG xchg ARG2(b,a)) +#define XCHG_W(a, b) CHOICE(xchgw ARG2(a,b), xchgw ARG2(a,b), _WTOG xchg ARG2(b,a)) +#define XCHG_B(a, b) CHOICE(xchgb ARG2(a,b), xchgb ARG2(a,b), xchgb ARG2(b,a)) +#define XLAT CHOICE(xlat, xlat, xlat) +#define XOR_L(a, b) CHOICE(xorl ARG2(a,b), xorl ARG2(a,b), _LTOG xor ARG2(b,a)) +#define XOR_W(a, b) CHOICE(xorw ARG2(a,b), xorw ARG2(a,b), _WTOG xor ARG2(b,a)) +#define XOR_B(a, b) CHOICE(xorb ARG2(a,b), xorb ARG2(a,b), xorb ARG2(b,a)) + + +/* Floating Point Instructions */ +#define F2XM1 CHOICE(f2xm1, f2xm1, f2xm1) +#define FABS CHOICE(fabs, fabs, fabs) +#define FADD_D(a) CHOICE(faddl a, faddl a, faddd a) +#define FADD_S(a) CHOICE(fadds a, fadds a, fadds a) +#define FADD2(a, b) CHOICE(fadd ARG2(a,b), fadd ARG2(a,b), fadd ARG2(b,a)) +#define FADDP(a, b) CHOICE(faddp ARG2(a,b), faddp ARG2(a,b), faddp ARG2(b,a)) +#define FIADD_L(a) CHOICE(fiaddl a, fiaddl a, fiaddl a) +#define FIADD_W(a) CHOICE(fiadd a, fiadds a, fiadds a) +#define FBLD(a) CHOICE(fbld a, fbld a, fbld a) +#define FBSTP(a) CHOICE(fbstp a, fbstp a, fbstp a) +#define FCHS CHOICE(fchs, fchs, fchs) +#define FCLEX CHOICE(fclex, wait; fnclex, wait; fclex) +#define FNCLEX CHOICE(fnclex, fnclex, fclex) +#define FCOM(a) CHOICE(fcom a, fcom a, fcom a) +#define FCOM_D(a) CHOICE(fcoml a, fcoml a, fcomd a) +#define FCOM_S(a) CHOICE(fcoms a, fcoms a, fcoms a) +#define FCOMP(a) CHOICE(fcomp a, fcomp a, fcomp a) +#define FCOMP_D(a) CHOICE(fcompl a, fcompl a, fcompd a) +#define FCOMP_S(a) CHOICE(fcomps a, fcomps a, fcomps a) +#define FCOMPP CHOICE(fcompp, fcompp, fcompp) +#define FCOS CHOICE(fcos, fcos, fcos) +#define FDECSTP CHOICE(fdecstp, fdecstp, fdecstp) +#define FDIV_D(a) CHOICE(fdivl a, fdivl a, fdivd a) +#define FDIV_S(a) CHOICE(fdivs a, fdivs a, fdivs a) +#define FDIV2(a, b) CHOICE(fdiv ARG2(a,b), fdiv ARG2(a,b), fdiv ARG2(b,a)) +#define FDIVP(a, b) CHOICE(fdivp ARG2(a,b), fdivp ARG2(a,b), fdivp ARG2(b,a)) +#define FIDIV_L(a) CHOICE(fidivl a, fidivl a, fidivl a) +#define FIDIV_W(a) CHOICE(fidiv a, fidivs a, fidivs a) +#define FDIVR_D(a) CHOICE(fdivrl a, fdivrl a, fdivrd a) +#define FDIVR_S(a) CHOICE(fdivrs a, fdivrs a, fdivrs a) +#define FDIVR2(a, b) CHOICE(fdivr ARG2(a,b), fdivr ARG2(a,b), fdivr ARG2(b,a)) +#define FDIVRP(a, b) CHOICE(fdivrp ARG2(a,b), fdivrp ARG2(a,b), fdivrp ARG2(b,a)) +#define FIDIVR_L(a) CHOICE(fidivrl a, fidivrl a, fidivrl a) +#define FIDIVR_W(a) CHOICE(fidivr a, fidivrs a, fidivrs a) +#define FFREE(a) CHOICE(ffree a, ffree a, ffree a) +#define FICOM_L(a) CHOICE(ficoml a, ficoml a, ficoml a) +#define FICOM_W(a) CHOICE(ficom a, ficoms a, ficoms a) +#define FICOMP_L(a) CHOICE(ficompl a, ficompl a, ficompl a) +#define FICOMP_W(a) CHOICE(ficomp a, ficomps a, ficomps a) +#define FILD_Q(a) CHOICE(fildll a, fildq a, fildq a) +#define FILD_L(a) CHOICE(fildl a, fildl a, fildl a) +#define FILD_W(a) CHOICE(fild a, filds a, filds a) +#define FINCSTP CHOICE(fincstp, fincstp, fincstp) +#define FINIT CHOICE(finit, wait; fninit, wait; finit) +#define FNINIT CHOICE(fninit, fninit, finit) +#define FIST_L(a) CHOICE(fistl a, fistl a, fistl a) +#define FIST_W(a) CHOICE(fist a, fists a, fists a) +#define FISTP_Q(a) CHOICE(fistpll a, fistpq a, fistpq a) +#define FISTP_L(a) CHOICE(fistpl a, fistpl a, fistpl a) +#define FISTP_W(a) CHOICE(fistp a, fistps a, fistps a) +#define FLD_X(a) CHOICE(fldt a, fldt a, fldx a) /* 80 bit data type! */ +#define FLD_D(a) CHOICE(fldl a, fldl a, fldd a) +#define FLD_S(a) CHOICE(flds a, flds a, flds a) +#define FLD1 CHOICE(fld1, fld1, fld1) +#define FLDL2T CHOICE(fldl2t, fldl2t, fldl2t) +#define FLDL2E CHOICE(fldl2e, fldl2e, fldl2e) +#define FLDPI CHOICE(fldpi, fldpi, fldpi) +#define FLDLG2 CHOICE(fldlg2, fldlg2, fldlg2) +#define FLDLN2 CHOICE(fldln2, fldln2, fldln2) +#define FLDZ CHOICE(fldz, fldz, fldz) +#define FLDCW(a) CHOICE(fldcw a, fldcw a, fldcw a) +#define FLDENV(a) CHOICE(fldenv a, fldenv a, fldenv a) +#define FMUL_S(a) CHOICE(fmuls a, fmuls a, fmuls a) +#define FMUL_D(a) CHOICE(fmull a, fmull a, fmuld a) +#define FMUL2(a, b) CHOICE(fmul ARG2(a,b), fmul ARG2(a,b), fmul ARG2(b,a)) +#define FMULP(a, b) CHOICE(fmulp ARG2(a,b), fmulp ARG2(a,b), fmulp ARG2(b,a)) +#define FIMUL_L(a) CHOICE(fimull a, fimull a, fimull a) +#define FIMUL_W(a) CHOICE(fimul a, fimuls a, fimuls a) +#define FNOP CHOICE(fnop, fnop, fnop) +#define FPATAN CHOICE(fpatan, fpatan, fpatan) +#define FPREM CHOICE(fprem, fprem, fprem) +#define FPREM1 CHOICE(fprem1, fprem1, fprem1) +#define FPTAN CHOICE(fptan, fptan, fptan) +#define FRNDINT CHOICE(frndint, frndint, frndint) +#define FRSTOR(a) CHOICE(frstor a, frstor a, frstor a) +#define FSAVE(a) CHOICE(fsave a, wait; fnsave a, wait; fsave a) +#define FNSAVE(a) CHOICE(fnsave a, fnsave a, fsave a) +#define FSCALE CHOICE(fscale, fscale, fscale) +#define FSIN CHOICE(fsin, fsin, fsin) +#define FSINCOS CHOICE(fsincos, fsincos, fsincos) +#define FSQRT CHOICE(fsqrt, fsqrt, fsqrt) +#define FST_D(a) CHOICE(fstl a, fstl a, fstd a) +#define FST_S(a) CHOICE(fsts a, fsts a, fsts a) +#define FSTP_X(a) CHOICE(fstpt a, fstpt a, fstpx a) +#define FSTP_D(a) CHOICE(fstpl a, fstpl a, fstpd a) +#define FSTP_S(a) CHOICE(fstps a, fstps a, fstps a) +#define FSTP(a) CHOICE(fstp a, fstp a, fstp a) +#define FSTCW(a) CHOICE(fstcw a, wait; fnstcw a, wait; fstcw a) +#define FNSTCW(a) CHOICE(fnstcw a, fnstcw a, fstcw a) +#define FSTENV(a) CHOICE(fstenv a, wait; fnstenv a, fstenv a) +#define FNSTENV(a) CHOICE(fnstenv a, fnstenv a, fstenv a) +#define FSTSW(a) CHOICE(fstsw a, wait; fnstsw a, wait; fstsw a) +#define FNSTSW(a) CHOICE(fnstsw a, fnstsw a, fstsw a) +#define FSUB_S(a) CHOICE(fsubs a, fsubs a, fsubs a) +#define FSUB_D(a) CHOICE(fsubl a, fsubl a, fsubd a) +#define FSUB2(a, b) CHOICE(fsub ARG2(a,b), fsub ARG2(a,b), fsub ARG2(b,a)) +#define FSUBP(a, b) CHOICE(fsubp ARG2(a,b), fsubp ARG2(a,b), fsubp ARG2(b,a)) +#define FISUB_L(a) CHOICE(fisubl a, fisubl a, fisubl a) +#define FISUB_W(a) CHOICE(fisub a, fisubs a, fisubs a) +#define FSUBR_S(a) CHOICE(fsubrs a, fsubrs a, fsubrs a) +#define FSUBR_D(a) CHOICE(fsubrl a, fsubrl a, fsubrd a) +#define FSUBR2(a, b) CHOICE(fsubr ARG2(a,b), fsubr ARG2(a,b), fsubr ARG2(b,a)) +#define FSUBRP(a, b) CHOICE(fsubrp ARG2(a,b), fsubrp ARG2(a,b), fsubrp ARG2(b,a)) +#define FISUBR_L(a) CHOICE(fisubrl a, fisubrl a, fisubrl a) +#define FISUBR_W(a) CHOICE(fisubr a, fisubrs a, fisubrs a) +#define FTST CHOICE(ftst, ftst, ftst) +#define FUCOM(a) CHOICE(fucom a, fucom a, fucom a) +#define FUCOMP(a) CHOICE(fucomp a, fucomp a, fucomp a) +#define FUCOMPP CHOICE(fucompp, fucompp, fucompp) +#define FWAIT CHOICE(wait, wait, wait) +#define FXAM CHOICE(fxam, fxam, fxam) +#define FXCH(a) CHOICE(fxch a, fxch a, fxch a) +#define FXTRACT CHOICE(fxtract, fxtract, fxtract) +#define FYL2X CHOICE(fyl2x, fyl2x, fyl2x) +#define FYL2XP1 CHOICE(fyl2xp1, fyl2xp1, fyl2xp1) + +/* New instructions */ +#define CPUID CHOICE(D_BYTE ARG2(15, 162), cpuid, D_BYTE ARG2(15, 162)) +#define RDTSC CHOICE(D_BYTE ARG2(15, 49), rdtsc, D_BYTE ARG2(15, 49)) + +#else /* NASM_ASSEMBLER || MASM_ASSEMBLER is defined */ + + /****************************************/ + /* */ + /* Intel style assemblers. */ + /* (NASM and MASM) */ + /* */ + /****************************************/ + +#define P_EAX EAX +#define L_EAX EAX +#define W_AX AX +#define B_AH AH +#define B_AL AL + +#define P_EBX EBX +#define L_EBX EBX +#define W_BX BX +#define B_BH BH +#define B_BL BL + +#define P_ECX ECX +#define L_ECX ECX +#define W_CX CX +#define B_CH CH +#define B_CL CL + +#define P_EDX EDX +#define L_EDX EDX +#define W_DX DX +#define B_DH DH +#define B_DL DL + +#define P_EBP EBP +#define L_EBP EBP +#define W_BP BP + +#define P_ESI ESI +#define L_ESI ESI +#define W_SI SI + +#define P_EDI EDI +#define L_EDI EDI +#define W_DI DI + +#define P_ESP ESP +#define L_ESP ESP +#define W_SP SP + +#define W_CS CS +#define W_SS SS +#define W_DS DS +#define W_ES ES +#define W_FS FS +#define W_GS GS + +#define X_ST ST +#define D_ST ST +#define L_ST ST + +#define P_MM0 mm0 +#define P_MM1 mm1 +#define P_MM2 mm2 +#define P_MM3 mm3 +#define P_MM4 mm4 +#define P_MM5 mm5 +#define P_MM6 mm6 +#define P_MM7 mm7 + +#define P_XMM0 xmm0 +#define P_XMM1 xmm1 +#define P_XMM2 xmm2 +#define P_XMM3 xmm3 +#define P_XMM4 xmm4 +#define P_XMM5 xmm5 +#define P_XMM6 xmm6 +#define P_XMM7 xmm7 + +#define CONCAT(x, y) x ## y +#define CONCAT3(x, y, z) x ## y ## z + +#if defined(NASM_ASSEMBLER) + +#define ST(n) st ## n +#define ST0 st0 + +#define TBYTE_PTR tword +#define QWORD_PTR qword +#define DWORD_PTR dword +#define WORD_PTR word +#define BYTE_PTR byte + +#define OFFSET + +#define GLOBL GLOBAL +#define ALIGNTEXT32 ALIGN 32 +#define ALIGNTEXT16 ALIGN 16 +#define ALIGNTEXT8 ALIGN 8 +#define ALIGNTEXT4 ALIGN 4 +#define ALIGNTEXT2 ALIGN 2 +#define ALIGNTEXT32ifNOP ALIGN 32 +#define ALIGNTEXT16ifNOP ALIGN 16 +#define ALIGNTEXT8ifNOP ALIGN 8 +#define ALIGNTEXT4ifNOP ALIGN 4 +#define ALIGNDATA32 ALIGN 32 +#define ALIGNDATA16 ALIGN 16 +#define ALIGNDATA8 ALIGN 8 +#define ALIGNDATA4 ALIGN 4 +#define ALIGNDATA2 ALIGN 2 +#define FILE(s) +#define STRING(s) db s +#define D_LONG dd +#define D_WORD dw +#define D_BYTE db +/* #define SPACE */ +/* #define COMM */ +#if defined(__WATCOMC__) +SECTION _TEXT public align=16 class=CODE use32 flat +SECTION _DATA public align=16 class=DATA use32 flat +#define SEG_TEXT SECTION _TEXT +#define SEG_DATA SECTION _DATA +#define SEG_BSS SECTION .bss +#else +#define SEG_DATA SECTION .data +#define SEG_TEXT SECTION .text +#define SEG_BSS SECTION .bss +#endif + +#define D_SPACE(n) db n REP 0 + +#define AS_BEGIN + +/* Jcc's should be handled better than this... */ +#define NEAR near + +#else /* MASM */ + +#define TBYTE_PTR tbyte ptr +#define QWORD_PTR qword ptr +#define DWORD_PTR dword ptr +#define WORD_PTR word ptr +#define BYTE_PTR byte ptr + +#define OFFSET offset + +#define GLOBL GLOBAL +#define ALIGNTEXT32 ALIGN 32 +#define ALIGNTEXT16 ALIGN 16 +#define ALIGNTEXT8 ALIGN 8 +#define ALIGNTEXT4 ALIGN 4 +#define ALIGNTEXT2 ALIGN 2 +#define ALIGNTEXT32ifNOP ALIGN 32 +#define ALIGNTEXT16ifNOP ALIGN 16 +#define ALIGNTEXT8ifNOP ALIGN 8 +#define ALIGNTEXT4ifNOP ALIGN 4 +#define ALIGNDATA32 ALIGN 32 +#define ALIGNDATA16 ALIGN 16 +#define ALIGNDATA8 ALIGN 8 +#define ALIGNDATA4 ALIGN 4 +#define ALIGNDATA2 ALIGN 2 +#define FILE(s) +#define STRING(s) db s +#define D_LONG dd +#define D_WORD dw +#define D_BYTE db +/* #define SPACE */ +/* #define COMM */ +#define SEG_DATA .DATA +#define SEG_TEXT .CODE +#define SEG_BSS .DATA + +#define D_SPACE(n) db n REP 0 + +#define AS_BEGIN + +#define NEAR + +#endif + +#if defined(Lynx) || (defined(SYSV) || defined(SVR4)) \ + || (defined(__linux__) || defined(__OS2ELF__)) && defined(__ELF__) \ + || (defined(__FreeBSD__) && __FreeBSD__ >= 3) \ + || (defined(__NetBSD__) && defined(__ELF__)) +#define GLNAME(a) a +#else +#define GLNAME(a) CONCAT(_, a) +#endif + +/* + * Addressing Modes + */ + +/* Immediate Mode */ +#define P_ADDR(a) OFFSET a +#define X_ADDR(a) OFFSET a +#define D_ADDR(a) OFFSET a +#define L_ADDR(a) OFFSET a +#define W_ADDR(a) OFFSET a +#define B_ADDR(a) OFFSET a + +#define P_CONST(a) a +#define X_CONST(a) a +#define D_CONST(a) a +#define L_CONST(a) a +#define W_CONST(a) a +#define B_CONST(a) a + +/* Indirect Mode */ +#ifdef NASM_ASSEMBLER +#define P_CONTENT(a) [a] +#define X_CONTENT(a) TBYTE_PTR [a] +#define D_CONTENT(a) QWORD_PTR [a] +#define L_CONTENT(a) DWORD_PTR [a] +#define W_CONTENT(a) WORD_PTR [a] +#define B_CONTENT(a) BYTE_PTR [a] +#else +#define P_CONTENT(a) a +#define X_CONTENT(a) TBYTE_PTR a +#define D_CONTENT(a) QWORD_PTR a +#define L_CONTENT(a) DWORD_PTR a +#define W_CONTENT(a) WORD_PTR a +#define B_CONTENT(a) BYTE_PTR a +#endif + +/* Register a indirect */ +#define P_REGIND(a) [a] +#define X_REGIND(a) TBYTE_PTR [a] +#define D_REGIND(a) QWORD_PTR [a] +#define L_REGIND(a) DWORD_PTR [a] +#define W_REGIND(a) WORD_PTR [a] +#define B_REGIND(a) BYTE_PTR [a] + +/* Register b indirect plus displacement a */ +#define P_REGOFF(a, b) [b + a] +#define X_REGOFF(a, b) TBYTE_PTR [b + a] +#define D_REGOFF(a, b) QWORD_PTR [b + a] +#define L_REGOFF(a, b) DWORD_PTR [b + a] +#define W_REGOFF(a, b) WORD_PTR [b + a] +#define B_REGOFF(a, b) BYTE_PTR [b + a] + +/* Reg indirect Base + Index + Displacement - this is mainly for 16-bit mode + * which has no scaling + */ +#define P_REGBID(b, i, d) [b + i + d] +#define X_REGBID(b, i, d) TBYTE_PTR [b + i + d] +#define D_REGBID(b, i, d) QWORD_PTR [b + i + d] +#define L_REGBID(b, i, d) DWORD_PTR [b + i + d] +#define W_REGBID(b, i, d) WORD_PTR [b + i + d] +#define B_REGBID(b, i, d) BYTE_PTR [b + i + d] + +/* Reg indirect Base + (Index * Scale) */ +#define P_REGBIS(b, i, s) [b + i * s] +#define X_REGBIS(b, i, s) TBYTE_PTR [b + i * s] +#define D_REGBIS(b, i, s) QWORD_PTR [b + i * s] +#define L_REGBIS(b, i, s) DWORD_PTR [b + i * s] +#define W_REGBIS(b, i, s) WORD_PTR [b + i * s] +#define B_REGBIS(b, i, s) BYTE_PTR [b + i * s] + +/* Reg indirect Base + (Index * Scale) + Displacement */ +#define P_REGBISD(b, i, s, d) [b + i * s + d] +#define X_REGBISD(b, i, s, d) TBYTE_PTR [b + i * s + d] +#define D_REGBISD(b, i, s, d) QWORD_PTR [b + i * s + d] +#define L_REGBISD(b, i, s, d) DWORD_PTR [b + i * s + d] +#define W_REGBISD(b, i, s, d) WORD_PTR [b + i * s + d] +#define B_REGBISD(b, i, s, d) BYTE_PTR [b + i * s + d] + +/* Displaced Scaled Index: */ +#define P_REGDIS(d, i, s) [i * s + d] +#define X_REGDIS(d, i, s) TBYTE_PTR [i * s + d] +#define D_REGDIS(d, i, s) QWORD_PTR [i * s + d] +#define L_REGDIS(d, i, s) DWORD_PTR [i * s + d] +#define W_REGDIS(d, i, s) WORD_PTR [i * s + d] +#define B_REGDIS(d, i, s) BYTE_PTR [i * s + d] + +/* Indexed Base: */ +#define P_REGBI(b, i) [b + i] +#define X_REGBI(b, i) TBYTE_PTR [b + i] +#define D_REGBI(b, i) QWORD_PTR [b + i] +#define L_REGBI(b, i) DWORD_PTR [b + i] +#define W_REGBI(b, i) WORD_PTR [b + i] +#define B_REGBI(b, i) BYTE_PTR [b + i] + +/* Displaced Base: */ +#define P_REGDB(d, b) [b + d] +#define X_REGDB(d, b) TBYTE_PTR [b + d] +#define D_REGDB(d, b) QWORD_PTR [b + d] +#define L_REGDB(d, b) DWORD_PTR [b + d] +#define W_REGDB(d, b) WORD_PTR [b + d] +#define B_REGDB(d, b) BYTE_PTR [b + d] + +/* Variable indirect: */ +#define VARINDIRECT(var) [var] + +/* Use register contents as jump/call target: */ +#define CODEPTR(reg) P_(reg) + +/* + * Redefine assembler commands + */ + +#define P_(a) P_ ## a +#define X_(a) X_ ## a +#define D_(a) D_ ## a +#define SR_(a) W_ ## a +#define S_(a) L_ ## a +#define L_(a) L_ ## a +#define W_(a) W_ ## a +#define B_(a) B_ ## a + +#define AAA aaa +#define AAD aad +#define AAM aam +#define AAS aas +#define ADC_L(a, b) adc L_(b), L_(a) +#define ADC_W(a, b) adc W_(b), W_(a) +#define ADC_B(a, b) adc B_(b), B_(a) +#define ADD_L(a, b) add L_(b), L_(a) +#define ADD_W(a, b) add W_(b), W_(a) +#define ADD_B(a, b) add B_(b), B_(a) +#define AND_L(a, b) and L_(b), L_(a) +#define AND_W(a, b) and W_(b), W_(a) +#define AND_B(a, b) and B_(b), B_(a) +#define ARPL(a,b) arpl W_(b), a +#define BOUND_L(a, b) bound L_(b), L_(a) +#define BOUND_W(a, b) bound W_(b), W_(a) +#define BSF_L(a, b) bsf L_(b), L_(a) +#define BSF_W(a, b) bsf W_(b), W_(a) +#define BSR_L(a, b) bsr L_(b), L_(a) +#define BSR_W(a, b) bsr W_(b), W_(a) +#define BT_L(a, b) bt L_(b), L_(a) +#define BT_W(a, b) bt W_(b), W_(a) +#define BTC_L(a, b) btc L_(b), L_(a) +#define BTC_W(a, b) btc W_(b), W_(a) +#define BTR_L(a, b) btr L_(b), L_(a) +#define BTR_W(a, b) btr W_(b), W_(a) +#define BTS_L(a, b) bts L_(b), L_(a) +#define BTS_W(a, b) bts W_(b), W_(a) +#define CALL(a) call a +#define CALLF(s,a) call far s:a +#define CBW cbw +#define CWDE cwde +#define CLC clc +#define CLD cld +#define CLI cli +#define CLTS clts +#define CMC cmc +#define CMP_L(a, b) cmp L_(b), L_(a) +#define CMP_W(a, b) cmp W_(b), W_(a) +#define CMP_B(a, b) cmp B_(b), B_(a) +#define CMPS_L cmpsd +#define CMPS_W cmpsw +#define CMPS_B cmpsb +#define CPUID cpuid +#define CWD cwd +#define CDQ cdq +#define DAA daa +#define DAS das +#define DEC_L(a) dec L_(a) +#define DEC_W(a) dec W_(a) +#define DEC_B(a) dec B_(a) +#define DIV_L(a) div L_(a) +#define DIV_W(a) div W_(a) +#define DIV_B(a) div B_(a) +#define ENTER(a,b) enter b, a +#define HLT hlt +#define IDIV_L(a) idiv L_(a) +#define IDIV_W(a) idiv W_(a) +#define IDIV_B(a) idiv B_(a) +#define IMUL_L(a, b) imul L_(b), L_(a) +#define IMUL_W(a, b) imul W_(b), W_(a) +#define IMUL_B(a) imul B_(a) +#define IN_L in EAX, DX +#define IN_W in AX, DX +#define IN_B in AL, DX +#define IN1_L(a) in1 L_(a) +#define IN1_W(a) in1 W_(a) +#define IN1_B(a) in1 B_(a) +#define INC_L(a) inc L_(a) +#define INC_W(a) inc W_(a) +#define INC_B(a) inc B_(a) +#define INS_L ins +#define INS_W ins +#define INS_B ins +#define INT(a) int B_(a) +#define INT3 int3 +#define INTO into +#define IRET iret +#define IRETD iretd +#define JA(a) ja NEAR a +#define JAE(a) jae NEAR a +#define JB(a) jb NEAR a +#define JBE(a) jbe NEAR a +#define JC(a) jc NEAR a +#define JE(a) je NEAR a +#define JG(a) jg NEAR a +#define JGE(a) jge NEAR a +#define JL(a) jl NEAR a +#define JLE(a) jle NEAR a +#define JNA(a) jna NEAR a +#define JNAE(a) jnae NEAR a +#define JNB(a) jnb NEAR a +#define JNBE(a) jnbe NEAR a +#define JNC(a) jnc NEAR a +#define JNE(a) jne NEAR a +#define JNG(a) jng NEAR a +#define JNGE(a) jnge NEAR a +#define JNL(a) jnl NEAR a +#define JNLE(a) jnle NEAR a +#define JNO(a) jno NEAR a +#define JNP(a) jnp NEAR a +#define JNS(a) jns NEAR a +#define JNZ(a) jnz NEAR a +#define JO(a) jo NEAR a +#define JP(a) jp NEAR a +#define JPE(a) jpe NEAR a +#define JPO(a) jpo NEAR a +#define JS(a) js NEAR a +#define JZ(a) jz NEAR a +#define JMP(a) jmp a +#define JMPF(s,a) jmp far s:a +#define LAHF lahf +#define LAR(a, b) lar b, a +#define LEA_L(a, b) lea P_(b), P_(a) +#define LEA_W(a, b) lea P_(b), P_(a) +#define LEAVE leave +#define LGDT(a) lgdt a +#define LIDT(a) lidt a +#define LDS(a, b) lds b, P_(a) +#define LES(a, b) les b, P_(a) +#define LFS(a, b) lfs b, P_(a) +#define LGS(a, b) lgs b, P_(a) +#define LSS(a, b) lss b, P_(a) +#define LLDT(a) lldt a +#define LMSW(a) lmsw a +#define LOCK lock +#define LODS_L lodsd +#define LODS_W lodsw +#define LODS_B lodsb +#define LOOP(a) loop a +#define LOOPE(a) loope a +#define LOOPZ(a) loopz a +#define LOOPNE(a) loopne a +#define LOOPNZ(a) loopnz a +#define LSL(a, b) lsl b, a +#define LTR(a) ltr a +#define MOV_SR(a, b) mov SR_(b), SR_(a) +#define MOV_L(a, b) mov L_(b), L_(a) +#define MOV_W(a, b) mov W_(b), W_(a) +#define MOV_B(a, b) mov B_(b), B_(a) +#define MOVS_L movsd +#define MOVS_W movsw +#define MOVS_B movsb +#define MOVSX_BL(a, b) movsx B_(b), B_(a) +#define MOVSX_BW(a, b) movsx B_(b), B_(a) +#define MOVSX_WL(a, b) movsx W_(b), W_(a) +#define MOVZX_BL(a, b) movzx B_(b), B_(a) +#define MOVZX_BW(a, b) movzx B_(b), B_(a) +#define MOVZX_WL(a, b) movzx W_(b), W_(a) +#define MUL_L(a) mul L_(a) +#define MUL_W(a) mul W_(a) +#define MUL_B(a) mul B_(a) +#define NEG_L(a) neg L_(a) +#define NEG_W(a) neg W_(a) +#define NEG_B(a) neg B_(a) +#define NOP nop +#define NOT_L(a) not L_(a) +#define NOT_W(a) not W_(a) +#define NOT_B(a) not B_(a) +#define OR_L(a,b) or L_(b), L_(a) +#define OR_W(a,b) or W_(b), W_(a) +#define OR_B(a,b) or B_(b), B_(a) +#define OUT_L out DX, EAX +#define OUT_W out DX, AX +#define OUT_B out DX, AL +#define OUT1_L(a) out1 L_(a) +#define OUT1_W(a) out1 W_(a) +#define OUT1_B(a) out1 B_(a) +#define OUTS_L outsd +#define OUTS_W outsw +#define OUTS_B outsb +#define POP_SR(a) pop SR_(a) +#define POP_L(a) pop L_(a) +#define POP_W(a) pop W_(a) +#define POPA_L popad +#define POPA_W popa +#define POPF_L popfd +#define POPF_W popf +#define PUSH_SR(a) push SR_(a) +#define PUSH_L(a) push L_(a) +#define PUSH_W(a) push W_(a) +#define PUSH_B(a) push B_(a) +#define PUSHA_L pushad +#define PUSHA_W pusha +#define PUSHF_L pushfd +#define PUSHF_W pushf +#define RCL_L(a, b) rcl L_(b), L_(a) +#define RCL_W(a, b) rcl W_(b), W_(a) +#define RCL_B(a, b) rcl B_(b), B_(a) +#define RCR_L(a, b) rcr L_(b), L_(a) +#define RCR_W(a, b) rcr W_(b), W_(a) +#define RCR_B(a, b) rcr B_(b), B_(a) +#define RDTSC rdtsc +#define ROL_L(a, b) rol L_(b), L_(a) +#define ROL_W(a, b) rol W_(b), W_(a) +#define ROL_B(a, b) rol B_(b), B_(a) +#define ROR_L(a, b) ror L_(b), L_(a) +#define ROR_W(a, b) ror W_(b), W_(a) +#define ROR_B(a, b) ror B_(b), B_(a) +#define REP rep +#define REPE repe +#define REPNE repne +#define REPNZ REPNE +#define REPZ REPE +#define RET ret +#define SAHF sahf +#define SAL_L(a, b) sal L_(b), B_(a) +#define SAL_W(a, b) sal W_(b), B_(a) +#define SAL_B(a, b) sal B_(b), B_(a) +#define SAR_L(a, b) sar L_(b), B_(a) +#define SAR_W(a, b) sar W_(b), B_(a) +#define SAR_B(a, b) sar B_(b), B_(a) +#define SBB_L(a, b) sbb L_(b), L_(a) +#define SBB_W(a, b) sbb W_(b), W_(a) +#define SBB_B(a, b) sbb B_(b), B_(a) +#define SCAS_L scas +#define SCAS_W scas +#define SCAS_B scas +#define SETA(a) seta a +#define SETAE(a) setae a +#define SETB(a) setb a +#define SETBE(a) setbe a +#define SETC(a) setc a +#define SETE(a) sete a +#define SETG(a) setg a +#define SETGE(a) setge a +#define SETL(a) setl a +#define SETLE(a) setle a +#define SETNA(a) setna a +#define SETNAE(a) setnae a +#define SETNB(a) setnb a +#define SETNBE(a) setnbe a +#define SETNC(a) setnc a +#define SETNE(a) setne a +#define SETNG(a) setng a +#define SETNGE(a) setnge a +#define SETNL(a) setnl a +#define SETNLE(a) setnle a +#define SETNO(a) setno a +#define SETNP(a) setnp a +#define SETNS(a) setns a +#define SETNZ(a) setnz a +#define SETO(a) seto a +#define SETP(a) setp a +#define SETPE(a) setpe a +#define SETPO(a) setpo a +#define SETS(a) sets a +#define SETZ(a) setz a +#define SGDT(a) sgdt a +#define SIDT(a) sidt a +#define SHL_L(a, b) shl L_(b), B_(a) +#define SHL_W(a, b) shl W_(b), B_(a) +#define SHL_B(a, b) shl B_(b), B_(a) +#define SHLD_L(a,b,c) shld +#define SHLD2_L(a,b) shld L_(b), L_(a) +#define SHLD_W(a,b,c) shld +#define SHLD2_W(a,b) shld W_(b), W_(a) +#define SHR_L(a, b) shr L_(b), B_(a) +#define SHR_W(a, b) shr W_(b), B_(a) +#define SHR_B(a, b) shr B_(b), B_(a) +#define SHRD_L(a,b,c) shrd +#define SHRD2_L(a,b) shrd L_(b), L_(a) +#define SHRD_W(a,b,c) shrd +#define SHRD2_W(a,b) shrd W_(b), W_(a) +#define SLDT(a) sldt a +#define SMSW(a) smsw a +#define STC stc +#define STD std +#define STI sti +#define STOS_L stosd +#define STOS_W stosw +#define STOS_B stosb +#define STR(a) str a +#define SUB_L(a, b) sub L_(b), L_(a) +#define SUB_W(a, b) sub W_(b), W_(a) +#define SUB_B(a, b) sub B_(b), B_(a) +#define TEST_L(a, b) test L_(b), L_(a) +#define TEST_W(a, b) test W_(b), W_(a) +#define TEST_B(a, b) test B_(b), B_(a) +#define VERR(a) verr a +#define VERW(a) verw a +#define WAIT wait +#define XCHG_L(a, b) xchg L_(b), L_(a) +#define XCHG_W(a, b) xchg W_(b), W_(a) +#define XCHG_B(a, b) xchg B_(b), B_(a) +#define XLAT xlat +#define XOR_L(a, b) xor L_(b), L_(a) +#define XOR_W(a, b) xor W_(b), W_(a) +#define XOR_B(a, b) xor B_(b), B_(a) + + +/* Floating Point Instructions */ +#define F2XM1 f2xm1 +#define FABS fabs +#define FADD_D(a) fadd D_(a) +#define FADD_S(a) fadd S_(a) +#define FADD2(a, b) fadd b, a +#define FADDP(a, b) faddp b, a +#define FIADD_L(a) fiadd L_(a) +#define FIADD_W(a) fiadd W_(a) +#define FBLD(a) fbld a +#define FBSTP(a) fbstp a +#define FCHS fchs +#define FCLEX fclex +#define FNCLEX fnclex +#define FCOM(a) fcom a +#define FCOM_D(a) fcom D_(a) +#define FCOM_S(a) fcom S_(a) +#define FCOMP(a) fcomp a +#define FCOMP_D(a) fcomp D_(a) +#define FCOMP_S(a) fcomp S_(a) +#define FCOMPP fcompp +#define FCOS fcos +#define FDECSTP fdecstp +#define FDIV_D(a) fdiv D_(a) +#define FDIV_S(a) fdiv S_(a) +#define FDIV2(a, b) fdiv b, a +#define FDIVP(a, b) fdivp b, a +#define FIDIV_L(a) fidiv L_(a) +#define FIDIV_W(a) fidiv W_(a) +#define FDIVR_D(a) fdivr D_(a) +#define FDIVR_S(a) fdivr S_(a) +#define FDIVR2(a, b) fdivr b, a +#define FDIVRP(a, b) fdivrp b, a +#define FIDIVR_L(a) fidivr L_(a) +#define FIDIVR_W(a) fidivr W_(a) +#define FFREE(a) ffree a +#define FICOM_L(a) ficom L_(a) +#define FICOM_W(a) ficom W_(a) +#define FICOMP_L(a) ficomp L_(a) +#define FICOMP_W(a) ficomp W_(a) +#define FILD_Q(a) fild D_(a) +#define FILD_L(a) fild L_(a) +#define FILD_W(a) fild W_(a) +#define FINCSTP fincstp +#define FINIT finit +#define FNINIT fninit +#define FIST_L(a) fist L_(a) +#define FIST_W(a) fist W_(a) +#define FISTP_Q(a) fistp D_(a) +#define FISTP_L(a) fistp L_(a) +#define FISTP_W(a) fistp W_(a) +#define FLD_X(a) fld X_(a) +#define FLD_D(a) fld D_(a) +#define FLD_S(a) fld S_(a) +#define FLD1 fld1 +#define FLDL2T fldl2t +#define FLDL2E fldl2e +#define FLDPI fldpi +#define FLDLG2 fldlg2 +#define FLDLN2 fldln2 +#define FLDZ fldz +#define FLDCW(a) fldcw a +#define FLDENV(a) fldenv a +#define FMUL_S(a) fmul S_(a) +#define FMUL_D(a) fmul D_(a) +#define FMUL2(a, b) fmul b, a +#define FMULP(a, b) fmulp b, a +#define FIMUL_L(a) fimul L_(a) +#define FIMUL_W(a) fimul W_(a) +#define FNOP fnop +#define FPATAN fpatan +#define FPREM fprem +#define FPREM1 fprem1 +#define FPTAN fptan +#define FRNDINT frndint +#define FRSTOR(a) frstor a +#define FSAVE(a) fsave a +#define FNSAVE(a) fnsave a +#define FSCALE fscale +#define FSIN fsin +#define FSINCOS fsincos +#define FSQRT fsqrt +#define FST_D(a) fst D_(a) +#define FST_S(a) fst S_(a) +#define FSTP_X(a) fstp X_(a) +#define FSTP_D(a) fstp D_(a) +#define FSTP_S(a) fstp S_(a) +#define FSTP(a) fstp a +#define FSTCW(a) fstcw a +#define FNSTCW(a) fnstcw a +#define FSTENV(a) fstenv a +#define FNSTENV(a) fnstenv a +#define FSTSW(a) fstsw a +#define FNSTSW(a) fnstsw a +#define FSUB_S(a) fsub S_(a) +#define FSUB_D(a) fsub D_(a) +#define FSUB2(a, b) fsub b, a +#define FSUBP(a, b) fsubp b, a +#define FISUB_L(a) fisub L_(a) +#define FISUB_W(a) fisub W_(a) +#define FSUBR_S(a) fsubr S_(a) +#define FSUBR_D(a) fsubr D_(a) +#define FSUBR2(a, b) fsubr b, a +#define FSUBRP(a, b) fsubrp b, a +#define FISUBR_L(a) fisubr L_(a) +#define FISUBR_W(a) fisubr W_(a) +#define FTST ftst +#define FUCOM(a) fucom a +#define FUCOMP(a) fucomp a +#define FUCOMPP fucompp +#define FWAIT fwait +#define FXAM fxam +#define FXCH(a) fxch a +#define FXTRACT fxtract +#define FYL2X fyl2x +#define FYL2XP1 fyl2xp1 + +#endif /* NASM_ASSEMBLER, MASM_ASSEMBLER */ + + /****************************************/ + /* */ + /* Extensions to x86 insn set - */ + /* MMX, 3DNow! */ + /* */ + /****************************************/ + +#if defined(NASM_ASSEMBLER) || defined(MASM_ASSEMBLER) +#define P_ARG1(a) P_ ## a +#define P_ARG2(a, b) P_ ## b, P_ ## a +#define P_ARG3(a, b, c) P_ ## c, P_ ## b, P_ ## a +#else +#define P_ARG1(a) a +#define P_ARG2(a, b) a, b +#define P_ARG3(a, b, c) a, b, c +#endif + +/* MMX */ +#define MOVD(a, b) movd P_ARG2(a, b) +#define MOVQ(a, b) movq P_ARG2(a, b) + +#define PADDB(a, b) paddb P_ARG2(a, b) +#define PADDW(a, b) paddw P_ARG2(a, b) +#define PADDD(a, b) paddd P_ARG2(a, b) + +#define PADDSB(a, b) paddsb P_ARG2(a, b) +#define PADDSW(a, b) paddsw P_ARG2(a, b) + +#define PADDUSB(a, b) paddusb P_ARG2(a, b) +#define PADDUSW(a, b) paddusw P_ARG2(a, b) + +#define PSUBB(a, b) psubb P_ARG2(a, b) +#define PSUBW(a, b) psubw P_ARG2(a, b) +#define PSUBD(a, b) psubd P_ARG2(a, b) + +#define PSUBSB(a, b) psubsb P_ARG2(a, b) +#define PSUBSW(a, b) psubsw P_ARG2(a, b) + +#define PSUBUSB(a, b) psubusb P_ARG2(a, b) +#define PSUBUSW(a, b) psubusw P_ARG2(a, b) + +#define PCMPEQB(a, b) pcmpeqb P_ARG2(a, b) +#define PCMPEQW(a, b) pcmpeqw P_ARG2(a, b) +#define PCMPEQD(a, b) pcmpeqd P_ARG2(a, b) + +#define PCMPGTB(a, b) pcmpgtb P_ARG2(a, b) +#define PCMPGTW(a, b) pcmpgtw P_ARG2(a, b) +#define PCMPGTD(a, b) pcmpgtd P_ARG2(a, b) + +#define PMULHW(a, b) pmulhw P_ARG2(a, b) +#define PMULLW(a, b) pmullw P_ARG2(a, b) + +#define PMADDWD(a, b) pmaddwd P_ARG2(a, b) + +#define PAND(a, b) pand P_ARG2(a, b) + +#define PANDN(a, b) pandn P_ARG2(a, b) + +#define POR(a, b) por P_ARG2(a, b) + +#define PXOR(a, b) pxor P_ARG2(a, b) + +#define PSRAW(a, b) psraw P_ARG2(a, b) +#define PSRAD(a, b) psrad P_ARG2(a, b) + +#define PSRLW(a, b) psrlw P_ARG2(a, b) +#define PSRLD(a, b) psrld P_ARG2(a, b) +#define PSRLQ(a, b) psrlq P_ARG2(a, b) + +#define PSLLW(a, b) psllw P_ARG2(a, b) +#define PSLLD(a, b) pslld P_ARG2(a, b) +#define PSLLQ(a, b) psllq P_ARG2(a, b) + +#define PACKSSWB(a, b) packsswb P_ARG2(a, b) +#define PACKSSDW(a, b) packssdw P_ARG2(a, b) +#define PACKUSWB(a, b) packuswb P_ARG2(a, b) + +#define PUNPCKHBW(a, b) punpckhbw P_ARG2(a, b) +#define PUNPCKHWD(a, b) punpckhwd P_ARG2(a, b) +#define PUNPCKHDQ(a, b) punpckhdq P_ARG2(a, b) +#define PUNPCKLBW(a, b) punpcklbw P_ARG2(a, b) +#define PUNPCKLWD(a, b) punpcklwd P_ARG2(a, b) +#define PUNPCKLDQ(a, b) punpckldq P_ARG2(a, b) + +#define EMMS emms + +/* AMD 3DNow! */ +#define PAVGUSB(a, b) pavgusb P_ARG2(a, b) +#define PFADD(a, b) pfadd P_ARG2(a, b) +#define PFSUB(a, b) pfsub P_ARG2(a, b) +#define PFSUBR(a, b) pfsubr P_ARG2(a, b) +#define PFACC(a, b) pfacc P_ARG2(a, b) +#define PFCMPGE(a, b) pfcmpge P_ARG2(a, b) +#define PFCMPGT(a, b) pfcmpgt P_ARG2(a, b) +#define PFCMPEQ(a, b) pfcmpeq P_ARG2(a, b) +#define PFMIN(a, b) pfmin P_ARG2(a, b) +#define PFMAX(a, b) pfmax P_ARG2(a, b) +#define PI2FD(a, b) pi2fd P_ARG2(a, b) +#define PF2ID(a, b) pf2id P_ARG2(a, b) +#define PFRCP(a, b) pfrcp P_ARG2(a, b) +#define PFRSQRT(a, b) pfrsqrt P_ARG2(a, b) +#define PFMUL(a, b) pfmul P_ARG2(a, b) +#define PFRCPIT1(a, b) pfrcpit1 P_ARG2(a, b) +#define PFRSQIT1(a, b) pfrsqit1 P_ARG2(a, b) +#define PFRCPIT2(a, b) pfrcpit2 P_ARG2(a, b) +#define PMULHRW(a, b) pmulhrw P_ARG2(a, b) + +#define FEMMS femms +#define PREFETCH(a) prefetch P_ARG1(a) +#define PREFETCHW(a) prefetchw P_ARG1(a) + +/* Intel SSE */ +#define ADDPS(a, b) addps P_ARG2(a, b) +#define ADDSS(a, b) addss P_ARG2(a, b) +#define ANDNPS(a, b) andnps P_ARG2(a, b) +#define ANDPS(a, b) andps P_ARG2(a, b) +/* NASM only knows the pseudo ops for these. +#define CMPPS(a, b, c) cmpps P_ARG3(a, b, c) +#define CMPSS(a, b, c) cmpss P_ARG3(a, b, c) +*/ +#define CMPEQPS(a, b) cmpeqps P_ARG2(a, b) +#define CMPLTPS(a, b) cmpltps P_ARG2(a, b) +#define CMPLEPS(a, b) cmpleps P_ARG2(a, b) +#define CMPUNORDPS(a, b) cmpunordps P_ARG2(a, b) +#define CMPNEQPS(a, b) cmpneqps P_ARG2(a, b) +#define CMPNLTPS(a, b) cmpnltps P_ARG2(a, b) +#define CMPNLEPS(a, b) cmpnleps P_ARG2(a, b) +#define CMPORDPS(a, b) cmpordps P_ARG2(a, b) +#define CMPEQSS(a, b) cmpeqss P_ARG2(a, b) +#define CMPLTSS(a, b) cmpltss P_ARG2(a, b) +#define CMPLESS(a, b) cmpless P_ARG2(a, b) +#define CMPUNORDSS(a, b) cmpunordss P_ARG2(a, b) +#define CMPNEQSS(a, b) cmpneqss P_ARG2(a, b) +#define CMPNLTSS(a, b) cmpnltss P_ARG2(a, b) +#define CMPNLESS(a, b) cmpnless P_ARG2(a, b) +#define CMPORDSS(a, b) cmpordss P_ARG2(a, b) +#define COMISS(a, b) comiss P_ARG2(a, b) +#define CVTPI2PS(a, b) cvtpi2ps P_ARG2(a, b) +#define CVTPS2PI(a, b) cvtps2pi P_ARG2(a, b) +#define CVTSI2SS(a, b) cvtsi2ss P_ARG2(a, b) +#define CVTSS2SI(a, b) cvtss2si P_ARG2(a, b) +#define CVTTPS2PI(a, b) cvttps2pi P_ARG2(a, b) +#define CVTTSS2SI(a, b) cvttss2si P_ARG2(a, b) +#define DIVPS(a, b) divps P_ARG2(a, b) +#define DIVSS(a, b) divss P_ARG2(a, b) +#define FXRSTOR(a) fxrstor P_ARG1(a) +#define FXSAVE(a) fxsave P_ARG1(a) +#define LDMXCSR(a) ldmxcsr P_ARG1(a) +#define MAXPS(a, b) maxps P_ARG2(a, b) +#define MAXSS(a, b) maxss P_ARG2(a, b) +#define MINPS(a, b) minps P_ARG2(a, b) +#define MINSS(a, b) minss P_ARG2(a, b) +#define MOVAPS(a, b) movaps P_ARG2(a, b) +#define MOVHLPS(a, b) movhlps P_ARG2(a, b) +#define MOVHPS(a, b) movhps P_ARG2(a, b) +#define MOVLHPS(a, b) movlhps P_ARG2(a, b) +#define MOVLPS(a, b) movlps P_ARG2(a, b) +#define MOVMSKPS(a, b) movmskps P_ARG2(a, b) +#define MOVNTPS(a, b) movntps P_ARG2(a, b) +#define MOVNTQ(a, b) movntq P_ARG2(a, b) +#define MOVSS(a, b) movss P_ARG2(a, b) +#define MOVUPS(a, b) movups P_ARG2(a, b) +#define MULPS(a, b) mulps P_ARG2(a, b) +#define MULSS(a, b) mulss P_ARG2(a, b) +#define ORPS(a, b) orps P_ARG2(a, b) +#define RCPPS(a, b) rcpps P_ARG2(a, b) +#define RCPSS(a, b) rcpss P_ARG2(a, b) +#define RSQRTPS(a, b) rsqrtps P_ARG2(a, b) +#define RSQRTSS(a, b) rsqrtss P_ARG2(a, b) +#define SHUFPS(a, b, c) shufps P_ARG3(a, b, c) +#define SQRTPS(a, b) sqrtps P_ARG2(a, b) +#define SQRTSS(a, b) sqrtss P_ARG2(a, b) +#define STMXCSR(a) stmxcsr P_ARG1(a) +#define SUBPS(a, b) subps P_ARG2(a, b) +#define UCOMISS(a, b) ucomiss P_ARG2(a, b) +#define UNPCKHPS(a, b) unpckhps P_ARG2(a, b) +#define UNPCKLPS(a, b) unpcklps P_ARG2(a, b) +#define XORPS(a, b) xorps P_ARG2(a, b) + +#define PREFETCHNTA(a) prefetchnta P_ARG1(a) +#define PREFETCHT0(a) prefetcht0 P_ARG1(a) +#define PREFETCHT1(a) prefetcht1 P_ARG1(a) +#define PREFETCHT2(a) prefetcht2 P_ARG1(a) +#define SFENCE sfence + +/* Added by BrianP for FreeBSD (per David Dawes) */ +#if !defined(NASM_ASSEMBLER) && !defined(MASM_ASSEMBLER) && !defined(__bsdi__) +#define LLBL(a) CONCAT(.L,a) +#define LLBL2(a,b) CONCAT3(.L,a,b) +#else +#define LLBL(a) a +#define LLBL2(a,b) CONCAT(a,b) +#endif + +/* Segment overrides */ +#define SEGCS D_BYTE 46 +#define SEGDS D_BYTE 62 +#define SEGES D_BYTE 38 +#define SEGFS D_BYTE 100 +#define SEGGS D_BYTE 101 + +/* Temporary labels: valid until next non-local label */ +#ifdef NASM_ASSEMBLER +#define TLBL(a) CONCAT(.,a) +#else +#define TLBL(a) CONCAT(a,$) +#endif + +/* Hidden symbol visibility support. + * If we build with gcc's -fvisibility=hidden flag, we'll need to change + * the symbol visibility mode to 'default'. + */ +#if defined(GNU_ASSEMBLER) && !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) +# define HIDDEN(x) .hidden x +#elif defined(__GNUC__) && !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) +# pragma GCC visibility push(default) +# define HIDDEN(x) .hidden x +#else +# define HIDDEN(x) +#endif + +#endif /* __ASSYNTAX_H__ */ diff --git a/mesalib/src/mesa/x86/clip_args.h b/mesalib/src/mesa/x86/clip_args.h index faa3c6f87..796611fbf 100644 --- a/mesalib/src/mesa/x86/clip_args.h +++ b/mesalib/src/mesa/x86/clip_args.h @@ -1,59 +1,59 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * Clip test function interface for assembly code. Simply define - * FRAME_OFFSET to the number of bytes pushed onto the stack before - * using the ARG_* argument macros. - * - * Gareth Hughes - */ - -#ifndef __CLIP_ARGS_H__ -#define __CLIP_ARGS_H__ - -/* - * Offsets for clip_func arguments - * - * typedef GLvector4f *(*clip_func)( GLvector4f *clip_vec, - * GLvector4f *proj_vec, - * GLubyte clipMask[], - * GLubyte *orMask, - * GLubyte *andMask ); - */ - -#define OFFSET_SOURCE 4 -#define OFFSET_DEST 8 -#define OFFSET_CLIP 12 -#define OFFSET_OR 16 -#define OFFSET_AND 20 - -#define ARG_SOURCE REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP) -#define ARG_DEST REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP) -#define ARG_CLIP REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP) -#define ARG_OR REGOFF(FRAME_OFFSET+OFFSET_OR, ESP) -#define ARG_AND REGOFF(FRAME_OFFSET+OFFSET_AND, ESP) - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * Clip test function interface for assembly code. Simply define + * FRAME_OFFSET to the number of bytes pushed onto the stack before + * using the ARG_* argument macros. + * + * Gareth Hughes + */ + +#ifndef __CLIP_ARGS_H__ +#define __CLIP_ARGS_H__ + +/* + * Offsets for clip_func arguments + * + * typedef GLvector4f *(*clip_func)( GLvector4f *clip_vec, + * GLvector4f *proj_vec, + * GLubyte clipMask[], + * GLubyte *orMask, + * GLubyte *andMask ); + */ + +#define OFFSET_SOURCE 4 +#define OFFSET_DEST 8 +#define OFFSET_CLIP 12 +#define OFFSET_OR 16 +#define OFFSET_AND 20 + +#define ARG_SOURCE REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP) +#define ARG_DEST REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP) +#define ARG_CLIP REGOFF(FRAME_OFFSET+OFFSET_CLIP, ESP) +#define ARG_OR REGOFF(FRAME_OFFSET+OFFSET_OR, ESP) +#define ARG_AND REGOFF(FRAME_OFFSET+OFFSET_AND, ESP) + +#endif diff --git a/mesalib/src/mesa/x86/common_x86.c b/mesalib/src/mesa/x86/common_x86.c index f5b9357cc..b70ee5084 100644 --- a/mesalib/src/mesa/x86/common_x86.c +++ b/mesalib/src/mesa/x86/common_x86.c @@ -1,336 +1,336 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.1 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file common_x86.c - * - * Check CPU capabilities & initialize optimized funtions for this particular - * processor. - * - * Changed by Andre Werthmann for using the new SSE functions. - * - * \author Holger Waechtler - * \author Andre Werthmann - */ - -/* XXX these includes should probably go into imports.h or glheader.h */ -#if defined(USE_SSE_ASM) && defined(__linux__) -#include -#endif -#if defined(USE_SSE_ASM) && defined(__FreeBSD__) -#include -#include -#endif -#if defined(USE_SSE_ASM) && defined(__OpenBSD__) -#include -#include -#include -#endif - -#include "main/imports.h" -#include "common_x86_asm.h" - - -/** Bitmask of X86_FEATURE_x bits */ -int _mesa_x86_cpu_features = 0x0; - -static int detection_debug = GL_FALSE; - -/* No reason for this to be public. - */ -extern GLuint _ASMAPI _mesa_x86_has_cpuid(void); -extern void _ASMAPI _mesa_x86_cpuid(GLuint op, GLuint *reg_eax, GLuint *reg_ebx, GLuint *reg_ecx, GLuint *reg_edx); -extern GLuint _ASMAPI _mesa_x86_cpuid_eax(GLuint op); -extern GLuint _ASMAPI _mesa_x86_cpuid_ebx(GLuint op); -extern GLuint _ASMAPI _mesa_x86_cpuid_ecx(GLuint op); -extern GLuint _ASMAPI _mesa_x86_cpuid_edx(GLuint op); - - -#if defined(USE_SSE_ASM) -/* - * We must verify that the Streaming SIMD Extensions are truly supported - * on this processor before we go ahead and hook out the optimized code. - * - * However, I have been told by Alan Cox that all 2.4 (and later) Linux - * kernels provide full SSE support on all processors that expose SSE via - * the CPUID mechanism. - */ - -/* These are assembly functions: */ -extern void _mesa_test_os_sse_support( void ); -extern void _mesa_test_os_sse_exception_support( void ); - - -#if defined(WIN32) -#ifndef STATUS_FLOAT_MULTIPLE_TRAPS -# define STATUS_FLOAT_MULTIPLE_TRAPS (0xC00002B5L) -#endif -static LONG WINAPI ExceptionFilter(LPEXCEPTION_POINTERS exp) -{ - PEXCEPTION_RECORD rec = exp->ExceptionRecord; - PCONTEXT ctx = exp->ContextRecord; - - if ( rec->ExceptionCode == EXCEPTION_ILLEGAL_INSTRUCTION ) { - _mesa_debug(NULL, "EXCEPTION_ILLEGAL_INSTRUCTION\n" ); - _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); - } else if ( rec->ExceptionCode == STATUS_FLOAT_MULTIPLE_TRAPS ) { - _mesa_debug(NULL, "STATUS_FLOAT_MULTIPLE_TRAPS\n"); - /* Windows seems to clear the exception flag itself, we just have to increment Eip */ - } else { - _mesa_debug(NULL, "UNEXPECTED EXCEPTION (0x%08x), terminating!\n" ); - return EXCEPTION_EXECUTE_HANDLER; - } - - if ( (ctx->ContextFlags & CONTEXT_CONTROL) != CONTEXT_CONTROL ) { - _mesa_debug(NULL, "Context does not contain control registers, terminating!\n"); - return EXCEPTION_EXECUTE_HANDLER; - } - ctx->Eip += 3; - - return EXCEPTION_CONTINUE_EXECUTION; -} -#endif /* WIN32 */ - - -/** - * Check if SSE is supported. - * If not, turn off the X86_FEATURE_XMM flag in _mesa_x86_cpu_features. - */ -void _mesa_check_os_sse_support( void ) -{ -#if defined(__FreeBSD__) - { - int ret, enabled; - unsigned int len; - len = sizeof(enabled); - ret = sysctlbyname("hw.instruction_sse", &enabled, &len, NULL, 0); - if (ret || !enabled) - _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); - } -#elif defined (__NetBSD__) - { - int ret, enabled; - size_t len = sizeof(enabled); - ret = sysctlbyname("machdep.sse", &enabled, &len, (void *)NULL, 0); - if (ret || !enabled) - _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); - } -#elif defined(__OpenBSD__) - { - int mib[2]; - int ret, enabled; - size_t len = sizeof(enabled); - - mib[0] = CTL_MACHDEP; - mib[1] = CPU_SSE; - - ret = sysctl(mib, 2, &enabled, &len, NULL, 0); - if (ret || !enabled) - _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); - } -#elif defined(WIN32) - LPTOP_LEVEL_EXCEPTION_FILTER oldFilter; - - /* Install our ExceptionFilter */ - oldFilter = SetUnhandledExceptionFilter( ExceptionFilter ); - - if ( cpu_has_xmm ) { - _mesa_debug(NULL, "Testing OS support for SSE...\n"); - - _mesa_test_os_sse_support(); - - if ( cpu_has_xmm ) { - _mesa_debug(NULL, "Yes.\n"); - } else { - _mesa_debug(NULL, "No!\n"); - } - } - - if ( cpu_has_xmm ) { - _mesa_debug(NULL, "Testing OS support for SSE unmasked exceptions...\n"); - - _mesa_test_os_sse_exception_support(); - - if ( cpu_has_xmm ) { - _mesa_debug(NULL, "Yes.\n"); - } else { - _mesa_debug(NULL, "No!\n"); - } - } - - /* Restore previous exception filter */ - SetUnhandledExceptionFilter( oldFilter ); - - if ( cpu_has_xmm ) { - _mesa_debug(NULL, "Tests of OS support for SSE passed.\n"); - } else { - _mesa_debug(NULL, "Tests of OS support for SSE failed!\n"); - } -#else - /* Do nothing on other platforms for now. - */ - if (detection_debug) - _mesa_debug(NULL, "Not testing OS support for SSE, leaving enabled.\n"); -#endif /* __FreeBSD__ */ -} - -#endif /* USE_SSE_ASM */ - - -/** - * Initialize the _mesa_x86_cpu_features bitfield. - * This is a no-op if called more than once. - */ -void -_mesa_get_x86_features(void) -{ - static int called = 0; - - if (called) - return; - - called = 1; - -#ifdef USE_X86_ASM - _mesa_x86_cpu_features = 0x0; - - if (_mesa_getenv( "MESA_NO_ASM")) { - return; - } - - if (!_mesa_x86_has_cpuid()) { - _mesa_debug(NULL, "CPUID not detected\n"); - } - else { - GLuint cpu_features; - GLuint cpu_ext_features; - GLuint cpu_ext_info; - char cpu_vendor[13]; - GLuint result; - - /* get vendor name */ - _mesa_x86_cpuid(0, &result, (GLuint *)(cpu_vendor + 0), (GLuint *)(cpu_vendor + 8), (GLuint *)(cpu_vendor + 4)); - cpu_vendor[12] = '\0'; - - if (detection_debug) - _mesa_debug(NULL, "CPU vendor: %s\n", cpu_vendor); - - /* get cpu features */ - cpu_features = _mesa_x86_cpuid_edx(1); - - if (cpu_features & X86_CPU_FPU) - _mesa_x86_cpu_features |= X86_FEATURE_FPU; - if (cpu_features & X86_CPU_CMOV) - _mesa_x86_cpu_features |= X86_FEATURE_CMOV; - -#ifdef USE_MMX_ASM - if (cpu_features & X86_CPU_MMX) - _mesa_x86_cpu_features |= X86_FEATURE_MMX; -#endif - -#ifdef USE_SSE_ASM - if (cpu_features & X86_CPU_XMM) - _mesa_x86_cpu_features |= X86_FEATURE_XMM; - if (cpu_features & X86_CPU_XMM2) - _mesa_x86_cpu_features |= X86_FEATURE_XMM2; -#endif - - /* query extended cpu features */ - if ((cpu_ext_info = _mesa_x86_cpuid_eax(0x80000000)) > 0x80000000) { - if (cpu_ext_info >= 0x80000001) { - - cpu_ext_features = _mesa_x86_cpuid_edx(0x80000001); - - if (cpu_features & X86_CPU_MMX) { - -#ifdef USE_3DNOW_ASM - if (cpu_ext_features & X86_CPUEXT_3DNOW) - _mesa_x86_cpu_features |= X86_FEATURE_3DNOW; - if (cpu_ext_features & X86_CPUEXT_3DNOW_EXT) - _mesa_x86_cpu_features |= X86_FEATURE_3DNOWEXT; -#endif - -#ifdef USE_MMX_ASM - if (cpu_ext_features & X86_CPUEXT_MMX_EXT) - _mesa_x86_cpu_features |= X86_FEATURE_MMXEXT; -#endif - } - } - - /* query cpu name */ - if (cpu_ext_info >= 0x80000002) { - GLuint ofs; - char cpu_name[49]; - for (ofs = 0; ofs < 3; ofs++) - _mesa_x86_cpuid(0x80000002+ofs, (GLuint *)(cpu_name + (16*ofs)+0), (GLuint *)(cpu_name + (16*ofs)+4), (GLuint *)(cpu_name + (16*ofs)+8), (GLuint *)(cpu_name + (16*ofs)+12)); - cpu_name[48] = '\0'; /* the name should be NULL terminated, but just to be sure */ - - if (detection_debug) - _mesa_debug(NULL, "CPU name: %s\n", cpu_name); - } - } - - } - -#ifdef USE_MMX_ASM - if ( cpu_has_mmx ) { - if ( _mesa_getenv( "MESA_NO_MMX" ) == 0 ) { - if (detection_debug) - _mesa_debug(NULL, "MMX cpu detected.\n"); - } else { - _mesa_x86_cpu_features &= ~(X86_FEATURE_MMX); - } - } -#endif - -#ifdef USE_3DNOW_ASM - if ( cpu_has_3dnow ) { - if ( _mesa_getenv( "MESA_NO_3DNOW" ) == 0 ) { - if (detection_debug) - _mesa_debug(NULL, "3DNow! cpu detected.\n"); - } else { - _mesa_x86_cpu_features &= ~(X86_FEATURE_3DNOW); - } - } -#endif - -#ifdef USE_SSE_ASM - if ( cpu_has_xmm ) { - if ( _mesa_getenv( "MESA_NO_SSE" ) == 0 ) { - if (detection_debug) - _mesa_debug(NULL, "SSE cpu detected.\n"); - if ( _mesa_getenv( "MESA_FORCE_SSE" ) == 0 ) { - _mesa_check_os_sse_support(); - } - } else { - _mesa_debug(NULL, "SSE cpu detected, but switched off by user.\n"); - _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); - } - } -#endif - -#endif /* USE_X86_ASM */ - - (void) detection_debug; -} +/* + * Mesa 3-D graphics library + * Version: 6.5.1 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file common_x86.c + * + * Check CPU capabilities & initialize optimized funtions for this particular + * processor. + * + * Changed by Andre Werthmann for using the new SSE functions. + * + * \author Holger Waechtler + * \author Andre Werthmann + */ + +/* XXX these includes should probably go into imports.h or glheader.h */ +#if defined(USE_SSE_ASM) && defined(__linux__) +#include +#endif +#if defined(USE_SSE_ASM) && defined(__FreeBSD__) +#include +#include +#endif +#if defined(USE_SSE_ASM) && defined(__OpenBSD__) +#include +#include +#include +#endif + +#include "main/imports.h" +#include "common_x86_asm.h" + + +/** Bitmask of X86_FEATURE_x bits */ +int _mesa_x86_cpu_features = 0x0; + +static int detection_debug = GL_FALSE; + +/* No reason for this to be public. + */ +extern GLuint _ASMAPI _mesa_x86_has_cpuid(void); +extern void _ASMAPI _mesa_x86_cpuid(GLuint op, GLuint *reg_eax, GLuint *reg_ebx, GLuint *reg_ecx, GLuint *reg_edx); +extern GLuint _ASMAPI _mesa_x86_cpuid_eax(GLuint op); +extern GLuint _ASMAPI _mesa_x86_cpuid_ebx(GLuint op); +extern GLuint _ASMAPI _mesa_x86_cpuid_ecx(GLuint op); +extern GLuint _ASMAPI _mesa_x86_cpuid_edx(GLuint op); + + +#if defined(USE_SSE_ASM) +/* + * We must verify that the Streaming SIMD Extensions are truly supported + * on this processor before we go ahead and hook out the optimized code. + * + * However, I have been told by Alan Cox that all 2.4 (and later) Linux + * kernels provide full SSE support on all processors that expose SSE via + * the CPUID mechanism. + */ + +/* These are assembly functions: */ +extern void _mesa_test_os_sse_support( void ); +extern void _mesa_test_os_sse_exception_support( void ); + + +#if defined(WIN32) +#ifndef STATUS_FLOAT_MULTIPLE_TRAPS +# define STATUS_FLOAT_MULTIPLE_TRAPS (0xC00002B5L) +#endif +static LONG WINAPI ExceptionFilter(LPEXCEPTION_POINTERS exp) +{ + PEXCEPTION_RECORD rec = exp->ExceptionRecord; + PCONTEXT ctx = exp->ContextRecord; + + if ( rec->ExceptionCode == EXCEPTION_ILLEGAL_INSTRUCTION ) { + _mesa_debug(NULL, "EXCEPTION_ILLEGAL_INSTRUCTION\n" ); + _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); + } else if ( rec->ExceptionCode == STATUS_FLOAT_MULTIPLE_TRAPS ) { + _mesa_debug(NULL, "STATUS_FLOAT_MULTIPLE_TRAPS\n"); + /* Windows seems to clear the exception flag itself, we just have to increment Eip */ + } else { + _mesa_debug(NULL, "UNEXPECTED EXCEPTION (0x%08x), terminating!\n" ); + return EXCEPTION_EXECUTE_HANDLER; + } + + if ( (ctx->ContextFlags & CONTEXT_CONTROL) != CONTEXT_CONTROL ) { + _mesa_debug(NULL, "Context does not contain control registers, terminating!\n"); + return EXCEPTION_EXECUTE_HANDLER; + } + ctx->Eip += 3; + + return EXCEPTION_CONTINUE_EXECUTION; +} +#endif /* WIN32 */ + + +/** + * Check if SSE is supported. + * If not, turn off the X86_FEATURE_XMM flag in _mesa_x86_cpu_features. + */ +void _mesa_check_os_sse_support( void ) +{ +#if defined(__FreeBSD__) + { + int ret, enabled; + unsigned int len; + len = sizeof(enabled); + ret = sysctlbyname("hw.instruction_sse", &enabled, &len, NULL, 0); + if (ret || !enabled) + _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); + } +#elif defined (__NetBSD__) + { + int ret, enabled; + size_t len = sizeof(enabled); + ret = sysctlbyname("machdep.sse", &enabled, &len, (void *)NULL, 0); + if (ret || !enabled) + _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); + } +#elif defined(__OpenBSD__) + { + int mib[2]; + int ret, enabled; + size_t len = sizeof(enabled); + + mib[0] = CTL_MACHDEP; + mib[1] = CPU_SSE; + + ret = sysctl(mib, 2, &enabled, &len, NULL, 0); + if (ret || !enabled) + _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); + } +#elif defined(WIN32) + LPTOP_LEVEL_EXCEPTION_FILTER oldFilter; + + /* Install our ExceptionFilter */ + oldFilter = SetUnhandledExceptionFilter( ExceptionFilter ); + + if ( cpu_has_xmm ) { + _mesa_debug(NULL, "Testing OS support for SSE...\n"); + + _mesa_test_os_sse_support(); + + if ( cpu_has_xmm ) { + _mesa_debug(NULL, "Yes.\n"); + } else { + _mesa_debug(NULL, "No!\n"); + } + } + + if ( cpu_has_xmm ) { + _mesa_debug(NULL, "Testing OS support for SSE unmasked exceptions...\n"); + + _mesa_test_os_sse_exception_support(); + + if ( cpu_has_xmm ) { + _mesa_debug(NULL, "Yes.\n"); + } else { + _mesa_debug(NULL, "No!\n"); + } + } + + /* Restore previous exception filter */ + SetUnhandledExceptionFilter( oldFilter ); + + if ( cpu_has_xmm ) { + _mesa_debug(NULL, "Tests of OS support for SSE passed.\n"); + } else { + _mesa_debug(NULL, "Tests of OS support for SSE failed!\n"); + } +#else + /* Do nothing on other platforms for now. + */ + if (detection_debug) + _mesa_debug(NULL, "Not testing OS support for SSE, leaving enabled.\n"); +#endif /* __FreeBSD__ */ +} + +#endif /* USE_SSE_ASM */ + + +/** + * Initialize the _mesa_x86_cpu_features bitfield. + * This is a no-op if called more than once. + */ +void +_mesa_get_x86_features(void) +{ + static int called = 0; + + if (called) + return; + + called = 1; + +#ifdef USE_X86_ASM + _mesa_x86_cpu_features = 0x0; + + if (_mesa_getenv( "MESA_NO_ASM")) { + return; + } + + if (!_mesa_x86_has_cpuid()) { + _mesa_debug(NULL, "CPUID not detected\n"); + } + else { + GLuint cpu_features; + GLuint cpu_ext_features; + GLuint cpu_ext_info; + char cpu_vendor[13]; + GLuint result; + + /* get vendor name */ + _mesa_x86_cpuid(0, &result, (GLuint *)(cpu_vendor + 0), (GLuint *)(cpu_vendor + 8), (GLuint *)(cpu_vendor + 4)); + cpu_vendor[12] = '\0'; + + if (detection_debug) + _mesa_debug(NULL, "CPU vendor: %s\n", cpu_vendor); + + /* get cpu features */ + cpu_features = _mesa_x86_cpuid_edx(1); + + if (cpu_features & X86_CPU_FPU) + _mesa_x86_cpu_features |= X86_FEATURE_FPU; + if (cpu_features & X86_CPU_CMOV) + _mesa_x86_cpu_features |= X86_FEATURE_CMOV; + +#ifdef USE_MMX_ASM + if (cpu_features & X86_CPU_MMX) + _mesa_x86_cpu_features |= X86_FEATURE_MMX; +#endif + +#ifdef USE_SSE_ASM + if (cpu_features & X86_CPU_XMM) + _mesa_x86_cpu_features |= X86_FEATURE_XMM; + if (cpu_features & X86_CPU_XMM2) + _mesa_x86_cpu_features |= X86_FEATURE_XMM2; +#endif + + /* query extended cpu features */ + if ((cpu_ext_info = _mesa_x86_cpuid_eax(0x80000000)) > 0x80000000) { + if (cpu_ext_info >= 0x80000001) { + + cpu_ext_features = _mesa_x86_cpuid_edx(0x80000001); + + if (cpu_features & X86_CPU_MMX) { + +#ifdef USE_3DNOW_ASM + if (cpu_ext_features & X86_CPUEXT_3DNOW) + _mesa_x86_cpu_features |= X86_FEATURE_3DNOW; + if (cpu_ext_features & X86_CPUEXT_3DNOW_EXT) + _mesa_x86_cpu_features |= X86_FEATURE_3DNOWEXT; +#endif + +#ifdef USE_MMX_ASM + if (cpu_ext_features & X86_CPUEXT_MMX_EXT) + _mesa_x86_cpu_features |= X86_FEATURE_MMXEXT; +#endif + } + } + + /* query cpu name */ + if (cpu_ext_info >= 0x80000002) { + GLuint ofs; + char cpu_name[49]; + for (ofs = 0; ofs < 3; ofs++) + _mesa_x86_cpuid(0x80000002+ofs, (GLuint *)(cpu_name + (16*ofs)+0), (GLuint *)(cpu_name + (16*ofs)+4), (GLuint *)(cpu_name + (16*ofs)+8), (GLuint *)(cpu_name + (16*ofs)+12)); + cpu_name[48] = '\0'; /* the name should be NULL terminated, but just to be sure */ + + if (detection_debug) + _mesa_debug(NULL, "CPU name: %s\n", cpu_name); + } + } + + } + +#ifdef USE_MMX_ASM + if ( cpu_has_mmx ) { + if ( _mesa_getenv( "MESA_NO_MMX" ) == 0 ) { + if (detection_debug) + _mesa_debug(NULL, "MMX cpu detected.\n"); + } else { + _mesa_x86_cpu_features &= ~(X86_FEATURE_MMX); + } + } +#endif + +#ifdef USE_3DNOW_ASM + if ( cpu_has_3dnow ) { + if ( _mesa_getenv( "MESA_NO_3DNOW" ) == 0 ) { + if (detection_debug) + _mesa_debug(NULL, "3DNow! cpu detected.\n"); + } else { + _mesa_x86_cpu_features &= ~(X86_FEATURE_3DNOW); + } + } +#endif + +#ifdef USE_SSE_ASM + if ( cpu_has_xmm ) { + if ( _mesa_getenv( "MESA_NO_SSE" ) == 0 ) { + if (detection_debug) + _mesa_debug(NULL, "SSE cpu detected.\n"); + if ( _mesa_getenv( "MESA_FORCE_SSE" ) == 0 ) { + _mesa_check_os_sse_support(); + } + } else { + _mesa_debug(NULL, "SSE cpu detected, but switched off by user.\n"); + _mesa_x86_cpu_features &= ~(X86_FEATURE_XMM); + } + } +#endif + +#endif /* USE_X86_ASM */ + + (void) detection_debug; +} diff --git a/mesalib/src/mesa/x86/common_x86_asm.S b/mesalib/src/mesa/x86/common_x86_asm.S index 1d61bde58..ea4047a0e 100644 --- a/mesalib/src/mesa/x86/common_x86_asm.S +++ b/mesalib/src/mesa/x86/common_x86_asm.S @@ -1,220 +1,220 @@ -/* - * Mesa 3-D graphics library - * Version: 6.3 - * - * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * Check extended CPU capabilities. Now justs returns the raw CPUID - * feature information, allowing the higher level code to interpret the - * results. - * - * Written by Holger Waechtler - * - * Cleaned up and simplified by Gareth Hughes - * - */ - -/* - * NOTE: Avoid using spaces in between '(' ')' and arguments, especially - * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces - * in there will break the build on some platforms. - */ - -#include "matypes.h" -#include "assyntax.h" -#include "common_x86_features.h" - - SEG_TEXT - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_x86_has_cpuid) -HIDDEN(_mesa_x86_has_cpuid) -GLNAME(_mesa_x86_has_cpuid): - - /* Test for the CPUID command. If the ID Flag bit in EFLAGS - * (bit 21) is writable, the CPUID command is present */ - PUSHF_L - POP_L (EAX) - MOV_L (EAX, ECX) - XOR_L (CONST(0x00200000), EAX) - PUSH_L (EAX) - POPF_L - PUSHF_L - POP_L (EAX) - - /* Verify the ID Flag bit has been written. */ - CMP_L (ECX, EAX) - SETNE (AL) - XOR_L (CONST(0xff), EAX) - - RET - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_x86_cpuid) -HIDDEN(_mesa_x86_cpuid) -GLNAME(_mesa_x86_cpuid): - - MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ - PUSH_L (EDI) - PUSH_L (EBX) - - CPUID - - MOV_L (REGOFF(16, ESP), EDI) /* *eax */ - MOV_L (EAX, REGIND(EDI)) - MOV_L (REGOFF(20, ESP), EDI) /* *ebx */ - MOV_L (EBX, REGIND(EDI)) - MOV_L (REGOFF(24, ESP), EDI) /* *ecx */ - MOV_L (ECX, REGIND(EDI)) - MOV_L (REGOFF(28, ESP), EDI) /* *edx */ - MOV_L (EDX, REGIND(EDI)) - - POP_L (EBX) - POP_L (EDI) - RET - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_x86_cpuid_eax) -HIDDEN(_mesa_x86_cpuid_eax) -GLNAME(_mesa_x86_cpuid_eax): - - MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ - PUSH_L (EBX) - - CPUID - - POP_L (EBX) - RET - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_x86_cpuid_ebx) -HIDDEN(_mesa_x86_cpuid_ebx) -GLNAME(_mesa_x86_cpuid_ebx): - - MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ - PUSH_L (EBX) - - CPUID - MOV_L (EBX, EAX) /* return EBX */ - - POP_L (EBX) - RET - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_x86_cpuid_ecx) -HIDDEN(_mesa_x86_cpuid_ecx) -GLNAME(_mesa_x86_cpuid_ecx): - - MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ - PUSH_L (EBX) - - CPUID - MOV_L (ECX, EAX) /* return ECX */ - - POP_L (EBX) - RET - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_x86_cpuid_edx) -HIDDEN(_mesa_x86_cpuid_edx) -GLNAME(_mesa_x86_cpuid_edx): - - MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ - PUSH_L (EBX) - - CPUID - MOV_L (EDX, EAX) /* return EDX */ - - POP_L (EBX) - RET - -#ifdef USE_SSE_ASM -/* Execute an SSE instruction to see if the operating system correctly - * supports SSE. A signal handler for SIGILL should have been set - * before calling this function, otherwise this could kill the client - * application. - * - * -----> !!!! ATTENTION DEVELOPERS !!!! <----- - * - * If you're debugging with gdb and you get stopped in this function, - * just type 'continue'! Execution will proceed normally. - * See freedesktop.org bug #1709 for more info. - */ -ALIGNTEXT4 -GLOBL GLNAME( _mesa_test_os_sse_support ) -HIDDEN(_mesa_test_os_sse_support) -GLNAME( _mesa_test_os_sse_support ): - - XORPS ( XMM0, XMM0 ) - - RET - - -/* Perform an SSE divide-by-zero to see if the operating system - * correctly supports unmasked SIMD FPU exceptions. Signal handlers for - * SIGILL and SIGFPE should have been set before calling this function, - * otherwise this could kill the client application. - */ -ALIGNTEXT4 -GLOBL GLNAME( _mesa_test_os_sse_exception_support ) -HIDDEN(_mesa_test_os_sse_exception_support) -GLNAME( _mesa_test_os_sse_exception_support ): - - PUSH_L ( EBP ) - MOV_L ( ESP, EBP ) - SUB_L ( CONST( 8 ), ESP ) - - /* Save the original MXCSR register value. - */ - STMXCSR ( REGOFF( -4, EBP ) ) - - /* Unmask the divide-by-zero exception and perform one. - */ - STMXCSR ( REGOFF( -8, EBP ) ) - AND_L ( CONST( 0xfffffdff ), REGOFF( -8, EBP ) ) - LDMXCSR ( REGOFF( -8, EBP ) ) - - XORPS ( XMM0, XMM0 ) - - PUSH_L ( CONST( 0x3f800000 ) ) - PUSH_L ( CONST( 0x3f800000 ) ) - PUSH_L ( CONST( 0x3f800000 ) ) - PUSH_L ( CONST( 0x3f800000 ) ) - - MOVUPS ( REGIND( ESP ), XMM1 ) - - DIVPS ( XMM0, XMM1 ) - - /* Restore the original MXCSR register value. - */ - LDMXCSR ( REGOFF( -4, EBP ) ) - - LEAVE - RET - -#endif - - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif +/* + * Mesa 3-D graphics library + * Version: 6.3 + * + * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * Check extended CPU capabilities. Now justs returns the raw CPUID + * feature information, allowing the higher level code to interpret the + * results. + * + * Written by Holger Waechtler + * + * Cleaned up and simplified by Gareth Hughes + * + */ + +/* + * NOTE: Avoid using spaces in between '(' ')' and arguments, especially + * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces + * in there will break the build on some platforms. + */ + +#include "matypes.h" +#include "assyntax.h" +#include "common_x86_features.h" + + SEG_TEXT + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_x86_has_cpuid) +HIDDEN(_mesa_x86_has_cpuid) +GLNAME(_mesa_x86_has_cpuid): + + /* Test for the CPUID command. If the ID Flag bit in EFLAGS + * (bit 21) is writable, the CPUID command is present */ + PUSHF_L + POP_L (EAX) + MOV_L (EAX, ECX) + XOR_L (CONST(0x00200000), EAX) + PUSH_L (EAX) + POPF_L + PUSHF_L + POP_L (EAX) + + /* Verify the ID Flag bit has been written. */ + CMP_L (ECX, EAX) + SETNE (AL) + XOR_L (CONST(0xff), EAX) + + RET + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_x86_cpuid) +HIDDEN(_mesa_x86_cpuid) +GLNAME(_mesa_x86_cpuid): + + MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ + PUSH_L (EDI) + PUSH_L (EBX) + + CPUID + + MOV_L (REGOFF(16, ESP), EDI) /* *eax */ + MOV_L (EAX, REGIND(EDI)) + MOV_L (REGOFF(20, ESP), EDI) /* *ebx */ + MOV_L (EBX, REGIND(EDI)) + MOV_L (REGOFF(24, ESP), EDI) /* *ecx */ + MOV_L (ECX, REGIND(EDI)) + MOV_L (REGOFF(28, ESP), EDI) /* *edx */ + MOV_L (EDX, REGIND(EDI)) + + POP_L (EBX) + POP_L (EDI) + RET + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_x86_cpuid_eax) +HIDDEN(_mesa_x86_cpuid_eax) +GLNAME(_mesa_x86_cpuid_eax): + + MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ + PUSH_L (EBX) + + CPUID + + POP_L (EBX) + RET + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_x86_cpuid_ebx) +HIDDEN(_mesa_x86_cpuid_ebx) +GLNAME(_mesa_x86_cpuid_ebx): + + MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ + PUSH_L (EBX) + + CPUID + MOV_L (EBX, EAX) /* return EBX */ + + POP_L (EBX) + RET + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_x86_cpuid_ecx) +HIDDEN(_mesa_x86_cpuid_ecx) +GLNAME(_mesa_x86_cpuid_ecx): + + MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ + PUSH_L (EBX) + + CPUID + MOV_L (ECX, EAX) /* return ECX */ + + POP_L (EBX) + RET + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_x86_cpuid_edx) +HIDDEN(_mesa_x86_cpuid_edx) +GLNAME(_mesa_x86_cpuid_edx): + + MOV_L (REGOFF(4, ESP), EAX) /* cpuid op */ + PUSH_L (EBX) + + CPUID + MOV_L (EDX, EAX) /* return EDX */ + + POP_L (EBX) + RET + +#ifdef USE_SSE_ASM +/* Execute an SSE instruction to see if the operating system correctly + * supports SSE. A signal handler for SIGILL should have been set + * before calling this function, otherwise this could kill the client + * application. + * + * -----> !!!! ATTENTION DEVELOPERS !!!! <----- + * + * If you're debugging with gdb and you get stopped in this function, + * just type 'continue'! Execution will proceed normally. + * See freedesktop.org bug #1709 for more info. + */ +ALIGNTEXT4 +GLOBL GLNAME( _mesa_test_os_sse_support ) +HIDDEN(_mesa_test_os_sse_support) +GLNAME( _mesa_test_os_sse_support ): + + XORPS ( XMM0, XMM0 ) + + RET + + +/* Perform an SSE divide-by-zero to see if the operating system + * correctly supports unmasked SIMD FPU exceptions. Signal handlers for + * SIGILL and SIGFPE should have been set before calling this function, + * otherwise this could kill the client application. + */ +ALIGNTEXT4 +GLOBL GLNAME( _mesa_test_os_sse_exception_support ) +HIDDEN(_mesa_test_os_sse_exception_support) +GLNAME( _mesa_test_os_sse_exception_support ): + + PUSH_L ( EBP ) + MOV_L ( ESP, EBP ) + SUB_L ( CONST( 8 ), ESP ) + + /* Save the original MXCSR register value. + */ + STMXCSR ( REGOFF( -4, EBP ) ) + + /* Unmask the divide-by-zero exception and perform one. + */ + STMXCSR ( REGOFF( -8, EBP ) ) + AND_L ( CONST( 0xfffffdff ), REGOFF( -8, EBP ) ) + LDMXCSR ( REGOFF( -8, EBP ) ) + + XORPS ( XMM0, XMM0 ) + + PUSH_L ( CONST( 0x3f800000 ) ) + PUSH_L ( CONST( 0x3f800000 ) ) + PUSH_L ( CONST( 0x3f800000 ) ) + PUSH_L ( CONST( 0x3f800000 ) ) + + MOVUPS ( REGIND( ESP ), XMM1 ) + + DIVPS ( XMM0, XMM1 ) + + /* Restore the original MXCSR register value. + */ + LDMXCSR ( REGOFF( -4, EBP ) ) + + LEAVE + RET + +#endif + + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/common_x86_asm.h b/mesalib/src/mesa/x86/common_x86_asm.h index efda4a0c8..0d39e3d23 100644 --- a/mesalib/src/mesa/x86/common_x86_asm.h +++ b/mesalib/src/mesa/x86/common_x86_asm.h @@ -1,53 +1,53 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * Check CPU capabilities & initialize optimized funtions for this particular - * processor. - * - * Written by Holger Waechtler - * Changed by Andre Werthmann for using the - * new SSE functions - * - * Reimplemented by Gareth Hughes in a more - * future-proof manner, based on code in the Linux kernel. - */ - -#ifndef __COMMON_X86_ASM_H__ -#define __COMMON_X86_ASM_H__ - -/* Do not reference mtypes.h from this file. - */ -#include "common_x86_features.h" - -extern int _mesa_x86_cpu_features; - -extern void _mesa_get_x86_features(void); - -extern void _mesa_check_os_sse_support(void); - -extern void _mesa_init_all_x86_transform_asm( void ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * Check CPU capabilities & initialize optimized funtions for this particular + * processor. + * + * Written by Holger Waechtler + * Changed by Andre Werthmann for using the + * new SSE functions + * + * Reimplemented by Gareth Hughes in a more + * future-proof manner, based on code in the Linux kernel. + */ + +#ifndef __COMMON_X86_ASM_H__ +#define __COMMON_X86_ASM_H__ + +/* Do not reference mtypes.h from this file. + */ +#include "common_x86_features.h" + +extern int _mesa_x86_cpu_features; + +extern void _mesa_get_x86_features(void); + +extern void _mesa_check_os_sse_support(void); + +extern void _mesa_init_all_x86_transform_asm( void ); + +#endif diff --git a/mesalib/src/mesa/x86/common_x86_features.h b/mesalib/src/mesa/x86/common_x86_features.h index 2482d00ab..676af8c1f 100644 --- a/mesalib/src/mesa/x86/common_x86_features.h +++ b/mesalib/src/mesa/x86/common_x86_features.h @@ -1,67 +1,67 @@ - -/* - * Mesa 3-D graphics library - * Version: 5.1 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * x86 CPUID feature information. The raw data is returned by - * _mesa_identify_x86_cpu_features() and interpreted with the cpu_has_* - * helper macros. - * - * Gareth Hughes - */ - -#ifndef __COMMON_X86_FEATURES_H__ -#define __COMMON_X86_FEATURES_H__ - -#define X86_FEATURE_FPU (1<<0) -#define X86_FEATURE_CMOV (1<<1) -#define X86_FEATURE_MMXEXT (1<<2) -#define X86_FEATURE_MMX (1<<3) -#define X86_FEATURE_FXSR (1<<4) -#define X86_FEATURE_XMM (1<<5) -#define X86_FEATURE_XMM2 (1<<6) -#define X86_FEATURE_3DNOWEXT (1<<7) -#define X86_FEATURE_3DNOW (1<<8) - -/* standard X86 CPU features */ -#define X86_CPU_FPU (1<<0) -#define X86_CPU_CMOV (1<<15) -#define X86_CPU_MMX (1<<23) -#define X86_CPU_XMM (1<<25) -#define X86_CPU_XMM2 (1<<26) - -/* extended X86 CPU features */ -#define X86_CPUEXT_MMX_EXT (1<<22) -#define X86_CPUEXT_3DNOW_EXT (1<<30) -#define X86_CPUEXT_3DNOW (1<<31) - -#define cpu_has_mmx (_mesa_x86_cpu_features & X86_FEATURE_MMX) -#define cpu_has_mmxext (_mesa_x86_cpu_features & X86_FEATURE_MMXEXT) -#define cpu_has_xmm (_mesa_x86_cpu_features & X86_FEATURE_XMM) -#define cpu_has_xmm2 (_mesa_x86_cpu_features & X86_FEATURE_XMM2) -#define cpu_has_3dnow (_mesa_x86_cpu_features & X86_FEATURE_3DNOW) -#define cpu_has_3dnowext (_mesa_x86_cpu_features & X86_FEATURE_3DNOWEXT) - -#endif - + +/* + * Mesa 3-D graphics library + * Version: 5.1 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * x86 CPUID feature information. The raw data is returned by + * _mesa_identify_x86_cpu_features() and interpreted with the cpu_has_* + * helper macros. + * + * Gareth Hughes + */ + +#ifndef __COMMON_X86_FEATURES_H__ +#define __COMMON_X86_FEATURES_H__ + +#define X86_FEATURE_FPU (1<<0) +#define X86_FEATURE_CMOV (1<<1) +#define X86_FEATURE_MMXEXT (1<<2) +#define X86_FEATURE_MMX (1<<3) +#define X86_FEATURE_FXSR (1<<4) +#define X86_FEATURE_XMM (1<<5) +#define X86_FEATURE_XMM2 (1<<6) +#define X86_FEATURE_3DNOWEXT (1<<7) +#define X86_FEATURE_3DNOW (1<<8) + +/* standard X86 CPU features */ +#define X86_CPU_FPU (1<<0) +#define X86_CPU_CMOV (1<<15) +#define X86_CPU_MMX (1<<23) +#define X86_CPU_XMM (1<<25) +#define X86_CPU_XMM2 (1<<26) + +/* extended X86 CPU features */ +#define X86_CPUEXT_MMX_EXT (1<<22) +#define X86_CPUEXT_3DNOW_EXT (1<<30) +#define X86_CPUEXT_3DNOW (1<<31) + +#define cpu_has_mmx (_mesa_x86_cpu_features & X86_FEATURE_MMX) +#define cpu_has_mmxext (_mesa_x86_cpu_features & X86_FEATURE_MMXEXT) +#define cpu_has_xmm (_mesa_x86_cpu_features & X86_FEATURE_XMM) +#define cpu_has_xmm2 (_mesa_x86_cpu_features & X86_FEATURE_XMM2) +#define cpu_has_3dnow (_mesa_x86_cpu_features & X86_FEATURE_3DNOW) +#define cpu_has_3dnowext (_mesa_x86_cpu_features & X86_FEATURE_3DNOWEXT) + +#endif + diff --git a/mesalib/src/mesa/x86/mmx.h b/mesalib/src/mesa/x86/mmx.h index 28a43d18d..74e9979d3 100644 --- a/mesalib/src/mesa/x86/mmx.h +++ b/mesalib/src/mesa/x86/mmx.h @@ -1,59 +1,59 @@ -/* - * Mesa 3-D graphics library - * Version: 6.5.2 - * - * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - - -#ifndef ASM_MMX_H -#define ASM_MMX_H - -#include "main/compiler.h" -#include "main/glheader.h" - -struct gl_context; - -extern void _ASMAPI -_mesa_mmx_blend_transparency( struct gl_context *ctx, GLuint n, const GLubyte mask[], - GLvoid *rgba, const GLvoid *dest, - GLenum chanType ); - -extern void _ASMAPI -_mesa_mmx_blend_add( struct gl_context *ctx, GLuint n, const GLubyte mask[], - GLvoid *rgba, const GLvoid *dest, - GLenum chanType ); - -extern void _ASMAPI -_mesa_mmx_blend_min( struct gl_context *ctx, GLuint n, const GLubyte mask[], - GLvoid *rgba, const GLvoid *dest, - GLenum chanType ); - -extern void _ASMAPI -_mesa_mmx_blend_max( struct gl_context *ctx, GLuint n, const GLubyte mask[], - GLvoid *rgba, const GLvoid *dest, - GLenum chanType ); - -extern void _ASMAPI -_mesa_mmx_blend_modulate( struct gl_context *ctx, GLuint n, const GLubyte mask[], - GLvoid *rgba, const GLvoid *dest, - GLenum chanType ); - -#endif +/* + * Mesa 3-D graphics library + * Version: 6.5.2 + * + * Copyright (C) 1999-2006 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + + +#ifndef ASM_MMX_H +#define ASM_MMX_H + +#include "main/compiler.h" +#include "main/glheader.h" + +struct gl_context; + +extern void _ASMAPI +_mesa_mmx_blend_transparency( struct gl_context *ctx, GLuint n, const GLubyte mask[], + GLvoid *rgba, const GLvoid *dest, + GLenum chanType ); + +extern void _ASMAPI +_mesa_mmx_blend_add( struct gl_context *ctx, GLuint n, const GLubyte mask[], + GLvoid *rgba, const GLvoid *dest, + GLenum chanType ); + +extern void _ASMAPI +_mesa_mmx_blend_min( struct gl_context *ctx, GLuint n, const GLubyte mask[], + GLvoid *rgba, const GLvoid *dest, + GLenum chanType ); + +extern void _ASMAPI +_mesa_mmx_blend_max( struct gl_context *ctx, GLuint n, const GLubyte mask[], + GLvoid *rgba, const GLvoid *dest, + GLenum chanType ); + +extern void _ASMAPI +_mesa_mmx_blend_modulate( struct gl_context *ctx, GLuint n, const GLubyte mask[], + GLvoid *rgba, const GLvoid *dest, + GLenum chanType ); + +#endif diff --git a/mesalib/src/mesa/x86/mmx_blend.S b/mesalib/src/mesa/x86/mmx_blend.S index 1e3954730..eeaf43ea9 100644 --- a/mesalib/src/mesa/x86/mmx_blend.S +++ b/mesalib/src/mesa/x86/mmx_blend.S @@ -1,402 +1,402 @@ - ; -/* - * Written by Jos� Fonseca - */ - - -#ifdef USE_MMX_ASM -#include "assyntax.h" -#include "matypes.h" - -/* integer multiplication - alpha plus one - * - * makes the following approximation to the division (Sree) - * - * rgb*a/255 ~= (rgb*(a+1)) >> 256 - * - * which is the fastest method that satisfies the following OpenGL criteria - * - * 0*0 = 0 and 255*255 = 255 - * - * note that MX1 is a register with 0xffffffffffffffff constant which can be easily obtained making - * - * PCMPEQW ( MX1, MX1 ) - */ -#define GMB_MULT_AP1( MP1, MA1, MP2, MA2, MX1 ) \ - PSUBW ( MX1, MA1 ) /* a1 + 1 | a1 + 1 | a1 + 1 | a1 + 1 */ ;\ - PMULLW ( MP1, MA1 ) /* t1 = p1*a1 */ ;\ - ;\ -TWO(PSUBW ( MX1, MA2 )) /* a2 + 1 | a2 + 1 | a2 + 1 | a2 + 1 */ ;\ -TWO(PMULLW ( MP2, MA2 )) /* t2 = p2*a2 */ ;\ - ;\ - PSRLW ( CONST(8), MA1 ) /* t1 >> 8 ~= t1/255 */ ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 ~= t2/255 */ - - -/* integer multiplication - geometric series - * - * takes the geometric series approximation to the division - * - * t/255 = (t >> 8) + (t >> 16) + (t >> 24) .. - * - * in this case just the first two terms to fit in 16bit arithmetic - * - * t/255 ~= (t + (t >> 8)) >> 8 - * - * note that just by itself it doesn't satisfies the OpenGL criteria, as 255*255 = 254, - * so the special case a = 255 must be accounted or roundoff must be used - */ -#define GMB_MULT_GS( MP1, MA1, MP2, MA2 ) \ - PMULLW ( MP1, MA1 ) /* t1 = p1*a1 */ ;\ -TWO(PMULLW ( MP2, MA2 )) /* t2 = p2*a2 */ ;\ - ;\ - MOVQ ( MA1, MP1 ) ;\ - PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ - ;\ -TWO(MOVQ ( MA2, MP2 )) ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ - ;\ - PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ - PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ - ;\ -TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ - - -/* integer multiplication - geometric series plus rounding - * - * when using a geometric series division instead of truncating the result - * use roundoff in the approximation (Jim Blinn) - * - * t = rgb*a + 0x80 - * - * achieving the exact results - * - * note that M80 is register with the 0x0080008000800080 constant - */ -#define GMB_MULT_GSR( MP1, MA1, MP2, MA2, M80 ) \ - PMULLW ( MP1, MA1 ) /* t1 = p1*a1 */ ;\ - PADDW ( M80, MA1 ) /* t1 += 0x80 */ ;\ - ;\ -TWO(PMULLW ( MP2, MA2 )) /* t2 = p2*a2 */ ;\ -TWO(PADDW ( M80, MA2 )) /* t2 += 0x80 */ ;\ - ;\ - MOVQ ( MA1, MP1 ) ;\ - PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ - ;\ -TWO(MOVQ ( MA2, MP2 )) ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ - ;\ - PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ - PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ - ;\ -TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ - - -/* linear interpolation - geometric series - */ -#define GMB_LERP_GS( MP1, MQ1, MA1, MP2, MQ2, MA2) \ - PSUBW ( MQ1, MP1 ) /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */ ;\ - PSLLW ( CONST(8), MQ1 ) /* q1 << 8 */ ;\ - PMULLW ( MP1, MA1 ) /* t1 = (q1 - p1)*pa1 */ ;\ - ;\ -TWO(PSUBW ( MQ2, MP2 )) /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */ ;\ -TWO(PSLLW ( CONST(8), MQ2 )) /* q2 << 8 */ ;\ -TWO(PMULLW ( MP2, MA2 )) /* t2 = (q2 - p2)*pa2 */ ;\ - ;\ - MOVQ ( MA1, MP1 ) ;\ - PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ - ;\ -TWO(MOVQ ( MA2, MP2 )) ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ - ;\ - PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ -TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ - ;\ - PADDW ( MQ1, MA1 ) /* (t1/255 + q1) << 8 */ ;\ -TWO(PADDW ( MQ2, MA2 )) /* (t2/255 + q2) << 8 */ ;\ - ;\ - PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ - - -/* linear interpolation - geometric series with roundoff - * - * this is a generalization of Blinn's formula to signed arithmetic - * - * note that M80 is a register with the 0x0080008000800080 constant - */ -#define GMB_LERP_GSR( MP1, MQ1, MA1, MP2, MQ2, MA2, M80) \ - PSUBW ( MQ1, MP1 ) /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */ ;\ - PSLLW ( CONST(8), MQ1 ) /* q1 << 8 */ ;\ - PMULLW ( MP1, MA1 ) /* t1 = (q1 - p1)*pa1 */ ;\ - ;\ -TWO(PSUBW ( MQ2, MP2 )) /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */ ;\ -TWO(PSLLW ( CONST(8), MQ2 )) /* q2 << 8 */ ;\ -TWO(PMULLW ( MP2, MA2 )) /* t2 = (q2 - p2)*pa2 */ ;\ - ;\ - PSRLW ( CONST(15), MP1 ) /* q1 > p1 ? 1 : 0 */ ;\ -TWO(PSRLW ( CONST(15), MP2 )) /* q2 > q2 ? 1 : 0 */ ;\ - ;\ - PSLLW ( CONST(8), MP1 ) /* q1 > p1 ? 0x100 : 0 */ ;\ -TWO(PSLLW ( CONST(8), MP2 )) /* q2 > q2 ? 0x100 : 0 */ ;\ - ;\ - PSUBW ( MP1, MA1 ) /* t1 -=? 0x100 */ ;\ -TWO(PSUBW ( MP2, MA2 )) /* t2 -=? 0x100 */ ;\ - ;\ - PADDW ( M80, MA1 ) /* t1 += 0x80 */ ;\ -TWO(PADDW ( M80, MA2 )) /* t2 += 0x80 */ ;\ - ;\ - MOVQ ( MA1, MP1 ) ;\ - PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ - ;\ -TWO(MOVQ ( MA2, MP2 )) ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ - ;\ - PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ -TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ - ;\ - PADDW ( MQ1, MA1 ) /* (t1/255 + q1) << 8 */ ;\ -TWO(PADDW ( MQ2, MA2 )) /* (t2/255 + q2) << 8 */ ;\ - ;\ - PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ - - -/* linear interpolation - geometric series with correction - * - * instead of the roundoff this adds a small correction to satisfy the OpenGL criteria - * - * t/255 ~= (t + (t >> 8) + (t >> 15)) >> 8 - * - * note that although is faster than rounding off it doesn't give always the exact results - */ -#define GMB_LERP_GSC( MP1, MQ1, MA1, MP2, MQ2, MA2) \ - PSUBW ( MQ1, MP1 ) /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */ ;\ - PSLLW ( CONST(8), MQ1 ) /* q1 << 8 */ ;\ - PMULLW ( MP1, MA1 ) /* t1 = (q1 - p1)*pa1 */ ;\ - ;\ -TWO(PSUBW ( MQ2, MP2 )) /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */ ;\ -TWO(PSLLW ( CONST(8), MQ2 )) /* q2 << 8 */ ;\ -TWO(PMULLW ( MP2, MA2 )) /* t2 = (q2 - p2)*pa2 */ ;\ - ;\ - MOVQ ( MA1, MP1 ) ;\ - PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ - ;\ -TWO(MOVQ ( MA2, MP2 )) ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ - ;\ - PADDW ( MA1, MP1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ - PSRLW ( CONST(7), MA1 ) /* t1 >> 15 */ ;\ - ;\ -TWO(PADDW ( MA2, MP2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ -TWO(PSRLW ( CONST(7), MA2 )) /* t2 >> 15 */ ;\ - ;\ - PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) + (t1 >>15) ~= (t1/255) << 8 */ ;\ -TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) + (t2 >>15) ~= (t2/255) << 8 */ ;\ - ;\ - PADDW ( MQ1, MA1 ) /* (t1/255 + q1) << 8 */ ;\ -TWO(PADDW ( MQ2, MA2 )) /* (t2/255 + q2) << 8 */ ;\ - ;\ - PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ -TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ - - -/* common blending setup code - * - * note that M00 is a register with 0x0000000000000000 constant which can be easily obtained making - * - * PXOR ( M00, M00 ) - */ -#define GMB_LOAD(rgba, dest, MPP, MQQ) \ -ONE(MOVD ( REGIND(rgba), MPP )) /* | | | | qa1 | qb1 | qg1 | qr1 */ ;\ -ONE(MOVD ( REGIND(dest), MQQ )) /* | | | | pa1 | pb1 | pg1 | pr1 */ ;\ - ;\ -TWO(MOVQ ( REGIND(rgba), MPP )) /* qa2 | qb2 | qg2 | qr2 | qa1 | qb1 | qg1 | qr1 */ ;\ -TWO(MOVQ ( REGIND(dest), MQQ )) /* pa2 | pb2 | pg2 | pr2 | pa1 | pb1 | pg1 | pr1 */ - -#define GMB_UNPACK(MP1, MQ1, MP2, MQ2, M00) \ -TWO(MOVQ ( MP1, MP2 )) ;\ -TWO(MOVQ ( MQ1, MQ2 )) ;\ - ;\ - PUNPCKLBW ( M00, MQ1 ) /* qa1 | qb1 | qg1 | qr1 */ ;\ -TWO(PUNPCKHBW ( M00, MQ2 )) /* qa2 | qb2 | qg2 | qr2 */ ;\ - PUNPCKLBW ( M00, MP1 ) /* pa1 | pb1 | pg1 | pr1 */ ;\ -TWO(PUNPCKHBW ( M00, MP2 )) /* pa2 | pb2 | pg2 | pr2 */ - -#define GMB_ALPHA(MP1, MA1, MP2, MA2) \ - MOVQ ( MP1, MA1 ) ;\ -TWO(MOVQ ( MP2, MA2 )) ;\ - ;\ - PUNPCKHWD ( MA1, MA1 ) /* pa1 | pa1 | | */ ;\ -TWO(PUNPCKHWD ( MA2, MA2 )) /* pa2 | pa2 | | */ ;\ - PUNPCKHDQ ( MA1, MA1 ) /* pa1 | pa1 | pa1 | pa1 */ ;\ -TWO(PUNPCKHDQ ( MA2, MA2 )) /* pa2 | pa2 | pa2 | pa2 */ - -#define GMB_PACK( MS1, MS2 ) \ - PACKUSWB ( MS2, MS1 ) /* sa2 | sb2 | sg2 | sr2 | sa1 | sb1 | sg1 | sr1 */ ;\ - -#define GMB_STORE(rgba, MSS ) \ -ONE(MOVD ( MSS, REGIND(rgba) )) /* | | | | sa1 | sb1 | sg1 | sr1 */ ;\ -TWO(MOVQ ( MSS, REGIND(rgba) )) /* sa2 | sb2 | sg2 | sr2 | sa1 | sb1 | sg1 | sr1 */ - -/* Kevin F. Quinn 2 July 2006 - * Replace data segment constants with text-segment - * constants (via pushl/movq) - SEG_DATA - -ALIGNDATA8 -const_0080: - D_LONG 0x00800080, 0x00800080 - -const_80: - D_LONG 0x80808080, 0x80808080 -*/ -#define const_0080_l 0x00800080 -#define const_0080_h 0x00800080 -#define const_80_l 0x80808080 -#define const_80_h 0x80808080 - - SEG_TEXT - - -/* Blend transparency function - */ - -#define TAG(x) CONCAT(x,_transparency) -#define LLTAG(x) LLBL2(x,_transparency) - -#define INIT \ - PXOR ( MM0, MM0 ) /* 0x0000 | 0x0000 | 0x0000 | 0x0000 */ - -#define MAIN( rgba, dest ) \ - GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ - GMB_UNPACK( MM1, MM2, MM4, MM5, MM0 ) ;\ - GMB_ALPHA( MM1, MM3, MM4, MM6 ) ;\ - GMB_LERP_GSC( MM1, MM2, MM3, MM4, MM5, MM6 ) ;\ - GMB_PACK( MM3, MM6 ) ;\ - GMB_STORE( rgba, MM3 ) - -#include "mmx_blendtmp.h" - - -/* Blend add function - * - * FIXME: Add some loop unrolling here... - */ - -#define TAG(x) CONCAT(x,_add) -#define LLTAG(x) LLBL2(x,_add) - -#define INIT - -#define MAIN( rgba, dest ) \ -ONE(MOVD ( REGIND(rgba), MM1 )) /* | | | | qa1 | qb1 | qg1 | qr1 */ ;\ -ONE(MOVD ( REGIND(dest), MM2 )) /* | | | | pa1 | pb1 | pg1 | pr1 */ ;\ -ONE(PADDUSB ( MM2, MM1 )) ;\ -ONE(MOVD ( MM1, REGIND(rgba) )) /* | | | | sa1 | sb1 | sg1 | sr1 */ ;\ - ;\ -TWO(MOVQ ( REGIND(rgba), MM1 )) /* qa2 | qb2 | qg2 | qr2 | qa1 | qb1 | qg1 | qr1 */ ;\ -TWO(PADDUSB ( REGIND(dest), MM1 )) /* sa2 | sb2 | sg2 | sr2 | sa1 | sb1 | sg1 | sr1 */ ;\ -TWO(MOVQ ( MM1, REGIND(rgba) )) - -#include "mmx_blendtmp.h" - - -/* Blend min function - */ - -#define TAG(x) CONCAT(x,_min) -#define LLTAG(x) LLBL2(x,_min) - -/* Kevin F. Quinn 2nd July 2006 - * Replace data segment constants with text-segment instructions -#define INIT \ - MOVQ ( CONTENT(const_80), MM7 ) - */ -#define INIT \ - PUSH_L ( CONST(const_80_h) ) /* 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80*/ ;\ - PUSH_L ( CONST(const_80_l) ) ;\ - MOVQ ( REGIND(ESP), MM7 ) ;\ - ADD_L ( CONST(8), ESP) - -#define MAIN( rgba, dest ) \ - GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ - MOVQ ( MM1, MM3 ) ;\ - MOVQ ( MM2, MM4 ) ;\ - PXOR ( MM7, MM3 ) /* unsigned -> signed */ ;\ - PXOR ( MM7, MM4 ) /* unsigned -> signed */ ;\ - PCMPGTB ( MM3, MM4 ) /* q > p ? 0xff : 0x00 */ ;\ - PAND ( MM4, MM1 ) /* q > p ? p : 0 */ ;\ - PANDN ( MM2, MM4 ) /* q > p ? 0 : q */ ;\ - POR ( MM1, MM4 ) /* q > p ? p : q */ ;\ - GMB_STORE( rgba, MM4 ) - -#include "mmx_blendtmp.h" - - -/* Blend max function - */ - -#define TAG(x) CONCAT(x,_max) -#define LLTAG(x) LLBL2(x,_max) - -/* Kevin F. Quinn 2nd July 2006 - * Replace data segment constants with text-segment instructions -#define INIT \ - MOVQ ( CONTENT(const_80), MM7 ) - */ -#define INIT \ - PUSH_L ( CONST(const_80_l) ) /* 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80*/ ;\ - PUSH_L ( CONST(const_80_h) ) ;\ - MOVQ ( REGIND(ESP), MM7 ) ;\ - ADD_L ( CONST(8), ESP) - -#define MAIN( rgba, dest ) \ - GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ - MOVQ ( MM1, MM3 ) ;\ - MOVQ ( MM2, MM4 ) ;\ - PXOR ( MM7, MM3 ) /* unsigned -> signed */ ;\ - PXOR ( MM7, MM4 ) /* unsigned -> signed */ ;\ - PCMPGTB ( MM3, MM4 ) /* q > p ? 0xff : 0x00 */ ;\ - PAND ( MM4, MM2 ) /* q > p ? q : 0 */ ;\ - PANDN ( MM1, MM4 ) /* q > p ? 0 : p */ ;\ - POR ( MM2, MM4 ) /* q > p ? p : q */ ;\ - GMB_STORE( rgba, MM4 ) - -#include "mmx_blendtmp.h" - - -/* Blend modulate function - */ - -#define TAG(x) CONCAT(x,_modulate) -#define LLTAG(x) LLBL2(x,_modulate) - -/* Kevin F. Quinn 2nd July 2006 - * Replace data segment constants with text-segment instructions -#define INIT \ - MOVQ ( CONTENT(const_0080), MM7 ) - */ -#define INIT \ - PXOR ( MM0, MM0 ) /* 0x0000 | 0x0000 | 0x0000 | 0x0000 */ ;\ - PUSH_L ( CONST(const_0080_l) ) /* 0x0080 | 0x0080 | 0x0080 | 0x0080 */ ;\ - PUSH_L ( CONST(const_0080_h) ) ;\ - MOVQ ( REGIND(ESP), MM7 ) ;\ - ADD_L ( CONST(8), ESP) - -#define MAIN( rgba, dest ) \ - GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ - GMB_UNPACK( MM1, MM2, MM4, MM5, MM0 ) ;\ - GMB_MULT_GSR( MM1, MM2, MM4, MM5, MM7 ) ;\ - GMB_PACK( MM2, MM5 ) ;\ - GMB_STORE( rgba, MM2 ) - -#include "mmx_blendtmp.h" - -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + ; +/* + * Written by Jos� Fonseca + */ + + +#ifdef USE_MMX_ASM +#include "assyntax.h" +#include "matypes.h" + +/* integer multiplication - alpha plus one + * + * makes the following approximation to the division (Sree) + * + * rgb*a/255 ~= (rgb*(a+1)) >> 256 + * + * which is the fastest method that satisfies the following OpenGL criteria + * + * 0*0 = 0 and 255*255 = 255 + * + * note that MX1 is a register with 0xffffffffffffffff constant which can be easily obtained making + * + * PCMPEQW ( MX1, MX1 ) + */ +#define GMB_MULT_AP1( MP1, MA1, MP2, MA2, MX1 ) \ + PSUBW ( MX1, MA1 ) /* a1 + 1 | a1 + 1 | a1 + 1 | a1 + 1 */ ;\ + PMULLW ( MP1, MA1 ) /* t1 = p1*a1 */ ;\ + ;\ +TWO(PSUBW ( MX1, MA2 )) /* a2 + 1 | a2 + 1 | a2 + 1 | a2 + 1 */ ;\ +TWO(PMULLW ( MP2, MA2 )) /* t2 = p2*a2 */ ;\ + ;\ + PSRLW ( CONST(8), MA1 ) /* t1 >> 8 ~= t1/255 */ ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 ~= t2/255 */ + + +/* integer multiplication - geometric series + * + * takes the geometric series approximation to the division + * + * t/255 = (t >> 8) + (t >> 16) + (t >> 24) .. + * + * in this case just the first two terms to fit in 16bit arithmetic + * + * t/255 ~= (t + (t >> 8)) >> 8 + * + * note that just by itself it doesn't satisfies the OpenGL criteria, as 255*255 = 254, + * so the special case a = 255 must be accounted or roundoff must be used + */ +#define GMB_MULT_GS( MP1, MA1, MP2, MA2 ) \ + PMULLW ( MP1, MA1 ) /* t1 = p1*a1 */ ;\ +TWO(PMULLW ( MP2, MA2 )) /* t2 = p2*a2 */ ;\ + ;\ + MOVQ ( MA1, MP1 ) ;\ + PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ + ;\ +TWO(MOVQ ( MA2, MP2 )) ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ + ;\ + PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ + PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ + ;\ +TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ + + +/* integer multiplication - geometric series plus rounding + * + * when using a geometric series division instead of truncating the result + * use roundoff in the approximation (Jim Blinn) + * + * t = rgb*a + 0x80 + * + * achieving the exact results + * + * note that M80 is register with the 0x0080008000800080 constant + */ +#define GMB_MULT_GSR( MP1, MA1, MP2, MA2, M80 ) \ + PMULLW ( MP1, MA1 ) /* t1 = p1*a1 */ ;\ + PADDW ( M80, MA1 ) /* t1 += 0x80 */ ;\ + ;\ +TWO(PMULLW ( MP2, MA2 )) /* t2 = p2*a2 */ ;\ +TWO(PADDW ( M80, MA2 )) /* t2 += 0x80 */ ;\ + ;\ + MOVQ ( MA1, MP1 ) ;\ + PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ + ;\ +TWO(MOVQ ( MA2, MP2 )) ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ + ;\ + PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ + PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ + ;\ +TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ + + +/* linear interpolation - geometric series + */ +#define GMB_LERP_GS( MP1, MQ1, MA1, MP2, MQ2, MA2) \ + PSUBW ( MQ1, MP1 ) /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */ ;\ + PSLLW ( CONST(8), MQ1 ) /* q1 << 8 */ ;\ + PMULLW ( MP1, MA1 ) /* t1 = (q1 - p1)*pa1 */ ;\ + ;\ +TWO(PSUBW ( MQ2, MP2 )) /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */ ;\ +TWO(PSLLW ( CONST(8), MQ2 )) /* q2 << 8 */ ;\ +TWO(PMULLW ( MP2, MA2 )) /* t2 = (q2 - p2)*pa2 */ ;\ + ;\ + MOVQ ( MA1, MP1 ) ;\ + PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ + ;\ +TWO(MOVQ ( MA2, MP2 )) ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ + ;\ + PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ +TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ + ;\ + PADDW ( MQ1, MA1 ) /* (t1/255 + q1) << 8 */ ;\ +TWO(PADDW ( MQ2, MA2 )) /* (t2/255 + q2) << 8 */ ;\ + ;\ + PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ + + +/* linear interpolation - geometric series with roundoff + * + * this is a generalization of Blinn's formula to signed arithmetic + * + * note that M80 is a register with the 0x0080008000800080 constant + */ +#define GMB_LERP_GSR( MP1, MQ1, MA1, MP2, MQ2, MA2, M80) \ + PSUBW ( MQ1, MP1 ) /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */ ;\ + PSLLW ( CONST(8), MQ1 ) /* q1 << 8 */ ;\ + PMULLW ( MP1, MA1 ) /* t1 = (q1 - p1)*pa1 */ ;\ + ;\ +TWO(PSUBW ( MQ2, MP2 )) /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */ ;\ +TWO(PSLLW ( CONST(8), MQ2 )) /* q2 << 8 */ ;\ +TWO(PMULLW ( MP2, MA2 )) /* t2 = (q2 - p2)*pa2 */ ;\ + ;\ + PSRLW ( CONST(15), MP1 ) /* q1 > p1 ? 1 : 0 */ ;\ +TWO(PSRLW ( CONST(15), MP2 )) /* q2 > q2 ? 1 : 0 */ ;\ + ;\ + PSLLW ( CONST(8), MP1 ) /* q1 > p1 ? 0x100 : 0 */ ;\ +TWO(PSLLW ( CONST(8), MP2 )) /* q2 > q2 ? 0x100 : 0 */ ;\ + ;\ + PSUBW ( MP1, MA1 ) /* t1 -=? 0x100 */ ;\ +TWO(PSUBW ( MP2, MA2 )) /* t2 -=? 0x100 */ ;\ + ;\ + PADDW ( M80, MA1 ) /* t1 += 0x80 */ ;\ +TWO(PADDW ( M80, MA2 )) /* t2 += 0x80 */ ;\ + ;\ + MOVQ ( MA1, MP1 ) ;\ + PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ + ;\ +TWO(MOVQ ( MA2, MP2 )) ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ + ;\ + PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ +TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ + ;\ + PADDW ( MQ1, MA1 ) /* (t1/255 + q1) << 8 */ ;\ +TWO(PADDW ( MQ2, MA2 )) /* (t2/255 + q2) << 8 */ ;\ + ;\ + PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ + + +/* linear interpolation - geometric series with correction + * + * instead of the roundoff this adds a small correction to satisfy the OpenGL criteria + * + * t/255 ~= (t + (t >> 8) + (t >> 15)) >> 8 + * + * note that although is faster than rounding off it doesn't give always the exact results + */ +#define GMB_LERP_GSC( MP1, MQ1, MA1, MP2, MQ2, MA2) \ + PSUBW ( MQ1, MP1 ) /* pa1 - qa1 | pb1 - qb1 | pg1 - qg1 | pr1 - qr1 */ ;\ + PSLLW ( CONST(8), MQ1 ) /* q1 << 8 */ ;\ + PMULLW ( MP1, MA1 ) /* t1 = (q1 - p1)*pa1 */ ;\ + ;\ +TWO(PSUBW ( MQ2, MP2 )) /* pa2 - qa2 | pb2 - qb2 | pg2 - qg2 | pr2 - qr2 */ ;\ +TWO(PSLLW ( CONST(8), MQ2 )) /* q2 << 8 */ ;\ +TWO(PMULLW ( MP2, MA2 )) /* t2 = (q2 - p2)*pa2 */ ;\ + ;\ + MOVQ ( MA1, MP1 ) ;\ + PSRLW ( CONST(8), MA1 ) /* t1 >> 8 */ ;\ + ;\ +TWO(MOVQ ( MA2, MP2 )) ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* t2 >> 8 */ ;\ + ;\ + PADDW ( MA1, MP1 ) /* t1 + (t1 >> 8) ~= (t1/255) << 8 */ ;\ + PSRLW ( CONST(7), MA1 ) /* t1 >> 15 */ ;\ + ;\ +TWO(PADDW ( MA2, MP2 )) /* t2 + (t2 >> 8) ~= (t2/255) << 8 */ ;\ +TWO(PSRLW ( CONST(7), MA2 )) /* t2 >> 15 */ ;\ + ;\ + PADDW ( MP1, MA1 ) /* t1 + (t1 >> 8) + (t1 >>15) ~= (t1/255) << 8 */ ;\ +TWO(PADDW ( MP2, MA2 )) /* t2 + (t2 >> 8) + (t2 >>15) ~= (t2/255) << 8 */ ;\ + ;\ + PADDW ( MQ1, MA1 ) /* (t1/255 + q1) << 8 */ ;\ +TWO(PADDW ( MQ2, MA2 )) /* (t2/255 + q2) << 8 */ ;\ + ;\ + PSRLW ( CONST(8), MA1 ) /* sa1 | sb1 | sg1 | sr1 */ ;\ +TWO(PSRLW ( CONST(8), MA2 )) /* sa2 | sb2 | sg2 | sr2 */ + + +/* common blending setup code + * + * note that M00 is a register with 0x0000000000000000 constant which can be easily obtained making + * + * PXOR ( M00, M00 ) + */ +#define GMB_LOAD(rgba, dest, MPP, MQQ) \ +ONE(MOVD ( REGIND(rgba), MPP )) /* | | | | qa1 | qb1 | qg1 | qr1 */ ;\ +ONE(MOVD ( REGIND(dest), MQQ )) /* | | | | pa1 | pb1 | pg1 | pr1 */ ;\ + ;\ +TWO(MOVQ ( REGIND(rgba), MPP )) /* qa2 | qb2 | qg2 | qr2 | qa1 | qb1 | qg1 | qr1 */ ;\ +TWO(MOVQ ( REGIND(dest), MQQ )) /* pa2 | pb2 | pg2 | pr2 | pa1 | pb1 | pg1 | pr1 */ + +#define GMB_UNPACK(MP1, MQ1, MP2, MQ2, M00) \ +TWO(MOVQ ( MP1, MP2 )) ;\ +TWO(MOVQ ( MQ1, MQ2 )) ;\ + ;\ + PUNPCKLBW ( M00, MQ1 ) /* qa1 | qb1 | qg1 | qr1 */ ;\ +TWO(PUNPCKHBW ( M00, MQ2 )) /* qa2 | qb2 | qg2 | qr2 */ ;\ + PUNPCKLBW ( M00, MP1 ) /* pa1 | pb1 | pg1 | pr1 */ ;\ +TWO(PUNPCKHBW ( M00, MP2 )) /* pa2 | pb2 | pg2 | pr2 */ + +#define GMB_ALPHA(MP1, MA1, MP2, MA2) \ + MOVQ ( MP1, MA1 ) ;\ +TWO(MOVQ ( MP2, MA2 )) ;\ + ;\ + PUNPCKHWD ( MA1, MA1 ) /* pa1 | pa1 | | */ ;\ +TWO(PUNPCKHWD ( MA2, MA2 )) /* pa2 | pa2 | | */ ;\ + PUNPCKHDQ ( MA1, MA1 ) /* pa1 | pa1 | pa1 | pa1 */ ;\ +TWO(PUNPCKHDQ ( MA2, MA2 )) /* pa2 | pa2 | pa2 | pa2 */ + +#define GMB_PACK( MS1, MS2 ) \ + PACKUSWB ( MS2, MS1 ) /* sa2 | sb2 | sg2 | sr2 | sa1 | sb1 | sg1 | sr1 */ ;\ + +#define GMB_STORE(rgba, MSS ) \ +ONE(MOVD ( MSS, REGIND(rgba) )) /* | | | | sa1 | sb1 | sg1 | sr1 */ ;\ +TWO(MOVQ ( MSS, REGIND(rgba) )) /* sa2 | sb2 | sg2 | sr2 | sa1 | sb1 | sg1 | sr1 */ + +/* Kevin F. Quinn 2 July 2006 + * Replace data segment constants with text-segment + * constants (via pushl/movq) + SEG_DATA + +ALIGNDATA8 +const_0080: + D_LONG 0x00800080, 0x00800080 + +const_80: + D_LONG 0x80808080, 0x80808080 +*/ +#define const_0080_l 0x00800080 +#define const_0080_h 0x00800080 +#define const_80_l 0x80808080 +#define const_80_h 0x80808080 + + SEG_TEXT + + +/* Blend transparency function + */ + +#define TAG(x) CONCAT(x,_transparency) +#define LLTAG(x) LLBL2(x,_transparency) + +#define INIT \ + PXOR ( MM0, MM0 ) /* 0x0000 | 0x0000 | 0x0000 | 0x0000 */ + +#define MAIN( rgba, dest ) \ + GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ + GMB_UNPACK( MM1, MM2, MM4, MM5, MM0 ) ;\ + GMB_ALPHA( MM1, MM3, MM4, MM6 ) ;\ + GMB_LERP_GSC( MM1, MM2, MM3, MM4, MM5, MM6 ) ;\ + GMB_PACK( MM3, MM6 ) ;\ + GMB_STORE( rgba, MM3 ) + +#include "mmx_blendtmp.h" + + +/* Blend add function + * + * FIXME: Add some loop unrolling here... + */ + +#define TAG(x) CONCAT(x,_add) +#define LLTAG(x) LLBL2(x,_add) + +#define INIT + +#define MAIN( rgba, dest ) \ +ONE(MOVD ( REGIND(rgba), MM1 )) /* | | | | qa1 | qb1 | qg1 | qr1 */ ;\ +ONE(MOVD ( REGIND(dest), MM2 )) /* | | | | pa1 | pb1 | pg1 | pr1 */ ;\ +ONE(PADDUSB ( MM2, MM1 )) ;\ +ONE(MOVD ( MM1, REGIND(rgba) )) /* | | | | sa1 | sb1 | sg1 | sr1 */ ;\ + ;\ +TWO(MOVQ ( REGIND(rgba), MM1 )) /* qa2 | qb2 | qg2 | qr2 | qa1 | qb1 | qg1 | qr1 */ ;\ +TWO(PADDUSB ( REGIND(dest), MM1 )) /* sa2 | sb2 | sg2 | sr2 | sa1 | sb1 | sg1 | sr1 */ ;\ +TWO(MOVQ ( MM1, REGIND(rgba) )) + +#include "mmx_blendtmp.h" + + +/* Blend min function + */ + +#define TAG(x) CONCAT(x,_min) +#define LLTAG(x) LLBL2(x,_min) + +/* Kevin F. Quinn 2nd July 2006 + * Replace data segment constants with text-segment instructions +#define INIT \ + MOVQ ( CONTENT(const_80), MM7 ) + */ +#define INIT \ + PUSH_L ( CONST(const_80_h) ) /* 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80*/ ;\ + PUSH_L ( CONST(const_80_l) ) ;\ + MOVQ ( REGIND(ESP), MM7 ) ;\ + ADD_L ( CONST(8), ESP) + +#define MAIN( rgba, dest ) \ + GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ + MOVQ ( MM1, MM3 ) ;\ + MOVQ ( MM2, MM4 ) ;\ + PXOR ( MM7, MM3 ) /* unsigned -> signed */ ;\ + PXOR ( MM7, MM4 ) /* unsigned -> signed */ ;\ + PCMPGTB ( MM3, MM4 ) /* q > p ? 0xff : 0x00 */ ;\ + PAND ( MM4, MM1 ) /* q > p ? p : 0 */ ;\ + PANDN ( MM2, MM4 ) /* q > p ? 0 : q */ ;\ + POR ( MM1, MM4 ) /* q > p ? p : q */ ;\ + GMB_STORE( rgba, MM4 ) + +#include "mmx_blendtmp.h" + + +/* Blend max function + */ + +#define TAG(x) CONCAT(x,_max) +#define LLTAG(x) LLBL2(x,_max) + +/* Kevin F. Quinn 2nd July 2006 + * Replace data segment constants with text-segment instructions +#define INIT \ + MOVQ ( CONTENT(const_80), MM7 ) + */ +#define INIT \ + PUSH_L ( CONST(const_80_l) ) /* 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80| 0x80*/ ;\ + PUSH_L ( CONST(const_80_h) ) ;\ + MOVQ ( REGIND(ESP), MM7 ) ;\ + ADD_L ( CONST(8), ESP) + +#define MAIN( rgba, dest ) \ + GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ + MOVQ ( MM1, MM3 ) ;\ + MOVQ ( MM2, MM4 ) ;\ + PXOR ( MM7, MM3 ) /* unsigned -> signed */ ;\ + PXOR ( MM7, MM4 ) /* unsigned -> signed */ ;\ + PCMPGTB ( MM3, MM4 ) /* q > p ? 0xff : 0x00 */ ;\ + PAND ( MM4, MM2 ) /* q > p ? q : 0 */ ;\ + PANDN ( MM1, MM4 ) /* q > p ? 0 : p */ ;\ + POR ( MM2, MM4 ) /* q > p ? p : q */ ;\ + GMB_STORE( rgba, MM4 ) + +#include "mmx_blendtmp.h" + + +/* Blend modulate function + */ + +#define TAG(x) CONCAT(x,_modulate) +#define LLTAG(x) LLBL2(x,_modulate) + +/* Kevin F. Quinn 2nd July 2006 + * Replace data segment constants with text-segment instructions +#define INIT \ + MOVQ ( CONTENT(const_0080), MM7 ) + */ +#define INIT \ + PXOR ( MM0, MM0 ) /* 0x0000 | 0x0000 | 0x0000 | 0x0000 */ ;\ + PUSH_L ( CONST(const_0080_l) ) /* 0x0080 | 0x0080 | 0x0080 | 0x0080 */ ;\ + PUSH_L ( CONST(const_0080_h) ) ;\ + MOVQ ( REGIND(ESP), MM7 ) ;\ + ADD_L ( CONST(8), ESP) + +#define MAIN( rgba, dest ) \ + GMB_LOAD( rgba, dest, MM1, MM2 ) ;\ + GMB_UNPACK( MM1, MM2, MM4, MM5, MM0 ) ;\ + GMB_MULT_GSR( MM1, MM2, MM4, MM5, MM7 ) ;\ + GMB_PACK( MM2, MM5 ) ;\ + GMB_STORE( rgba, MM2 ) + +#include "mmx_blendtmp.h" + +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/mmx_blendtmp.h b/mesalib/src/mesa/x86/mmx_blendtmp.h index f282c0c21..8534792e2 100644 --- a/mesalib/src/mesa/x86/mmx_blendtmp.h +++ b/mesalib/src/mesa/x86/mmx_blendtmp.h @@ -1,114 +1,114 @@ -/* - * Written by José Fonseca - */ - - -/* - * void _mesa_mmx_blend( struct gl_context *ctx, - * GLuint n, - * const GLubyte mask[], - * GLchan rgba[][4], - * CONST GLchan dest[][4] ) - * - */ -ALIGNTEXT16 -GLOBL GLNAME( TAG(_mesa_mmx_blend) ) -HIDDEN( TAG(_mesa_mmx_blend) ) -GLNAME( TAG(_mesa_mmx_blend) ): - - PUSH_L ( EBP ) - MOV_L ( ESP, EBP ) - PUSH_L ( ESI ) - PUSH_L ( EDI ) - PUSH_L ( EBX ) - - MOV_L ( REGOFF(12, EBP), ECX ) /* n */ - CMP_L ( CONST(0), ECX) - JE ( LLTAG(GMB_return) ) - - MOV_L ( REGOFF(16, EBP), EBX ) /* mask */ - MOV_L ( REGOFF(20, EBP), EDI ) /* rgba */ - MOV_L ( REGOFF(24, EBP), ESI ) /* dest */ - - INIT - - TEST_L ( CONST(4), EDI ) /* align rgba on an 8-byte boundary */ - JZ ( LLTAG(GMB_align_end) ) - - CMP_B ( CONST(0), REGIND(EBX) ) /* *mask == 0 */ - JE ( LLTAG(GMB_align_continue) ) - - /* runin */ -#define ONE(x) x -#define TWO(x) - MAIN ( EDI, ESI ) -#undef ONE -#undef TWO - -LLTAG(GMB_align_continue): - - DEC_L ( ECX ) /* n -= 1 */ - INC_L ( EBX ) /* mask += 1 */ - ADD_L ( CONST(4), EDI ) /* rgba += 1 */ - ADD_L ( CONST(4), ESI ) /* dest += 1 */ - -LLTAG(GMB_align_end): - - CMP_L ( CONST(2), ECX) - JB ( LLTAG(GMB_loop_end) ) - -ALIGNTEXT16 -LLTAG(GMB_loop_begin): - - CMP_W ( CONST(0), REGIND(EBX) ) /* *mask == 0 && *(mask + 1) == 0 */ - JE ( LLTAG(GMB_loop_continue) ) - - /* main loop */ -#define ONE(x) -#define TWO(x) x - MAIN ( EDI, ESI ) -#undef ONE -#undef TWO - -LLTAG(GMB_loop_continue): - - DEC_L ( ECX ) - DEC_L ( ECX ) /* n -= 2 */ - ADD_L ( CONST(2), EBX ) /* mask += 2 */ - ADD_L ( CONST(8), EDI ) /* rgba += 2 */ - ADD_L ( CONST(8), ESI ) /* dest += 2 */ - CMP_L ( CONST(2), ECX ) - JAE ( LLTAG(GMB_loop_begin) ) - -LLTAG(GMB_loop_end): - - CMP_L ( CONST(1), ECX ) - JB ( LLTAG(GMB_done) ) - - CMP_B ( CONST(0), REGIND(EBX) ) /* *mask == 0 */ - JE ( LLTAG(GMB_done) ) - - /* runout */ -#define ONE(x) x -#define TWO(x) - MAIN ( EDI, ESI ) -#undef ONE -#undef TWO - -LLTAG(GMB_done): - - EMMS - -LLTAG(GMB_return): - - POP_L ( EBX ) - POP_L ( EDI ) - POP_L ( ESI ) - MOV_L ( EBP, ESP ) - POP_L ( EBP ) - RET - -#undef TAG -#undef LLTAG -#undef INIT -#undef MAIN +/* + * Written by José Fonseca + */ + + +/* + * void _mesa_mmx_blend( struct gl_context *ctx, + * GLuint n, + * const GLubyte mask[], + * GLchan rgba[][4], + * CONST GLchan dest[][4] ) + * + */ +ALIGNTEXT16 +GLOBL GLNAME( TAG(_mesa_mmx_blend) ) +HIDDEN( TAG(_mesa_mmx_blend) ) +GLNAME( TAG(_mesa_mmx_blend) ): + + PUSH_L ( EBP ) + MOV_L ( ESP, EBP ) + PUSH_L ( ESI ) + PUSH_L ( EDI ) + PUSH_L ( EBX ) + + MOV_L ( REGOFF(12, EBP), ECX ) /* n */ + CMP_L ( CONST(0), ECX) + JE ( LLTAG(GMB_return) ) + + MOV_L ( REGOFF(16, EBP), EBX ) /* mask */ + MOV_L ( REGOFF(20, EBP), EDI ) /* rgba */ + MOV_L ( REGOFF(24, EBP), ESI ) /* dest */ + + INIT + + TEST_L ( CONST(4), EDI ) /* align rgba on an 8-byte boundary */ + JZ ( LLTAG(GMB_align_end) ) + + CMP_B ( CONST(0), REGIND(EBX) ) /* *mask == 0 */ + JE ( LLTAG(GMB_align_continue) ) + + /* runin */ +#define ONE(x) x +#define TWO(x) + MAIN ( EDI, ESI ) +#undef ONE +#undef TWO + +LLTAG(GMB_align_continue): + + DEC_L ( ECX ) /* n -= 1 */ + INC_L ( EBX ) /* mask += 1 */ + ADD_L ( CONST(4), EDI ) /* rgba += 1 */ + ADD_L ( CONST(4), ESI ) /* dest += 1 */ + +LLTAG(GMB_align_end): + + CMP_L ( CONST(2), ECX) + JB ( LLTAG(GMB_loop_end) ) + +ALIGNTEXT16 +LLTAG(GMB_loop_begin): + + CMP_W ( CONST(0), REGIND(EBX) ) /* *mask == 0 && *(mask + 1) == 0 */ + JE ( LLTAG(GMB_loop_continue) ) + + /* main loop */ +#define ONE(x) +#define TWO(x) x + MAIN ( EDI, ESI ) +#undef ONE +#undef TWO + +LLTAG(GMB_loop_continue): + + DEC_L ( ECX ) + DEC_L ( ECX ) /* n -= 2 */ + ADD_L ( CONST(2), EBX ) /* mask += 2 */ + ADD_L ( CONST(8), EDI ) /* rgba += 2 */ + ADD_L ( CONST(8), ESI ) /* dest += 2 */ + CMP_L ( CONST(2), ECX ) + JAE ( LLTAG(GMB_loop_begin) ) + +LLTAG(GMB_loop_end): + + CMP_L ( CONST(1), ECX ) + JB ( LLTAG(GMB_done) ) + + CMP_B ( CONST(0), REGIND(EBX) ) /* *mask == 0 */ + JE ( LLTAG(GMB_done) ) + + /* runout */ +#define ONE(x) x +#define TWO(x) + MAIN ( EDI, ESI ) +#undef ONE +#undef TWO + +LLTAG(GMB_done): + + EMMS + +LLTAG(GMB_return): + + POP_L ( EBX ) + POP_L ( EDI ) + POP_L ( ESI ) + MOV_L ( EBP, ESP ) + POP_L ( EBP ) + RET + +#undef TAG +#undef LLTAG +#undef INIT +#undef MAIN diff --git a/mesalib/src/mesa/x86/norm_args.h b/mesalib/src/mesa/x86/norm_args.h index bad1a1975..5d352838b 100644 --- a/mesalib/src/mesa/x86/norm_args.h +++ b/mesalib/src/mesa/x86/norm_args.h @@ -1,57 +1,57 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * Normal transform function interface for assembly code. Simply define - * FRAME_OFFSET to the number of bytes pushed onto the stack before - * using the ARG_* argument macros. - * - * Gareth Hughes - */ - -#ifndef __NORM_ARGS_H__ -#define __NORM_ARGS_H__ - -/* Offsets for normal_func arguments - * - * typedef void (*normal_func)( CONST GLmatrix *mat, - * GLfloat scale, - * CONST GLvector4f *in, - * CONST GLfloat lengths[], - * GLvector4f *dest ); - */ -#define OFFSET_MAT 4 -#define OFFSET_SCALE 8 -#define OFFSET_IN 12 -#define OFFSET_LENGTHS 16 -#define OFFSET_DEST 20 - -#define ARG_MAT REGOFF(FRAME_OFFSET+OFFSET_MAT, ESP) -#define ARG_SCALE REGOFF(FRAME_OFFSET+OFFSET_SCALE, ESP) -#define ARG_IN REGOFF(FRAME_OFFSET+OFFSET_IN, ESP) -#define ARG_LENGTHS REGOFF(FRAME_OFFSET+OFFSET_LENGTHS, ESP) -#define ARG_DEST REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP) - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * Normal transform function interface for assembly code. Simply define + * FRAME_OFFSET to the number of bytes pushed onto the stack before + * using the ARG_* argument macros. + * + * Gareth Hughes + */ + +#ifndef __NORM_ARGS_H__ +#define __NORM_ARGS_H__ + +/* Offsets for normal_func arguments + * + * typedef void (*normal_func)( CONST GLmatrix *mat, + * GLfloat scale, + * CONST GLvector4f *in, + * CONST GLfloat lengths[], + * GLvector4f *dest ); + */ +#define OFFSET_MAT 4 +#define OFFSET_SCALE 8 +#define OFFSET_IN 12 +#define OFFSET_LENGTHS 16 +#define OFFSET_DEST 20 + +#define ARG_MAT REGOFF(FRAME_OFFSET+OFFSET_MAT, ESP) +#define ARG_SCALE REGOFF(FRAME_OFFSET+OFFSET_SCALE, ESP) +#define ARG_IN REGOFF(FRAME_OFFSET+OFFSET_IN, ESP) +#define ARG_LENGTHS REGOFF(FRAME_OFFSET+OFFSET_LENGTHS, ESP) +#define ARG_DEST REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP) + +#endif diff --git a/mesalib/src/mesa/x86/read_rgba_span_x86.S b/mesalib/src/mesa/x86/read_rgba_span_x86.S index 04571afb7..3be4515b1 100644 --- a/mesalib/src/mesa/x86/read_rgba_span_x86.S +++ b/mesalib/src/mesa/x86/read_rgba_span_x86.S @@ -1,686 +1,686 @@ -/* - * (C) Copyright IBM Corporation 2004 - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * on the rights to use, copy, modify, merge, publish, distribute, sub - * license, and/or sell copies of the Software, and to permit persons to whom - * the Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL - * IBM AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM, - * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR - * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE - * USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file read_rgba_span_x86.S - * Optimized routines to transfer pixel data from the framebuffer to a - * buffer in main memory. - * - * \author Ian Romanick - */ - - .file "read_rgba_span_x86.S" -#if !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) /* this one cries for assyntax.h */ -/* Kevin F. Quinn 2nd July 2006 - * Replaced data segment constants with text-segment instructions. - */ -#define LOAD_MASK(mvins,m1,m2) \ - pushl $0xff00ff00 ;\ - pushl $0xff00ff00 ;\ - pushl $0xff00ff00 ;\ - pushl $0xff00ff00 ;\ - mvins (%esp), m1 ;\ - pushl $0x00ff0000 ;\ - pushl $0x00ff0000 ;\ - pushl $0x00ff0000 ;\ - pushl $0x00ff0000 ;\ - mvins (%esp), m2 ;\ - addl $32, %esp - -/* I implemented these as macros because they appear in several places, - * and I've tweaked them a number of times. I got tired of changing every - * place they appear. :) - */ - -#define DO_ONE_PIXEL() \ - movl (%ebx), %eax ; \ - addl $4, %ebx ; \ - bswap %eax /* ARGB -> BGRA */ ; \ - rorl $8, %eax /* BGRA -> ABGR */ ; \ - movl %eax, (%ecx) /* ABGR -> R, G, B, A */ ; \ - addl $4, %ecx - -#define DO_ONE_LAST_PIXEL() \ - movl (%ebx), %eax ; \ - bswap %eax /* ARGB -> BGRA */ ; \ - rorl $8, %eax /* BGRA -> ABGR */ ; \ - movl %eax, (%ecx) /* ABGR -> R, G, B, A */ ; \ - - -/** - * MMX optimized version of the BGRA8888_REV to RGBA copy routine. - * - * \warning - * This function assumes that the caller will issue the EMMS instruction - * at the correct places. - */ - -.globl _generic_read_RGBA_span_BGRA8888_REV_MMX -#ifndef USE_DRICORE -.hidden _generic_read_RGBA_span_BGRA8888_REV_MMX -#endif - .type _generic_read_RGBA_span_BGRA8888_REV_MMX, @function -_generic_read_RGBA_span_BGRA8888_REV_MMX: - pushl %ebx - -#ifdef USE_INNER_EMMS - emms -#endif - LOAD_MASK(movq,%mm1,%mm2) - - movl 8(%esp), %ebx /* source pointer */ - movl 16(%esp), %edx /* number of pixels to copy */ - movl 12(%esp), %ecx /* destination pointer */ - - testl %edx, %edx - jle .L20 /* Bail if there's nothing to do. */ - - movl %ebx, %eax - - negl %eax - sarl $2, %eax - andl $1, %eax - je .L17 - - subl %eax, %edx - DO_ONE_PIXEL() -.L17: - - /* Would it be faster to unroll this loop once and process 4 pixels - * per pass, instead of just two? - */ - - movl %edx, %eax - shrl %eax - jmp .L18 -.L19: - movq (%ebx), %mm0 - addl $8, %ebx - - /* These 9 instructions do what PSHUFB (if there were such an - * instruction) could do in 1. :( - */ - - movq %mm0, %mm3 - movq %mm0, %mm4 - - pand %mm2, %mm3 - psllq $16, %mm4 - psrlq $16, %mm3 - pand %mm2, %mm4 - - pand %mm1, %mm0 - por %mm4, %mm3 - por %mm3, %mm0 - - movq %mm0, (%ecx) - addl $8, %ecx - subl $1, %eax -.L18: - jne .L19 - -#ifdef USE_INNER_EMMS - emms -#endif - - /* At this point there are either 1 or 0 pixels remaining to be - * converted. Convert the last pixel, if needed. - */ - - testl $1, %edx - je .L20 - - DO_ONE_LAST_PIXEL() - -.L20: - popl %ebx - ret - .size _generic_read_RGBA_span_BGRA8888_REV_MMX, .-_generic_read_RGBA_span_BGRA8888_REV_MMX - - -/** - * SSE optimized version of the BGRA8888_REV to RGBA copy routine. SSE - * instructions are only actually used to read data from the framebuffer. - * In practice, the speed-up is pretty small. - * - * \todo - * Do some more testing and determine if there's any reason to have this - * function in addition to the MMX version. - * - * \warning - * This function assumes that the caller will issue the EMMS instruction - * at the correct places. - */ - -.globl _generic_read_RGBA_span_BGRA8888_REV_SSE -#ifndef USE_DRICORE -.hidden _generic_read_RGBA_span_BGRA8888_REV_SSE -#endif - .type _generic_read_RGBA_span_BGRA8888_REV_SSE, @function -_generic_read_RGBA_span_BGRA8888_REV_SSE: - pushl %esi - pushl %ebx - pushl %ebp - -#ifdef USE_INNER_EMMS - emms -#endif - - LOAD_MASK(movq,%mm1,%mm2) - - movl 16(%esp), %ebx /* source pointer */ - movl 24(%esp), %edx /* number of pixels to copy */ - movl 20(%esp), %ecx /* destination pointer */ - - testl %edx, %edx - jle .L35 /* Bail if there's nothing to do. */ - - movl %esp, %ebp - subl $16, %esp - andl $0xfffffff0, %esp - - movl %ebx, %eax - movl %edx, %esi - - negl %eax - andl $15, %eax - sarl $2, %eax - cmpl %edx, %eax - cmovle %eax, %esi - - subl %esi, %edx - - testl $1, %esi - je .L32 - - DO_ONE_PIXEL() -.L32: - - testl $2, %esi - je .L31 - - movq (%ebx), %mm0 - addl $8, %ebx - - movq %mm0, %mm3 - movq %mm0, %mm4 - - pand %mm2, %mm3 - psllq $16, %mm4 - psrlq $16, %mm3 - pand %mm2, %mm4 - - pand %mm1, %mm0 - por %mm4, %mm3 - por %mm3, %mm0 - - movq %mm0, (%ecx) - addl $8, %ecx -.L31: - - movl %edx, %eax - shrl $2, %eax - jmp .L33 -.L34: - movaps (%ebx), %xmm0 - addl $16, %ebx - - /* This would be so much better if we could just move directly from - * an SSE register to an MMX register. Unfortunately, that - * functionality wasn't introduced until SSE2 with the MOVDQ2Q - * instruction. - */ - - movaps %xmm0, (%esp) - movq (%esp), %mm0 - movq 8(%esp), %mm5 - - movq %mm0, %mm3 - movq %mm0, %mm4 - movq %mm5, %mm6 - movq %mm5, %mm7 - - pand %mm2, %mm3 - pand %mm2, %mm6 - - psllq $16, %mm4 - psllq $16, %mm7 - - psrlq $16, %mm3 - psrlq $16, %mm6 - - pand %mm2, %mm4 - pand %mm2, %mm7 - - pand %mm1, %mm0 - pand %mm1, %mm5 - - por %mm4, %mm3 - por %mm7, %mm6 - - por %mm3, %mm0 - por %mm6, %mm5 - - movq %mm0, (%ecx) - movq %mm5, 8(%ecx) - addl $16, %ecx - - subl $1, %eax -.L33: - jne .L34 - -#ifdef USE_INNER_EMMS - emms -#endif - movl %ebp, %esp - - /* At this point there are either [0, 3] pixels remaining to be - * converted. - */ - - testl $2, %edx - je .L36 - - movq (%ebx), %mm0 - addl $8, %ebx - - movq %mm0, %mm3 - movq %mm0, %mm4 - - pand %mm2, %mm3 - psllq $16, %mm4 - psrlq $16, %mm3 - pand %mm2, %mm4 - - pand %mm1, %mm0 - por %mm4, %mm3 - por %mm3, %mm0 - - movq %mm0, (%ecx) - addl $8, %ecx -.L36: - - testl $1, %edx - je .L35 - - DO_ONE_LAST_PIXEL() -.L35: - popl %ebp - popl %ebx - popl %esi - ret - .size _generic_read_RGBA_span_BGRA8888_REV_SSE, .-_generic_read_RGBA_span_BGRA8888_REV_SSE - - -/** - * SSE2 optimized version of the BGRA8888_REV to RGBA copy routine. - */ - - .text -.globl _generic_read_RGBA_span_BGRA8888_REV_SSE2 -#ifndef USE_DRICORE -.hidden _generic_read_RGBA_span_BGRA8888_REV_SSE2 -#endif - .type _generic_read_RGBA_span_BGRA8888_REV_SSE2, @function -_generic_read_RGBA_span_BGRA8888_REV_SSE2: - pushl %esi - pushl %ebx - - LOAD_MASK(movdqu,%xmm1,%xmm2) - - movl 12(%esp), %ebx /* source pointer */ - movl 20(%esp), %edx /* number of pixels to copy */ - movl 16(%esp), %ecx /* destination pointer */ - - movl %ebx, %eax - movl %edx, %esi - - testl %edx, %edx - jle .L46 /* Bail if there's nothing to do. */ - - /* If the source pointer isn't a multiple of 16 we have to process - * a few pixels the "slow" way to get the address aligned for - * the SSE fetch intsructions. - */ - - negl %eax - andl $15, %eax - sarl $2, %eax - - cmpl %edx, %eax - cmovbe %eax, %esi - subl %esi, %edx - - testl $1, %esi - je .L41 - - DO_ONE_PIXEL() -.L41: - testl $2, %esi - je .L40 - - movq (%ebx), %xmm0 - addl $8, %ebx - - movdqa %xmm0, %xmm3 - movdqa %xmm0, %xmm4 - andps %xmm1, %xmm0 - - andps %xmm2, %xmm3 - pslldq $2, %xmm4 - psrldq $2, %xmm3 - andps %xmm2, %xmm4 - - orps %xmm4, %xmm3 - orps %xmm3, %xmm0 - - movq %xmm0, (%ecx) - addl $8, %ecx -.L40: - - /* Would it be worth having a specialized version of this loop for - * the case where the destination is 16-byte aligned? That version - * would be identical except that it could use movedqa instead of - * movdqu. - */ - - movl %edx, %eax - shrl $2, %eax - jmp .L42 -.L43: - movdqa (%ebx), %xmm0 - addl $16, %ebx - - movdqa %xmm0, %xmm3 - movdqa %xmm0, %xmm4 - andps %xmm1, %xmm0 - - andps %xmm2, %xmm3 - pslldq $2, %xmm4 - psrldq $2, %xmm3 - andps %xmm2, %xmm4 - - orps %xmm4, %xmm3 - orps %xmm3, %xmm0 - - movdqu %xmm0, (%ecx) - addl $16, %ecx - subl $1, %eax -.L42: - jne .L43 - - - /* There may be upto 3 pixels remaining to be copied. Take care - * of them now. We do the 2 pixel case first because the data - * will be aligned. - */ - - testl $2, %edx - je .L47 - - movq (%ebx), %xmm0 - addl $8, %ebx - - movdqa %xmm0, %xmm3 - movdqa %xmm0, %xmm4 - andps %xmm1, %xmm0 - - andps %xmm2, %xmm3 - pslldq $2, %xmm4 - psrldq $2, %xmm3 - andps %xmm2, %xmm4 - - orps %xmm4, %xmm3 - orps %xmm3, %xmm0 - - movq %xmm0, (%ecx) - addl $8, %ecx -.L47: - - testl $1, %edx - je .L46 - - DO_ONE_LAST_PIXEL() -.L46: - - popl %ebx - popl %esi - ret - .size _generic_read_RGBA_span_BGRA8888_REV_SSE2, .-_generic_read_RGBA_span_BGRA8888_REV_SSE2 - - - -#define MASK_565_L 0x07e0f800 -#define MASK_565_H 0x0000001f -/* Setting SCALE_ADJUST to 5 gives a perfect match with the - * classic C implementation in Mesa. Setting SCALE_ADJUST - * to 0 is slightly faster but at a small cost to accuracy. - */ -#define SCALE_ADJUST 5 -#if SCALE_ADJUST == 5 -#define PRESCALE_L 0x00100001 -#define PRESCALE_H 0x00000200 -#define SCALE_L 0x40C620E8 -#define SCALE_H 0x0000839d -#elif SCALE_ADJUST == 0 -#define PRESCALE_L 0x00200001 -#define PRESCALE_H 0x00000800 -#define SCALE_L 0x01040108 -#define SCALE_H 0x00000108 -#else -#error SCALE_ADJUST must either be 5 or 0. -#endif -#define ALPHA_L 0x00000000 -#define ALPHA_H 0x00ff0000 - -/** - * MMX optimized version of the RGB565 to RGBA copy routine. - */ - - .text - .globl _generic_read_RGBA_span_RGB565_MMX -#ifndef USE_DRICORE - .hidden _generic_read_RGBA_span_RGB565_MMX -#endif - .type _generic_read_RGBA_span_RGB565_MMX, @function - -_generic_read_RGBA_span_RGB565_MMX: - -#ifdef USE_INNER_EMMS - emms -#endif - - movl 4(%esp), %eax /* source pointer */ - movl 8(%esp), %edx /* destination pointer */ - movl 12(%esp), %ecx /* number of pixels to copy */ - - pushl $MASK_565_H - pushl $MASK_565_L - movq (%esp), %mm5 - pushl $PRESCALE_H - pushl $PRESCALE_L - movq (%esp), %mm6 - pushl $SCALE_H - pushl $SCALE_L - movq (%esp), %mm7 - pushl $ALPHA_H - pushl $ALPHA_L - movq (%esp), %mm3 - addl $32,%esp - - sarl $2, %ecx - jl .L01 /* Bail early if the count is negative. */ - jmp .L02 - -.L03: - /* Fetch 4 RGB565 pixels into %mm4. Distribute the first and - * second pixels into the four words of %mm0 and %mm2. - */ - - movq (%eax), %mm4 - addl $8, %eax - - pshufw $0x00, %mm4, %mm0 - pshufw $0x55, %mm4, %mm2 - - - /* Mask the pixels so that each word of each register contains only - * one color component. - */ - - pand %mm5, %mm0 - pand %mm5, %mm2 - - - /* Adjust the component values so that they are as small as possible, - * but large enough so that we can multiply them by an unsigned 16-bit - * number and get a value as large as 0x00ff0000. - */ - - pmullw %mm6, %mm0 - pmullw %mm6, %mm2 -#if SCALE_ADJUST > 0 - psrlw $SCALE_ADJUST, %mm0 - psrlw $SCALE_ADJUST, %mm2 -#endif - - /* Scale the input component values to be on the range - * [0, 0x00ff0000]. This it the real magic of the whole routine. - */ - - pmulhuw %mm7, %mm0 - pmulhuw %mm7, %mm2 - - - /* Always set the alpha value to 0xff. - */ - - por %mm3, %mm0 - por %mm3, %mm2 - - - /* Pack the 16-bit values to 8-bit values and store the converted - * pixel data. - */ - - packuswb %mm2, %mm0 - movq %mm0, (%edx) - addl $8, %edx - - pshufw $0xaa, %mm4, %mm0 - pshufw $0xff, %mm4, %mm2 - - pand %mm5, %mm0 - pand %mm5, %mm2 - pmullw %mm6, %mm0 - pmullw %mm6, %mm2 -#if SCALE_ADJUST > 0 - psrlw $SCALE_ADJUST, %mm0 - psrlw $SCALE_ADJUST, %mm2 -#endif - pmulhuw %mm7, %mm0 - pmulhuw %mm7, %mm2 - - por %mm3, %mm0 - por %mm3, %mm2 - - packuswb %mm2, %mm0 - - movq %mm0, (%edx) - addl $8, %edx - - subl $1, %ecx -.L02: - jne .L03 - - - /* At this point there can be at most 3 pixels left to process. If - * there is either 2 or 3 left, process 2. - */ - - movl 12(%esp), %ecx - testl $0x02, %ecx - je .L04 - - movd (%eax), %mm4 - addl $4, %eax - - pshufw $0x00, %mm4, %mm0 - pshufw $0x55, %mm4, %mm2 - - pand %mm5, %mm0 - pand %mm5, %mm2 - pmullw %mm6, %mm0 - pmullw %mm6, %mm2 -#if SCALE_ADJUST > 0 - psrlw $SCALE_ADJUST, %mm0 - psrlw $SCALE_ADJUST, %mm2 -#endif - pmulhuw %mm7, %mm0 - pmulhuw %mm7, %mm2 - - por %mm3, %mm0 - por %mm3, %mm2 - - packuswb %mm2, %mm0 - - movq %mm0, (%edx) - addl $8, %edx - -.L04: - /* At this point there can be at most 1 pixel left to process. - * Process it if needed. - */ - - testl $0x01, %ecx - je .L01 - - movzwl (%eax), %ecx - movd %ecx, %mm4 - - pshufw $0x00, %mm4, %mm0 - - pand %mm5, %mm0 - pmullw %mm6, %mm0 -#if SCALE_ADJUST > 0 - psrlw $SCALE_ADJUST, %mm0 -#endif - pmulhuw %mm7, %mm0 - - por %mm3, %mm0 - - packuswb %mm0, %mm0 - - movd %mm0, (%edx) - -.L01: -#ifdef USE_INNER_EMMS - emms -#endif - ret -#endif /* !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) */ - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif +/* + * (C) Copyright IBM Corporation 2004 + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * on the rights to use, copy, modify, merge, publish, distribute, sub + * license, and/or sell copies of the Software, and to permit persons to whom + * the Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * IBM AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file read_rgba_span_x86.S + * Optimized routines to transfer pixel data from the framebuffer to a + * buffer in main memory. + * + * \author Ian Romanick + */ + + .file "read_rgba_span_x86.S" +#if !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) /* this one cries for assyntax.h */ +/* Kevin F. Quinn 2nd July 2006 + * Replaced data segment constants with text-segment instructions. + */ +#define LOAD_MASK(mvins,m1,m2) \ + pushl $0xff00ff00 ;\ + pushl $0xff00ff00 ;\ + pushl $0xff00ff00 ;\ + pushl $0xff00ff00 ;\ + mvins (%esp), m1 ;\ + pushl $0x00ff0000 ;\ + pushl $0x00ff0000 ;\ + pushl $0x00ff0000 ;\ + pushl $0x00ff0000 ;\ + mvins (%esp), m2 ;\ + addl $32, %esp + +/* I implemented these as macros because they appear in several places, + * and I've tweaked them a number of times. I got tired of changing every + * place they appear. :) + */ + +#define DO_ONE_PIXEL() \ + movl (%ebx), %eax ; \ + addl $4, %ebx ; \ + bswap %eax /* ARGB -> BGRA */ ; \ + rorl $8, %eax /* BGRA -> ABGR */ ; \ + movl %eax, (%ecx) /* ABGR -> R, G, B, A */ ; \ + addl $4, %ecx + +#define DO_ONE_LAST_PIXEL() \ + movl (%ebx), %eax ; \ + bswap %eax /* ARGB -> BGRA */ ; \ + rorl $8, %eax /* BGRA -> ABGR */ ; \ + movl %eax, (%ecx) /* ABGR -> R, G, B, A */ ; \ + + +/** + * MMX optimized version of the BGRA8888_REV to RGBA copy routine. + * + * \warning + * This function assumes that the caller will issue the EMMS instruction + * at the correct places. + */ + +.globl _generic_read_RGBA_span_BGRA8888_REV_MMX +#ifndef USE_DRICORE +.hidden _generic_read_RGBA_span_BGRA8888_REV_MMX +#endif + .type _generic_read_RGBA_span_BGRA8888_REV_MMX, @function +_generic_read_RGBA_span_BGRA8888_REV_MMX: + pushl %ebx + +#ifdef USE_INNER_EMMS + emms +#endif + LOAD_MASK(movq,%mm1,%mm2) + + movl 8(%esp), %ebx /* source pointer */ + movl 16(%esp), %edx /* number of pixels to copy */ + movl 12(%esp), %ecx /* destination pointer */ + + testl %edx, %edx + jle .L20 /* Bail if there's nothing to do. */ + + movl %ebx, %eax + + negl %eax + sarl $2, %eax + andl $1, %eax + je .L17 + + subl %eax, %edx + DO_ONE_PIXEL() +.L17: + + /* Would it be faster to unroll this loop once and process 4 pixels + * per pass, instead of just two? + */ + + movl %edx, %eax + shrl %eax + jmp .L18 +.L19: + movq (%ebx), %mm0 + addl $8, %ebx + + /* These 9 instructions do what PSHUFB (if there were such an + * instruction) could do in 1. :( + */ + + movq %mm0, %mm3 + movq %mm0, %mm4 + + pand %mm2, %mm3 + psllq $16, %mm4 + psrlq $16, %mm3 + pand %mm2, %mm4 + + pand %mm1, %mm0 + por %mm4, %mm3 + por %mm3, %mm0 + + movq %mm0, (%ecx) + addl $8, %ecx + subl $1, %eax +.L18: + jne .L19 + +#ifdef USE_INNER_EMMS + emms +#endif + + /* At this point there are either 1 or 0 pixels remaining to be + * converted. Convert the last pixel, if needed. + */ + + testl $1, %edx + je .L20 + + DO_ONE_LAST_PIXEL() + +.L20: + popl %ebx + ret + .size _generic_read_RGBA_span_BGRA8888_REV_MMX, .-_generic_read_RGBA_span_BGRA8888_REV_MMX + + +/** + * SSE optimized version of the BGRA8888_REV to RGBA copy routine. SSE + * instructions are only actually used to read data from the framebuffer. + * In practice, the speed-up is pretty small. + * + * \todo + * Do some more testing and determine if there's any reason to have this + * function in addition to the MMX version. + * + * \warning + * This function assumes that the caller will issue the EMMS instruction + * at the correct places. + */ + +.globl _generic_read_RGBA_span_BGRA8888_REV_SSE +#ifndef USE_DRICORE +.hidden _generic_read_RGBA_span_BGRA8888_REV_SSE +#endif + .type _generic_read_RGBA_span_BGRA8888_REV_SSE, @function +_generic_read_RGBA_span_BGRA8888_REV_SSE: + pushl %esi + pushl %ebx + pushl %ebp + +#ifdef USE_INNER_EMMS + emms +#endif + + LOAD_MASK(movq,%mm1,%mm2) + + movl 16(%esp), %ebx /* source pointer */ + movl 24(%esp), %edx /* number of pixels to copy */ + movl 20(%esp), %ecx /* destination pointer */ + + testl %edx, %edx + jle .L35 /* Bail if there's nothing to do. */ + + movl %esp, %ebp + subl $16, %esp + andl $0xfffffff0, %esp + + movl %ebx, %eax + movl %edx, %esi + + negl %eax + andl $15, %eax + sarl $2, %eax + cmpl %edx, %eax + cmovle %eax, %esi + + subl %esi, %edx + + testl $1, %esi + je .L32 + + DO_ONE_PIXEL() +.L32: + + testl $2, %esi + je .L31 + + movq (%ebx), %mm0 + addl $8, %ebx + + movq %mm0, %mm3 + movq %mm0, %mm4 + + pand %mm2, %mm3 + psllq $16, %mm4 + psrlq $16, %mm3 + pand %mm2, %mm4 + + pand %mm1, %mm0 + por %mm4, %mm3 + por %mm3, %mm0 + + movq %mm0, (%ecx) + addl $8, %ecx +.L31: + + movl %edx, %eax + shrl $2, %eax + jmp .L33 +.L34: + movaps (%ebx), %xmm0 + addl $16, %ebx + + /* This would be so much better if we could just move directly from + * an SSE register to an MMX register. Unfortunately, that + * functionality wasn't introduced until SSE2 with the MOVDQ2Q + * instruction. + */ + + movaps %xmm0, (%esp) + movq (%esp), %mm0 + movq 8(%esp), %mm5 + + movq %mm0, %mm3 + movq %mm0, %mm4 + movq %mm5, %mm6 + movq %mm5, %mm7 + + pand %mm2, %mm3 + pand %mm2, %mm6 + + psllq $16, %mm4 + psllq $16, %mm7 + + psrlq $16, %mm3 + psrlq $16, %mm6 + + pand %mm2, %mm4 + pand %mm2, %mm7 + + pand %mm1, %mm0 + pand %mm1, %mm5 + + por %mm4, %mm3 + por %mm7, %mm6 + + por %mm3, %mm0 + por %mm6, %mm5 + + movq %mm0, (%ecx) + movq %mm5, 8(%ecx) + addl $16, %ecx + + subl $1, %eax +.L33: + jne .L34 + +#ifdef USE_INNER_EMMS + emms +#endif + movl %ebp, %esp + + /* At this point there are either [0, 3] pixels remaining to be + * converted. + */ + + testl $2, %edx + je .L36 + + movq (%ebx), %mm0 + addl $8, %ebx + + movq %mm0, %mm3 + movq %mm0, %mm4 + + pand %mm2, %mm3 + psllq $16, %mm4 + psrlq $16, %mm3 + pand %mm2, %mm4 + + pand %mm1, %mm0 + por %mm4, %mm3 + por %mm3, %mm0 + + movq %mm0, (%ecx) + addl $8, %ecx +.L36: + + testl $1, %edx + je .L35 + + DO_ONE_LAST_PIXEL() +.L35: + popl %ebp + popl %ebx + popl %esi + ret + .size _generic_read_RGBA_span_BGRA8888_REV_SSE, .-_generic_read_RGBA_span_BGRA8888_REV_SSE + + +/** + * SSE2 optimized version of the BGRA8888_REV to RGBA copy routine. + */ + + .text +.globl _generic_read_RGBA_span_BGRA8888_REV_SSE2 +#ifndef USE_DRICORE +.hidden _generic_read_RGBA_span_BGRA8888_REV_SSE2 +#endif + .type _generic_read_RGBA_span_BGRA8888_REV_SSE2, @function +_generic_read_RGBA_span_BGRA8888_REV_SSE2: + pushl %esi + pushl %ebx + + LOAD_MASK(movdqu,%xmm1,%xmm2) + + movl 12(%esp), %ebx /* source pointer */ + movl 20(%esp), %edx /* number of pixels to copy */ + movl 16(%esp), %ecx /* destination pointer */ + + movl %ebx, %eax + movl %edx, %esi + + testl %edx, %edx + jle .L46 /* Bail if there's nothing to do. */ + + /* If the source pointer isn't a multiple of 16 we have to process + * a few pixels the "slow" way to get the address aligned for + * the SSE fetch intsructions. + */ + + negl %eax + andl $15, %eax + sarl $2, %eax + + cmpl %edx, %eax + cmovbe %eax, %esi + subl %esi, %edx + + testl $1, %esi + je .L41 + + DO_ONE_PIXEL() +.L41: + testl $2, %esi + je .L40 + + movq (%ebx), %xmm0 + addl $8, %ebx + + movdqa %xmm0, %xmm3 + movdqa %xmm0, %xmm4 + andps %xmm1, %xmm0 + + andps %xmm2, %xmm3 + pslldq $2, %xmm4 + psrldq $2, %xmm3 + andps %xmm2, %xmm4 + + orps %xmm4, %xmm3 + orps %xmm3, %xmm0 + + movq %xmm0, (%ecx) + addl $8, %ecx +.L40: + + /* Would it be worth having a specialized version of this loop for + * the case where the destination is 16-byte aligned? That version + * would be identical except that it could use movedqa instead of + * movdqu. + */ + + movl %edx, %eax + shrl $2, %eax + jmp .L42 +.L43: + movdqa (%ebx), %xmm0 + addl $16, %ebx + + movdqa %xmm0, %xmm3 + movdqa %xmm0, %xmm4 + andps %xmm1, %xmm0 + + andps %xmm2, %xmm3 + pslldq $2, %xmm4 + psrldq $2, %xmm3 + andps %xmm2, %xmm4 + + orps %xmm4, %xmm3 + orps %xmm3, %xmm0 + + movdqu %xmm0, (%ecx) + addl $16, %ecx + subl $1, %eax +.L42: + jne .L43 + + + /* There may be upto 3 pixels remaining to be copied. Take care + * of them now. We do the 2 pixel case first because the data + * will be aligned. + */ + + testl $2, %edx + je .L47 + + movq (%ebx), %xmm0 + addl $8, %ebx + + movdqa %xmm0, %xmm3 + movdqa %xmm0, %xmm4 + andps %xmm1, %xmm0 + + andps %xmm2, %xmm3 + pslldq $2, %xmm4 + psrldq $2, %xmm3 + andps %xmm2, %xmm4 + + orps %xmm4, %xmm3 + orps %xmm3, %xmm0 + + movq %xmm0, (%ecx) + addl $8, %ecx +.L47: + + testl $1, %edx + je .L46 + + DO_ONE_LAST_PIXEL() +.L46: + + popl %ebx + popl %esi + ret + .size _generic_read_RGBA_span_BGRA8888_REV_SSE2, .-_generic_read_RGBA_span_BGRA8888_REV_SSE2 + + + +#define MASK_565_L 0x07e0f800 +#define MASK_565_H 0x0000001f +/* Setting SCALE_ADJUST to 5 gives a perfect match with the + * classic C implementation in Mesa. Setting SCALE_ADJUST + * to 0 is slightly faster but at a small cost to accuracy. + */ +#define SCALE_ADJUST 5 +#if SCALE_ADJUST == 5 +#define PRESCALE_L 0x00100001 +#define PRESCALE_H 0x00000200 +#define SCALE_L 0x40C620E8 +#define SCALE_H 0x0000839d +#elif SCALE_ADJUST == 0 +#define PRESCALE_L 0x00200001 +#define PRESCALE_H 0x00000800 +#define SCALE_L 0x01040108 +#define SCALE_H 0x00000108 +#else +#error SCALE_ADJUST must either be 5 or 0. +#endif +#define ALPHA_L 0x00000000 +#define ALPHA_H 0x00ff0000 + +/** + * MMX optimized version of the RGB565 to RGBA copy routine. + */ + + .text + .globl _generic_read_RGBA_span_RGB565_MMX +#ifndef USE_DRICORE + .hidden _generic_read_RGBA_span_RGB565_MMX +#endif + .type _generic_read_RGBA_span_RGB565_MMX, @function + +_generic_read_RGBA_span_RGB565_MMX: + +#ifdef USE_INNER_EMMS + emms +#endif + + movl 4(%esp), %eax /* source pointer */ + movl 8(%esp), %edx /* destination pointer */ + movl 12(%esp), %ecx /* number of pixels to copy */ + + pushl $MASK_565_H + pushl $MASK_565_L + movq (%esp), %mm5 + pushl $PRESCALE_H + pushl $PRESCALE_L + movq (%esp), %mm6 + pushl $SCALE_H + pushl $SCALE_L + movq (%esp), %mm7 + pushl $ALPHA_H + pushl $ALPHA_L + movq (%esp), %mm3 + addl $32,%esp + + sarl $2, %ecx + jl .L01 /* Bail early if the count is negative. */ + jmp .L02 + +.L03: + /* Fetch 4 RGB565 pixels into %mm4. Distribute the first and + * second pixels into the four words of %mm0 and %mm2. + */ + + movq (%eax), %mm4 + addl $8, %eax + + pshufw $0x00, %mm4, %mm0 + pshufw $0x55, %mm4, %mm2 + + + /* Mask the pixels so that each word of each register contains only + * one color component. + */ + + pand %mm5, %mm0 + pand %mm5, %mm2 + + + /* Adjust the component values so that they are as small as possible, + * but large enough so that we can multiply them by an unsigned 16-bit + * number and get a value as large as 0x00ff0000. + */ + + pmullw %mm6, %mm0 + pmullw %mm6, %mm2 +#if SCALE_ADJUST > 0 + psrlw $SCALE_ADJUST, %mm0 + psrlw $SCALE_ADJUST, %mm2 +#endif + + /* Scale the input component values to be on the range + * [0, 0x00ff0000]. This it the real magic of the whole routine. + */ + + pmulhuw %mm7, %mm0 + pmulhuw %mm7, %mm2 + + + /* Always set the alpha value to 0xff. + */ + + por %mm3, %mm0 + por %mm3, %mm2 + + + /* Pack the 16-bit values to 8-bit values and store the converted + * pixel data. + */ + + packuswb %mm2, %mm0 + movq %mm0, (%edx) + addl $8, %edx + + pshufw $0xaa, %mm4, %mm0 + pshufw $0xff, %mm4, %mm2 + + pand %mm5, %mm0 + pand %mm5, %mm2 + pmullw %mm6, %mm0 + pmullw %mm6, %mm2 +#if SCALE_ADJUST > 0 + psrlw $SCALE_ADJUST, %mm0 + psrlw $SCALE_ADJUST, %mm2 +#endif + pmulhuw %mm7, %mm0 + pmulhuw %mm7, %mm2 + + por %mm3, %mm0 + por %mm3, %mm2 + + packuswb %mm2, %mm0 + + movq %mm0, (%edx) + addl $8, %edx + + subl $1, %ecx +.L02: + jne .L03 + + + /* At this point there can be at most 3 pixels left to process. If + * there is either 2 or 3 left, process 2. + */ + + movl 12(%esp), %ecx + testl $0x02, %ecx + je .L04 + + movd (%eax), %mm4 + addl $4, %eax + + pshufw $0x00, %mm4, %mm0 + pshufw $0x55, %mm4, %mm2 + + pand %mm5, %mm0 + pand %mm5, %mm2 + pmullw %mm6, %mm0 + pmullw %mm6, %mm2 +#if SCALE_ADJUST > 0 + psrlw $SCALE_ADJUST, %mm0 + psrlw $SCALE_ADJUST, %mm2 +#endif + pmulhuw %mm7, %mm0 + pmulhuw %mm7, %mm2 + + por %mm3, %mm0 + por %mm3, %mm2 + + packuswb %mm2, %mm0 + + movq %mm0, (%edx) + addl $8, %edx + +.L04: + /* At this point there can be at most 1 pixel left to process. + * Process it if needed. + */ + + testl $0x01, %ecx + je .L01 + + movzwl (%eax), %ecx + movd %ecx, %mm4 + + pshufw $0x00, %mm4, %mm0 + + pand %mm5, %mm0 + pmullw %mm6, %mm0 +#if SCALE_ADJUST > 0 + psrlw $SCALE_ADJUST, %mm0 +#endif + pmulhuw %mm7, %mm0 + + por %mm3, %mm0 + + packuswb %mm0, %mm0 + + movd %mm0, (%edx) + +.L01: +#ifdef USE_INNER_EMMS + emms +#endif + ret +#endif /* !defined(__DJGPP__) && !defined(__MINGW32__) && !defined(__APPLE__) */ + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/read_rgba_span_x86.h b/mesalib/src/mesa/x86/read_rgba_span_x86.h index d255f6dcb..564b1bb0f 100644 --- a/mesalib/src/mesa/x86/read_rgba_span_x86.h +++ b/mesalib/src/mesa/x86/read_rgba_span_x86.h @@ -1,56 +1,56 @@ -/* - * (C) Copyright IBM Corporation 2004 - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * on the rights to use, copy, modify, merge, publish, distribute, sub - * license, and/or sell copies of the Software, and to permit persons to whom - * the Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice (including the next - * paragraph) shall be included in all copies or substantial portions of the - * Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL - * IBM AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM, - * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR - * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE - * USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** - * \file read_rgba_span_x86.h - * - * \author Ian Romanick - */ - -#ifndef READ_RGBA_SPAN_X86_H -#define READ_RGBA_SPAN_X86_H - -#if defined(USE_SSE_ASM) || defined(USE_MMX_ASM) -#include "x86/common_x86_asm.h" -#endif - -#if defined(USE_SSE_ASM) -extern void _generic_read_RGBA_span_BGRA8888_REV_SSE2( const unsigned char *, - unsigned char *, unsigned ); -#endif - -#if defined(USE_SSE_ASM) -extern void _generic_read_RGBA_span_BGRA8888_REV_SSE( const unsigned char *, - unsigned char *, unsigned ); -#endif - -#if defined(USE_MMX_ASM) -extern void _generic_read_RGBA_span_BGRA8888_REV_MMX( const unsigned char *, - unsigned char *, unsigned ); - -extern void _generic_read_RGBA_span_RGB565_MMX( const unsigned char *, - unsigned char *, unsigned ); -#endif - -#endif /* READ_RGBA_SPAN_X86_H */ +/* + * (C) Copyright IBM Corporation 2004 + * All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * on the rights to use, copy, modify, merge, publish, distribute, sub + * license, and/or sell copies of the Software, and to permit persons to whom + * the Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * IBM AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM, + * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR + * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE + * USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** + * \file read_rgba_span_x86.h + * + * \author Ian Romanick + */ + +#ifndef READ_RGBA_SPAN_X86_H +#define READ_RGBA_SPAN_X86_H + +#if defined(USE_SSE_ASM) || defined(USE_MMX_ASM) +#include "x86/common_x86_asm.h" +#endif + +#if defined(USE_SSE_ASM) +extern void _generic_read_RGBA_span_BGRA8888_REV_SSE2( const unsigned char *, + unsigned char *, unsigned ); +#endif + +#if defined(USE_SSE_ASM) +extern void _generic_read_RGBA_span_BGRA8888_REV_SSE( const unsigned char *, + unsigned char *, unsigned ); +#endif + +#if defined(USE_MMX_ASM) +extern void _generic_read_RGBA_span_BGRA8888_REV_MMX( const unsigned char *, + unsigned char *, unsigned ); + +extern void _generic_read_RGBA_span_RGB565_MMX( const unsigned char *, + unsigned char *, unsigned ); +#endif + +#endif /* READ_RGBA_SPAN_X86_H */ diff --git a/mesalib/src/mesa/x86/rtasm/x86sse.c b/mesalib/src/mesa/x86/rtasm/x86sse.c index 08e3969b2..c93faba79 100644 --- a/mesalib/src/mesa/x86/rtasm/x86sse.c +++ b/mesalib/src/mesa/x86/rtasm/x86sse.c @@ -1,1203 +1,1203 @@ -#ifdef USE_X86_ASM -#if defined(__i386__) || defined(__386__) - -#include "main/imports.h" -#include "x86sse.h" - -#define DISASSEM 0 -#define X86_TWOB 0x0f - -#if 0 -static unsigned char *cptr( void (*label)() ) -{ - return (unsigned char *)(unsigned long)label; -} -#endif - - -static void do_realloc( struct x86_function *p ) -{ - if (p->size == 0) { - p->size = 1024; - p->store = _mesa_exec_malloc(p->size); - p->csr = p->store; - } - else { - unsigned used = p->csr - p->store; - unsigned char *tmp = p->store; - p->size *= 2; - p->store = _mesa_exec_malloc(p->size); - memcpy(p->store, tmp, used); - p->csr = p->store + used; - _mesa_exec_free(tmp); - } -} - -/* Emit bytes to the instruction stream: - */ -static unsigned char *reserve( struct x86_function *p, int bytes ) -{ - if (p->csr + bytes - p->store > p->size) - do_realloc(p); - - { - unsigned char *csr = p->csr; - p->csr += bytes; - return csr; - } -} - - - -static void emit_1b( struct x86_function *p, char b0 ) -{ - char *csr = (char *)reserve(p, 1); - *csr = b0; -} - -static void emit_1i( struct x86_function *p, int i0 ) -{ - int *icsr = (int *)reserve(p, sizeof(i0)); - *icsr = i0; -} - -static void emit_1ub( struct x86_function *p, unsigned char b0 ) -{ - unsigned char *csr = reserve(p, 1); - *csr++ = b0; -} - -static void emit_2ub( struct x86_function *p, unsigned char b0, unsigned char b1 ) -{ - unsigned char *csr = reserve(p, 2); - *csr++ = b0; - *csr++ = b1; -} - -static void emit_3ub( struct x86_function *p, unsigned char b0, unsigned char b1, unsigned char b2 ) -{ - unsigned char *csr = reserve(p, 3); - *csr++ = b0; - *csr++ = b1; - *csr++ = b2; -} - - -/* Build a modRM byte + possible displacement. No treatment of SIB - * indexing. BZZT - no way to encode an absolute address. - */ -static void emit_modrm( struct x86_function *p, - struct x86_reg reg, - struct x86_reg regmem ) -{ - unsigned char val = 0; - - assert(reg.mod == mod_REG); - - val |= regmem.mod << 6; /* mod field */ - val |= reg.idx << 3; /* reg field */ - val |= regmem.idx; /* r/m field */ - - emit_1ub(p, val); - - /* Oh-oh we've stumbled into the SIB thing. - */ - if (regmem.file == file_REG32 && - regmem.idx == reg_SP) { - emit_1ub(p, 0x24); /* simplistic! */ - } - - switch (regmem.mod) { - case mod_REG: - case mod_INDIRECT: - break; - case mod_DISP8: - emit_1b(p, regmem.disp); - break; - case mod_DISP32: - emit_1i(p, regmem.disp); - break; - default: - assert(0); - break; - } -} - - -static void emit_modrm_noreg( struct x86_function *p, - unsigned op, - struct x86_reg regmem ) -{ - struct x86_reg dummy = x86_make_reg(file_REG32, op); - emit_modrm(p, dummy, regmem); -} - -/* Many x86 instructions have two opcodes to cope with the situations - * where the destination is a register or memory reference - * respectively. This function selects the correct opcode based on - * the arguments presented. - */ -static void emit_op_modrm( struct x86_function *p, - unsigned char op_dst_is_reg, - unsigned char op_dst_is_mem, - struct x86_reg dst, - struct x86_reg src ) -{ - switch (dst.mod) { - case mod_REG: - emit_1ub(p, op_dst_is_reg); - emit_modrm(p, dst, src); - break; - case mod_INDIRECT: - case mod_DISP32: - case mod_DISP8: - assert(src.mod == mod_REG); - emit_1ub(p, op_dst_is_mem); - emit_modrm(p, src, dst); - break; - default: - assert(0); - break; - } -} - - - - - - - -/* Create and manipulate registers and regmem values: - */ -struct x86_reg x86_make_reg( enum x86_reg_file file, - enum x86_reg_name idx ) -{ - struct x86_reg reg; - - reg.file = file; - reg.idx = idx; - reg.mod = mod_REG; - reg.disp = 0; - - return reg; -} - -struct x86_reg x86_make_disp( struct x86_reg reg, - int disp ) -{ - assert(reg.file == file_REG32); - - if (reg.mod == mod_REG) - reg.disp = disp; - else - reg.disp += disp; - - if (reg.disp == 0) - reg.mod = mod_INDIRECT; - else if (reg.disp <= 127 && reg.disp >= -128) - reg.mod = mod_DISP8; - else - reg.mod = mod_DISP32; - - return reg; -} - -struct x86_reg x86_deref( struct x86_reg reg ) -{ - return x86_make_disp(reg, 0); -} - -struct x86_reg x86_get_base_reg( struct x86_reg reg ) -{ - return x86_make_reg( reg.file, reg.idx ); -} - -unsigned char *x86_get_label( struct x86_function *p ) -{ - return p->csr; -} - - - -/*********************************************************************** - * x86 instructions - */ - - -void x86_jcc( struct x86_function *p, - enum x86_cc cc, - unsigned char *label ) -{ - int offset = label - (x86_get_label(p) + 2); - - if (offset <= 127 && offset >= -128) { - emit_1ub(p, 0x70 + cc); - emit_1b(p, (char) offset); - } - else { - offset = label - (x86_get_label(p) + 6); - emit_2ub(p, 0x0f, 0x80 + cc); - emit_1i(p, offset); - } -} - -/* Always use a 32bit offset for forward jumps: - */ -unsigned char *x86_jcc_forward( struct x86_function *p, - enum x86_cc cc ) -{ - emit_2ub(p, 0x0f, 0x80 + cc); - emit_1i(p, 0); - return x86_get_label(p); -} - -unsigned char *x86_jmp_forward( struct x86_function *p) -{ - emit_1ub(p, 0xe9); - emit_1i(p, 0); - return x86_get_label(p); -} - -unsigned char *x86_call_forward( struct x86_function *p) -{ - emit_1ub(p, 0xe8); - emit_1i(p, 0); - return x86_get_label(p); -} - -/* Fixup offset from forward jump: - */ -void x86_fixup_fwd_jump( struct x86_function *p, - unsigned char *fixup ) -{ - *(int *)(fixup - 4) = x86_get_label(p) - fixup; -} - -void x86_jmp( struct x86_function *p, unsigned char *label) -{ - emit_1ub(p, 0xe9); - emit_1i(p, label - x86_get_label(p) - 4); -} - -#if 0 -/* This doesn't work once we start reallocating & copying the - * generated code on buffer fills, because the call is relative to the - * current pc. - */ -void x86_call( struct x86_function *p, void (*label)()) -{ - emit_1ub(p, 0xe8); - emit_1i(p, cptr(label) - x86_get_label(p) - 4); -} -#else -void x86_call( struct x86_function *p, struct x86_reg reg) -{ - emit_1ub(p, 0xff); - emit_modrm_noreg(p, 2, reg); -} -#endif - - -/* michal: - * Temporary. As I need immediate operands, and dont want to mess with the codegen, - * I load the immediate into general purpose register and use it. - */ -void x86_mov_reg_imm( struct x86_function *p, struct x86_reg dst, int imm ) -{ - assert(dst.mod == mod_REG); - emit_1ub(p, 0xb8 + dst.idx); - emit_1i(p, imm); -} - -void x86_push( struct x86_function *p, - struct x86_reg reg ) -{ - assert(reg.mod == mod_REG); - emit_1ub(p, 0x50 + reg.idx); - p->stack_offset += 4; -} - -void x86_pop( struct x86_function *p, - struct x86_reg reg ) -{ - assert(reg.mod == mod_REG); - emit_1ub(p, 0x58 + reg.idx); - p->stack_offset -= 4; -} - -void x86_inc( struct x86_function *p, - struct x86_reg reg ) -{ - assert(reg.mod == mod_REG); - emit_1ub(p, 0x40 + reg.idx); -} - -void x86_dec( struct x86_function *p, - struct x86_reg reg ) -{ - assert(reg.mod == mod_REG); - emit_1ub(p, 0x48 + reg.idx); -} - -void x86_ret( struct x86_function *p ) -{ - emit_1ub(p, 0xc3); -} - -void x86_sahf( struct x86_function *p ) -{ - emit_1ub(p, 0x9e); -} - -void x86_mov( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_op_modrm( p, 0x8b, 0x89, dst, src ); -} - -void x86_xor( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_op_modrm( p, 0x33, 0x31, dst, src ); -} - -void x86_cmp( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_op_modrm( p, 0x3b, 0x39, dst, src ); -} - -void x86_lea( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_1ub(p, 0x8d); - emit_modrm( p, dst, src ); -} - -void x86_test( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_1ub(p, 0x85); - emit_modrm( p, dst, src ); -} - -void x86_add( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_op_modrm(p, 0x03, 0x01, dst, src ); -} - -void x86_mul( struct x86_function *p, - struct x86_reg src ) -{ - assert (src.file == file_REG32 && src.mod == mod_REG); - emit_op_modrm(p, 0xf7, 0, x86_make_reg (file_REG32, reg_SP), src ); -} - -void x86_sub( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_op_modrm(p, 0x2b, 0x29, dst, src ); -} - -void x86_or( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_op_modrm( p, 0x0b, 0x09, dst, src ); -} - -void x86_and( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_op_modrm( p, 0x23, 0x21, dst, src ); -} - - - -/*********************************************************************** - * SSE instructions - */ - - -void sse_movss( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, 0xF3, X86_TWOB); - emit_op_modrm( p, 0x10, 0x11, dst, src ); -} - -void sse_movaps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_1ub(p, X86_TWOB); - emit_op_modrm( p, 0x28, 0x29, dst, src ); -} - -void sse_movups( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_1ub(p, X86_TWOB); - emit_op_modrm( p, 0x10, 0x11, dst, src ); -} - -void sse_movhps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - assert(dst.mod != mod_REG || src.mod != mod_REG); - emit_1ub(p, X86_TWOB); - emit_op_modrm( p, 0x16, 0x17, dst, src ); /* cf movlhps */ -} - -void sse_movlps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - assert(dst.mod != mod_REG || src.mod != mod_REG); - emit_1ub(p, X86_TWOB); - emit_op_modrm( p, 0x12, 0x13, dst, src ); /* cf movhlps */ -} - -void sse_maxps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x5F); - emit_modrm( p, dst, src ); -} - -void sse_maxss( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0xF3, X86_TWOB, 0x5F); - emit_modrm( p, dst, src ); -} - -void sse_divss( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0xF3, X86_TWOB, 0x5E); - emit_modrm( p, dst, src ); -} - -void sse_minps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x5D); - emit_modrm( p, dst, src ); -} - -void sse_subps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x5C); - emit_modrm( p, dst, src ); -} - -void sse_mulps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x59); - emit_modrm( p, dst, src ); -} - -void sse_mulss( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0xF3, X86_TWOB, 0x59); - emit_modrm( p, dst, src ); -} - -void sse_addps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x58); - emit_modrm( p, dst, src ); -} - -void sse_addss( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0xF3, X86_TWOB, 0x58); - emit_modrm( p, dst, src ); -} - -void sse_andnps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x55); - emit_modrm( p, dst, src ); -} - -void sse_andps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x54); - emit_modrm( p, dst, src ); -} - -void sse_rsqrtps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x52); - emit_modrm( p, dst, src ); -} - -void sse_rsqrtss( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0xF3, X86_TWOB, 0x52); - emit_modrm( p, dst, src ); - -} - -void sse_movhlps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - assert(dst.mod == mod_REG && src.mod == mod_REG); - emit_2ub(p, X86_TWOB, 0x12); - emit_modrm( p, dst, src ); -} - -void sse_movlhps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - assert(dst.mod == mod_REG && src.mod == mod_REG); - emit_2ub(p, X86_TWOB, 0x16); - emit_modrm( p, dst, src ); -} - -void sse_orps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x56); - emit_modrm( p, dst, src ); -} - -void sse_xorps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x57); - emit_modrm( p, dst, src ); -} - -void sse_cvtps2pi( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - assert(dst.file == file_MMX && - (src.file == file_XMM || src.mod != mod_REG)); - - p->need_emms = 1; - - emit_2ub(p, X86_TWOB, 0x2d); - emit_modrm( p, dst, src ); -} - - -/* Shufps can also be used to implement a reduced swizzle when dest == - * arg0. - */ -void sse_shufps( struct x86_function *p, - struct x86_reg dest, - struct x86_reg arg0, - unsigned char shuf) -{ - emit_2ub(p, X86_TWOB, 0xC6); - emit_modrm(p, dest, arg0); - emit_1ub(p, shuf); -} - -void sse_cmpps( struct x86_function *p, - struct x86_reg dest, - struct x86_reg arg0, - unsigned char cc) -{ - emit_2ub(p, X86_TWOB, 0xC2); - emit_modrm(p, dest, arg0); - emit_1ub(p, cc); -} - -void sse_pmovmskb( struct x86_function *p, - struct x86_reg dest, - struct x86_reg src) -{ - emit_3ub(p, 0x66, X86_TWOB, 0xD7); - emit_modrm(p, dest, src); -} - -/*********************************************************************** - * SSE2 instructions - */ - -/** - * Perform a reduced swizzle: - */ -void sse2_pshufd( struct x86_function *p, - struct x86_reg dest, - struct x86_reg arg0, - unsigned char shuf) -{ - emit_3ub(p, 0x66, X86_TWOB, 0x70); - emit_modrm(p, dest, arg0); - emit_1ub(p, shuf); -} - -void sse2_cvttps2dq( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub( p, 0xF3, X86_TWOB, 0x5B ); - emit_modrm( p, dst, src ); -} - -void sse2_cvtps2dq( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0x66, X86_TWOB, 0x5B); - emit_modrm( p, dst, src ); -} - -void sse2_packssdw( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0x66, X86_TWOB, 0x6B); - emit_modrm( p, dst, src ); -} - -void sse2_packsswb( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0x66, X86_TWOB, 0x63); - emit_modrm( p, dst, src ); -} - -void sse2_packuswb( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0x66, X86_TWOB, 0x67); - emit_modrm( p, dst, src ); -} - -void sse2_rcpps( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, X86_TWOB, 0x53); - emit_modrm( p, dst, src ); -} - -void sse2_rcpss( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_3ub(p, 0xF3, X86_TWOB, 0x53); - emit_modrm( p, dst, src ); -} - -void sse2_movd( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - emit_2ub(p, 0x66, X86_TWOB); - emit_op_modrm( p, 0x6e, 0x7e, dst, src ); -} - - - - -/*********************************************************************** - * x87 instructions - */ -void x87_fist( struct x86_function *p, struct x86_reg dst ) -{ - emit_1ub(p, 0xdb); - emit_modrm_noreg(p, 2, dst); -} - -void x87_fistp( struct x86_function *p, struct x86_reg dst ) -{ - emit_1ub(p, 0xdb); - emit_modrm_noreg(p, 3, dst); -} - -void x87_fild( struct x86_function *p, struct x86_reg arg ) -{ - emit_1ub(p, 0xdf); - emit_modrm_noreg(p, 0, arg); -} - -void x87_fldz( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xee); -} - - -void x87_fldcw( struct x86_function *p, struct x86_reg arg ) -{ - assert(arg.file == file_REG32); - assert(arg.mod != mod_REG); - emit_1ub(p, 0xd9); - emit_modrm_noreg(p, 5, arg); -} - -void x87_fld1( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xe8); -} - -void x87_fldl2e( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xea); -} - -void x87_fldln2( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xed); -} - -void x87_fwait( struct x86_function *p ) -{ - emit_1ub(p, 0x9b); -} - -void x87_fnclex( struct x86_function *p ) -{ - emit_2ub(p, 0xdb, 0xe2); -} - -void x87_fclex( struct x86_function *p ) -{ - x87_fwait(p); - x87_fnclex(p); -} - - -static void x87_arith_op( struct x86_function *p, struct x86_reg dst, struct x86_reg arg, - unsigned char dst0ub0, - unsigned char dst0ub1, - unsigned char arg0ub0, - unsigned char arg0ub1, - unsigned char argmem_noreg) -{ - assert(dst.file == file_x87); - - if (arg.file == file_x87) { - if (dst.idx == 0) - emit_2ub(p, dst0ub0, dst0ub1+arg.idx); - else if (arg.idx == 0) - emit_2ub(p, arg0ub0, arg0ub1+arg.idx); - else - assert(0); - } - else if (dst.idx == 0) { - assert(arg.file == file_REG32); - emit_1ub(p, 0xd8); - emit_modrm_noreg(p, argmem_noreg, arg); - } - else - assert(0); -} - -void x87_fmul( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) -{ - x87_arith_op(p, dst, arg, - 0xd8, 0xc8, - 0xdc, 0xc8, - 4); -} - -void x87_fsub( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) -{ - x87_arith_op(p, dst, arg, - 0xd8, 0xe0, - 0xdc, 0xe8, - 4); -} - -void x87_fsubr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) -{ - x87_arith_op(p, dst, arg, - 0xd8, 0xe8, - 0xdc, 0xe0, - 5); -} - -void x87_fadd( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) -{ - x87_arith_op(p, dst, arg, - 0xd8, 0xc0, - 0xdc, 0xc0, - 0); -} - -void x87_fdiv( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) -{ - x87_arith_op(p, dst, arg, - 0xd8, 0xf0, - 0xdc, 0xf8, - 6); -} - -void x87_fdivr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) -{ - x87_arith_op(p, dst, arg, - 0xd8, 0xf8, - 0xdc, 0xf0, - 7); -} - -void x87_fmulp( struct x86_function *p, struct x86_reg dst ) -{ - assert(dst.file == file_x87); - assert(dst.idx >= 1); - emit_2ub(p, 0xde, 0xc8+dst.idx); -} - -void x87_fsubp( struct x86_function *p, struct x86_reg dst ) -{ - assert(dst.file == file_x87); - assert(dst.idx >= 1); - emit_2ub(p, 0xde, 0xe8+dst.idx); -} - -void x87_fsubrp( struct x86_function *p, struct x86_reg dst ) -{ - assert(dst.file == file_x87); - assert(dst.idx >= 1); - emit_2ub(p, 0xde, 0xe0+dst.idx); -} - -void x87_faddp( struct x86_function *p, struct x86_reg dst ) -{ - assert(dst.file == file_x87); - assert(dst.idx >= 1); - emit_2ub(p, 0xde, 0xc0+dst.idx); -} - -void x87_fdivp( struct x86_function *p, struct x86_reg dst ) -{ - assert(dst.file == file_x87); - assert(dst.idx >= 1); - emit_2ub(p, 0xde, 0xf8+dst.idx); -} - -void x87_fdivrp( struct x86_function *p, struct x86_reg dst ) -{ - assert(dst.file == file_x87); - assert(dst.idx >= 1); - emit_2ub(p, 0xde, 0xf0+dst.idx); -} - -void x87_fucom( struct x86_function *p, struct x86_reg arg ) -{ - assert(arg.file == file_x87); - emit_2ub(p, 0xdd, 0xe0+arg.idx); -} - -void x87_fucomp( struct x86_function *p, struct x86_reg arg ) -{ - assert(arg.file == file_x87); - emit_2ub(p, 0xdd, 0xe8+arg.idx); -} - -void x87_fucompp( struct x86_function *p ) -{ - emit_2ub(p, 0xda, 0xe9); -} - -void x87_fxch( struct x86_function *p, struct x86_reg arg ) -{ - assert(arg.file == file_x87); - emit_2ub(p, 0xd9, 0xc8+arg.idx); -} - -void x87_fabs( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xe1); -} - -void x87_fchs( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xe0); -} - -void x87_fcos( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xff); -} - - -void x87_fprndint( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xfc); -} - -void x87_fscale( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xfd); -} - -void x87_fsin( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xfe); -} - -void x87_fsincos( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xfb); -} - -void x87_fsqrt( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xfa); -} - -void x87_fxtract( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xf4); -} - -/* st0 = (2^st0)-1 - * - * Restrictions: -1.0 <= st0 <= 1.0 - */ -void x87_f2xm1( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xf0); -} - -/* st1 = st1 * log2(st0); - * pop_stack; - */ -void x87_fyl2x( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xf1); -} - -/* st1 = st1 * log2(st0 + 1.0); - * pop_stack; - * - * A fast operation, with restrictions: -.29 < st0 < .29 - */ -void x87_fyl2xp1( struct x86_function *p ) -{ - emit_2ub(p, 0xd9, 0xf9); -} - - -void x87_fld( struct x86_function *p, struct x86_reg arg ) -{ - if (arg.file == file_x87) - emit_2ub(p, 0xd9, 0xc0 + arg.idx); - else { - emit_1ub(p, 0xd9); - emit_modrm_noreg(p, 0, arg); - } -} - -void x87_fst( struct x86_function *p, struct x86_reg dst ) -{ - if (dst.file == file_x87) - emit_2ub(p, 0xdd, 0xd0 + dst.idx); - else { - emit_1ub(p, 0xd9); - emit_modrm_noreg(p, 2, dst); - } -} - -void x87_fstp( struct x86_function *p, struct x86_reg dst ) -{ - if (dst.file == file_x87) - emit_2ub(p, 0xdd, 0xd8 + dst.idx); - else { - emit_1ub(p, 0xd9); - emit_modrm_noreg(p, 3, dst); - } -} - -void x87_fcom( struct x86_function *p, struct x86_reg dst ) -{ - if (dst.file == file_x87) - emit_2ub(p, 0xd8, 0xd0 + dst.idx); - else { - emit_1ub(p, 0xd8); - emit_modrm_noreg(p, 2, dst); - } -} - -void x87_fcomp( struct x86_function *p, struct x86_reg dst ) -{ - if (dst.file == file_x87) - emit_2ub(p, 0xd8, 0xd8 + dst.idx); - else { - emit_1ub(p, 0xd8); - emit_modrm_noreg(p, 3, dst); - } -} - - -void x87_fnstsw( struct x86_function *p, struct x86_reg dst ) -{ - assert(dst.file == file_REG32); - - if (dst.idx == reg_AX && - dst.mod == mod_REG) - emit_2ub(p, 0xdf, 0xe0); - else { - emit_1ub(p, 0xdd); - emit_modrm_noreg(p, 7, dst); - } -} - - - - -/*********************************************************************** - * MMX instructions - */ - -void mmx_emms( struct x86_function *p ) -{ - assert(p->need_emms); - emit_2ub(p, 0x0f, 0x77); - p->need_emms = 0; -} - -void mmx_packssdw( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - assert(dst.file == file_MMX && - (src.file == file_MMX || src.mod != mod_REG)); - - p->need_emms = 1; - - emit_2ub(p, X86_TWOB, 0x6b); - emit_modrm( p, dst, src ); -} - -void mmx_packuswb( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - assert(dst.file == file_MMX && - (src.file == file_MMX || src.mod != mod_REG)); - - p->need_emms = 1; - - emit_2ub(p, X86_TWOB, 0x67); - emit_modrm( p, dst, src ); -} - -void mmx_movd( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - p->need_emms = 1; - emit_1ub(p, X86_TWOB); - emit_op_modrm( p, 0x6e, 0x7e, dst, src ); -} - -void mmx_movq( struct x86_function *p, - struct x86_reg dst, - struct x86_reg src ) -{ - p->need_emms = 1; - emit_1ub(p, X86_TWOB); - emit_op_modrm( p, 0x6f, 0x7f, dst, src ); -} - - -/*********************************************************************** - * Helper functions - */ - - -/* Retreive a reference to one of the function arguments, taking into - * account any push/pop activity: - */ -struct x86_reg x86_fn_arg( struct x86_function *p, - unsigned arg ) -{ - return x86_make_disp(x86_make_reg(file_REG32, reg_SP), - p->stack_offset + arg * 4); /* ??? */ -} - - -void x86_init_func( struct x86_function *p ) -{ - p->size = 0; - p->store = NULL; - p->csr = p->store; -} - -int x86_init_func_size( struct x86_function *p, unsigned code_size ) -{ - p->size = code_size; - p->store = _mesa_exec_malloc(code_size); - p->csr = p->store; - return p->store != NULL; -} - -void x86_release_func( struct x86_function *p ) -{ - _mesa_exec_free(p->store); - p->store = NULL; - p->csr = NULL; - p->size = 0; -} - - -void (*x86_get_func( struct x86_function *p ))(void) -{ - if (DISASSEM && p->store) - printf("disassemble %p %p\n", p->store, p->csr); - return (void (*)(void)) (unsigned long) p->store; -} - -#else - -void x86sse_dummy( void ) -{ -} - -#endif - -#else /* USE_X86_ASM */ - -int x86sse_c_dummy_var; /* silence warning */ - -#endif /* USE_X86_ASM */ +#ifdef USE_X86_ASM +#if defined(__i386__) || defined(__386__) + +#include "main/imports.h" +#include "x86sse.h" + +#define DISASSEM 0 +#define X86_TWOB 0x0f + +#if 0 +static unsigned char *cptr( void (*label)() ) +{ + return (unsigned char *)(unsigned long)label; +} +#endif + + +static void do_realloc( struct x86_function *p ) +{ + if (p->size == 0) { + p->size = 1024; + p->store = _mesa_exec_malloc(p->size); + p->csr = p->store; + } + else { + unsigned used = p->csr - p->store; + unsigned char *tmp = p->store; + p->size *= 2; + p->store = _mesa_exec_malloc(p->size); + memcpy(p->store, tmp, used); + p->csr = p->store + used; + _mesa_exec_free(tmp); + } +} + +/* Emit bytes to the instruction stream: + */ +static unsigned char *reserve( struct x86_function *p, int bytes ) +{ + if (p->csr + bytes - p->store > p->size) + do_realloc(p); + + { + unsigned char *csr = p->csr; + p->csr += bytes; + return csr; + } +} + + + +static void emit_1b( struct x86_function *p, char b0 ) +{ + char *csr = (char *)reserve(p, 1); + *csr = b0; +} + +static void emit_1i( struct x86_function *p, int i0 ) +{ + int *icsr = (int *)reserve(p, sizeof(i0)); + *icsr = i0; +} + +static void emit_1ub( struct x86_function *p, unsigned char b0 ) +{ + unsigned char *csr = reserve(p, 1); + *csr++ = b0; +} + +static void emit_2ub( struct x86_function *p, unsigned char b0, unsigned char b1 ) +{ + unsigned char *csr = reserve(p, 2); + *csr++ = b0; + *csr++ = b1; +} + +static void emit_3ub( struct x86_function *p, unsigned char b0, unsigned char b1, unsigned char b2 ) +{ + unsigned char *csr = reserve(p, 3); + *csr++ = b0; + *csr++ = b1; + *csr++ = b2; +} + + +/* Build a modRM byte + possible displacement. No treatment of SIB + * indexing. BZZT - no way to encode an absolute address. + */ +static void emit_modrm( struct x86_function *p, + struct x86_reg reg, + struct x86_reg regmem ) +{ + unsigned char val = 0; + + assert(reg.mod == mod_REG); + + val |= regmem.mod << 6; /* mod field */ + val |= reg.idx << 3; /* reg field */ + val |= regmem.idx; /* r/m field */ + + emit_1ub(p, val); + + /* Oh-oh we've stumbled into the SIB thing. + */ + if (regmem.file == file_REG32 && + regmem.idx == reg_SP) { + emit_1ub(p, 0x24); /* simplistic! */ + } + + switch (regmem.mod) { + case mod_REG: + case mod_INDIRECT: + break; + case mod_DISP8: + emit_1b(p, regmem.disp); + break; + case mod_DISP32: + emit_1i(p, regmem.disp); + break; + default: + assert(0); + break; + } +} + + +static void emit_modrm_noreg( struct x86_function *p, + unsigned op, + struct x86_reg regmem ) +{ + struct x86_reg dummy = x86_make_reg(file_REG32, op); + emit_modrm(p, dummy, regmem); +} + +/* Many x86 instructions have two opcodes to cope with the situations + * where the destination is a register or memory reference + * respectively. This function selects the correct opcode based on + * the arguments presented. + */ +static void emit_op_modrm( struct x86_function *p, + unsigned char op_dst_is_reg, + unsigned char op_dst_is_mem, + struct x86_reg dst, + struct x86_reg src ) +{ + switch (dst.mod) { + case mod_REG: + emit_1ub(p, op_dst_is_reg); + emit_modrm(p, dst, src); + break; + case mod_INDIRECT: + case mod_DISP32: + case mod_DISP8: + assert(src.mod == mod_REG); + emit_1ub(p, op_dst_is_mem); + emit_modrm(p, src, dst); + break; + default: + assert(0); + break; + } +} + + + + + + + +/* Create and manipulate registers and regmem values: + */ +struct x86_reg x86_make_reg( enum x86_reg_file file, + enum x86_reg_name idx ) +{ + struct x86_reg reg; + + reg.file = file; + reg.idx = idx; + reg.mod = mod_REG; + reg.disp = 0; + + return reg; +} + +struct x86_reg x86_make_disp( struct x86_reg reg, + int disp ) +{ + assert(reg.file == file_REG32); + + if (reg.mod == mod_REG) + reg.disp = disp; + else + reg.disp += disp; + + if (reg.disp == 0) + reg.mod = mod_INDIRECT; + else if (reg.disp <= 127 && reg.disp >= -128) + reg.mod = mod_DISP8; + else + reg.mod = mod_DISP32; + + return reg; +} + +struct x86_reg x86_deref( struct x86_reg reg ) +{ + return x86_make_disp(reg, 0); +} + +struct x86_reg x86_get_base_reg( struct x86_reg reg ) +{ + return x86_make_reg( reg.file, reg.idx ); +} + +unsigned char *x86_get_label( struct x86_function *p ) +{ + return p->csr; +} + + + +/*********************************************************************** + * x86 instructions + */ + + +void x86_jcc( struct x86_function *p, + enum x86_cc cc, + unsigned char *label ) +{ + int offset = label - (x86_get_label(p) + 2); + + if (offset <= 127 && offset >= -128) { + emit_1ub(p, 0x70 + cc); + emit_1b(p, (char) offset); + } + else { + offset = label - (x86_get_label(p) + 6); + emit_2ub(p, 0x0f, 0x80 + cc); + emit_1i(p, offset); + } +} + +/* Always use a 32bit offset for forward jumps: + */ +unsigned char *x86_jcc_forward( struct x86_function *p, + enum x86_cc cc ) +{ + emit_2ub(p, 0x0f, 0x80 + cc); + emit_1i(p, 0); + return x86_get_label(p); +} + +unsigned char *x86_jmp_forward( struct x86_function *p) +{ + emit_1ub(p, 0xe9); + emit_1i(p, 0); + return x86_get_label(p); +} + +unsigned char *x86_call_forward( struct x86_function *p) +{ + emit_1ub(p, 0xe8); + emit_1i(p, 0); + return x86_get_label(p); +} + +/* Fixup offset from forward jump: + */ +void x86_fixup_fwd_jump( struct x86_function *p, + unsigned char *fixup ) +{ + *(int *)(fixup - 4) = x86_get_label(p) - fixup; +} + +void x86_jmp( struct x86_function *p, unsigned char *label) +{ + emit_1ub(p, 0xe9); + emit_1i(p, label - x86_get_label(p) - 4); +} + +#if 0 +/* This doesn't work once we start reallocating & copying the + * generated code on buffer fills, because the call is relative to the + * current pc. + */ +void x86_call( struct x86_function *p, void (*label)()) +{ + emit_1ub(p, 0xe8); + emit_1i(p, cptr(label) - x86_get_label(p) - 4); +} +#else +void x86_call( struct x86_function *p, struct x86_reg reg) +{ + emit_1ub(p, 0xff); + emit_modrm_noreg(p, 2, reg); +} +#endif + + +/* michal: + * Temporary. As I need immediate operands, and dont want to mess with the codegen, + * I load the immediate into general purpose register and use it. + */ +void x86_mov_reg_imm( struct x86_function *p, struct x86_reg dst, int imm ) +{ + assert(dst.mod == mod_REG); + emit_1ub(p, 0xb8 + dst.idx); + emit_1i(p, imm); +} + +void x86_push( struct x86_function *p, + struct x86_reg reg ) +{ + assert(reg.mod == mod_REG); + emit_1ub(p, 0x50 + reg.idx); + p->stack_offset += 4; +} + +void x86_pop( struct x86_function *p, + struct x86_reg reg ) +{ + assert(reg.mod == mod_REG); + emit_1ub(p, 0x58 + reg.idx); + p->stack_offset -= 4; +} + +void x86_inc( struct x86_function *p, + struct x86_reg reg ) +{ + assert(reg.mod == mod_REG); + emit_1ub(p, 0x40 + reg.idx); +} + +void x86_dec( struct x86_function *p, + struct x86_reg reg ) +{ + assert(reg.mod == mod_REG); + emit_1ub(p, 0x48 + reg.idx); +} + +void x86_ret( struct x86_function *p ) +{ + emit_1ub(p, 0xc3); +} + +void x86_sahf( struct x86_function *p ) +{ + emit_1ub(p, 0x9e); +} + +void x86_mov( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_op_modrm( p, 0x8b, 0x89, dst, src ); +} + +void x86_xor( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_op_modrm( p, 0x33, 0x31, dst, src ); +} + +void x86_cmp( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_op_modrm( p, 0x3b, 0x39, dst, src ); +} + +void x86_lea( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_1ub(p, 0x8d); + emit_modrm( p, dst, src ); +} + +void x86_test( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_1ub(p, 0x85); + emit_modrm( p, dst, src ); +} + +void x86_add( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_op_modrm(p, 0x03, 0x01, dst, src ); +} + +void x86_mul( struct x86_function *p, + struct x86_reg src ) +{ + assert (src.file == file_REG32 && src.mod == mod_REG); + emit_op_modrm(p, 0xf7, 0, x86_make_reg (file_REG32, reg_SP), src ); +} + +void x86_sub( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_op_modrm(p, 0x2b, 0x29, dst, src ); +} + +void x86_or( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_op_modrm( p, 0x0b, 0x09, dst, src ); +} + +void x86_and( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_op_modrm( p, 0x23, 0x21, dst, src ); +} + + + +/*********************************************************************** + * SSE instructions + */ + + +void sse_movss( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, 0xF3, X86_TWOB); + emit_op_modrm( p, 0x10, 0x11, dst, src ); +} + +void sse_movaps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_1ub(p, X86_TWOB); + emit_op_modrm( p, 0x28, 0x29, dst, src ); +} + +void sse_movups( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_1ub(p, X86_TWOB); + emit_op_modrm( p, 0x10, 0x11, dst, src ); +} + +void sse_movhps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + assert(dst.mod != mod_REG || src.mod != mod_REG); + emit_1ub(p, X86_TWOB); + emit_op_modrm( p, 0x16, 0x17, dst, src ); /* cf movlhps */ +} + +void sse_movlps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + assert(dst.mod != mod_REG || src.mod != mod_REG); + emit_1ub(p, X86_TWOB); + emit_op_modrm( p, 0x12, 0x13, dst, src ); /* cf movhlps */ +} + +void sse_maxps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x5F); + emit_modrm( p, dst, src ); +} + +void sse_maxss( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0xF3, X86_TWOB, 0x5F); + emit_modrm( p, dst, src ); +} + +void sse_divss( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0xF3, X86_TWOB, 0x5E); + emit_modrm( p, dst, src ); +} + +void sse_minps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x5D); + emit_modrm( p, dst, src ); +} + +void sse_subps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x5C); + emit_modrm( p, dst, src ); +} + +void sse_mulps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x59); + emit_modrm( p, dst, src ); +} + +void sse_mulss( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0xF3, X86_TWOB, 0x59); + emit_modrm( p, dst, src ); +} + +void sse_addps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x58); + emit_modrm( p, dst, src ); +} + +void sse_addss( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0xF3, X86_TWOB, 0x58); + emit_modrm( p, dst, src ); +} + +void sse_andnps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x55); + emit_modrm( p, dst, src ); +} + +void sse_andps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x54); + emit_modrm( p, dst, src ); +} + +void sse_rsqrtps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x52); + emit_modrm( p, dst, src ); +} + +void sse_rsqrtss( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0xF3, X86_TWOB, 0x52); + emit_modrm( p, dst, src ); + +} + +void sse_movhlps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + assert(dst.mod == mod_REG && src.mod == mod_REG); + emit_2ub(p, X86_TWOB, 0x12); + emit_modrm( p, dst, src ); +} + +void sse_movlhps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + assert(dst.mod == mod_REG && src.mod == mod_REG); + emit_2ub(p, X86_TWOB, 0x16); + emit_modrm( p, dst, src ); +} + +void sse_orps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x56); + emit_modrm( p, dst, src ); +} + +void sse_xorps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x57); + emit_modrm( p, dst, src ); +} + +void sse_cvtps2pi( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + assert(dst.file == file_MMX && + (src.file == file_XMM || src.mod != mod_REG)); + + p->need_emms = 1; + + emit_2ub(p, X86_TWOB, 0x2d); + emit_modrm( p, dst, src ); +} + + +/* Shufps can also be used to implement a reduced swizzle when dest == + * arg0. + */ +void sse_shufps( struct x86_function *p, + struct x86_reg dest, + struct x86_reg arg0, + unsigned char shuf) +{ + emit_2ub(p, X86_TWOB, 0xC6); + emit_modrm(p, dest, arg0); + emit_1ub(p, shuf); +} + +void sse_cmpps( struct x86_function *p, + struct x86_reg dest, + struct x86_reg arg0, + unsigned char cc) +{ + emit_2ub(p, X86_TWOB, 0xC2); + emit_modrm(p, dest, arg0); + emit_1ub(p, cc); +} + +void sse_pmovmskb( struct x86_function *p, + struct x86_reg dest, + struct x86_reg src) +{ + emit_3ub(p, 0x66, X86_TWOB, 0xD7); + emit_modrm(p, dest, src); +} + +/*********************************************************************** + * SSE2 instructions + */ + +/** + * Perform a reduced swizzle: + */ +void sse2_pshufd( struct x86_function *p, + struct x86_reg dest, + struct x86_reg arg0, + unsigned char shuf) +{ + emit_3ub(p, 0x66, X86_TWOB, 0x70); + emit_modrm(p, dest, arg0); + emit_1ub(p, shuf); +} + +void sse2_cvttps2dq( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub( p, 0xF3, X86_TWOB, 0x5B ); + emit_modrm( p, dst, src ); +} + +void sse2_cvtps2dq( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0x66, X86_TWOB, 0x5B); + emit_modrm( p, dst, src ); +} + +void sse2_packssdw( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0x66, X86_TWOB, 0x6B); + emit_modrm( p, dst, src ); +} + +void sse2_packsswb( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0x66, X86_TWOB, 0x63); + emit_modrm( p, dst, src ); +} + +void sse2_packuswb( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0x66, X86_TWOB, 0x67); + emit_modrm( p, dst, src ); +} + +void sse2_rcpps( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, X86_TWOB, 0x53); + emit_modrm( p, dst, src ); +} + +void sse2_rcpss( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_3ub(p, 0xF3, X86_TWOB, 0x53); + emit_modrm( p, dst, src ); +} + +void sse2_movd( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + emit_2ub(p, 0x66, X86_TWOB); + emit_op_modrm( p, 0x6e, 0x7e, dst, src ); +} + + + + +/*********************************************************************** + * x87 instructions + */ +void x87_fist( struct x86_function *p, struct x86_reg dst ) +{ + emit_1ub(p, 0xdb); + emit_modrm_noreg(p, 2, dst); +} + +void x87_fistp( struct x86_function *p, struct x86_reg dst ) +{ + emit_1ub(p, 0xdb); + emit_modrm_noreg(p, 3, dst); +} + +void x87_fild( struct x86_function *p, struct x86_reg arg ) +{ + emit_1ub(p, 0xdf); + emit_modrm_noreg(p, 0, arg); +} + +void x87_fldz( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xee); +} + + +void x87_fldcw( struct x86_function *p, struct x86_reg arg ) +{ + assert(arg.file == file_REG32); + assert(arg.mod != mod_REG); + emit_1ub(p, 0xd9); + emit_modrm_noreg(p, 5, arg); +} + +void x87_fld1( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xe8); +} + +void x87_fldl2e( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xea); +} + +void x87_fldln2( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xed); +} + +void x87_fwait( struct x86_function *p ) +{ + emit_1ub(p, 0x9b); +} + +void x87_fnclex( struct x86_function *p ) +{ + emit_2ub(p, 0xdb, 0xe2); +} + +void x87_fclex( struct x86_function *p ) +{ + x87_fwait(p); + x87_fnclex(p); +} + + +static void x87_arith_op( struct x86_function *p, struct x86_reg dst, struct x86_reg arg, + unsigned char dst0ub0, + unsigned char dst0ub1, + unsigned char arg0ub0, + unsigned char arg0ub1, + unsigned char argmem_noreg) +{ + assert(dst.file == file_x87); + + if (arg.file == file_x87) { + if (dst.idx == 0) + emit_2ub(p, dst0ub0, dst0ub1+arg.idx); + else if (arg.idx == 0) + emit_2ub(p, arg0ub0, arg0ub1+arg.idx); + else + assert(0); + } + else if (dst.idx == 0) { + assert(arg.file == file_REG32); + emit_1ub(p, 0xd8); + emit_modrm_noreg(p, argmem_noreg, arg); + } + else + assert(0); +} + +void x87_fmul( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) +{ + x87_arith_op(p, dst, arg, + 0xd8, 0xc8, + 0xdc, 0xc8, + 4); +} + +void x87_fsub( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) +{ + x87_arith_op(p, dst, arg, + 0xd8, 0xe0, + 0xdc, 0xe8, + 4); +} + +void x87_fsubr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) +{ + x87_arith_op(p, dst, arg, + 0xd8, 0xe8, + 0xdc, 0xe0, + 5); +} + +void x87_fadd( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) +{ + x87_arith_op(p, dst, arg, + 0xd8, 0xc0, + 0xdc, 0xc0, + 0); +} + +void x87_fdiv( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) +{ + x87_arith_op(p, dst, arg, + 0xd8, 0xf0, + 0xdc, 0xf8, + 6); +} + +void x87_fdivr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ) +{ + x87_arith_op(p, dst, arg, + 0xd8, 0xf8, + 0xdc, 0xf0, + 7); +} + +void x87_fmulp( struct x86_function *p, struct x86_reg dst ) +{ + assert(dst.file == file_x87); + assert(dst.idx >= 1); + emit_2ub(p, 0xde, 0xc8+dst.idx); +} + +void x87_fsubp( struct x86_function *p, struct x86_reg dst ) +{ + assert(dst.file == file_x87); + assert(dst.idx >= 1); + emit_2ub(p, 0xde, 0xe8+dst.idx); +} + +void x87_fsubrp( struct x86_function *p, struct x86_reg dst ) +{ + assert(dst.file == file_x87); + assert(dst.idx >= 1); + emit_2ub(p, 0xde, 0xe0+dst.idx); +} + +void x87_faddp( struct x86_function *p, struct x86_reg dst ) +{ + assert(dst.file == file_x87); + assert(dst.idx >= 1); + emit_2ub(p, 0xde, 0xc0+dst.idx); +} + +void x87_fdivp( struct x86_function *p, struct x86_reg dst ) +{ + assert(dst.file == file_x87); + assert(dst.idx >= 1); + emit_2ub(p, 0xde, 0xf8+dst.idx); +} + +void x87_fdivrp( struct x86_function *p, struct x86_reg dst ) +{ + assert(dst.file == file_x87); + assert(dst.idx >= 1); + emit_2ub(p, 0xde, 0xf0+dst.idx); +} + +void x87_fucom( struct x86_function *p, struct x86_reg arg ) +{ + assert(arg.file == file_x87); + emit_2ub(p, 0xdd, 0xe0+arg.idx); +} + +void x87_fucomp( struct x86_function *p, struct x86_reg arg ) +{ + assert(arg.file == file_x87); + emit_2ub(p, 0xdd, 0xe8+arg.idx); +} + +void x87_fucompp( struct x86_function *p ) +{ + emit_2ub(p, 0xda, 0xe9); +} + +void x87_fxch( struct x86_function *p, struct x86_reg arg ) +{ + assert(arg.file == file_x87); + emit_2ub(p, 0xd9, 0xc8+arg.idx); +} + +void x87_fabs( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xe1); +} + +void x87_fchs( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xe0); +} + +void x87_fcos( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xff); +} + + +void x87_fprndint( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xfc); +} + +void x87_fscale( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xfd); +} + +void x87_fsin( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xfe); +} + +void x87_fsincos( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xfb); +} + +void x87_fsqrt( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xfa); +} + +void x87_fxtract( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xf4); +} + +/* st0 = (2^st0)-1 + * + * Restrictions: -1.0 <= st0 <= 1.0 + */ +void x87_f2xm1( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xf0); +} + +/* st1 = st1 * log2(st0); + * pop_stack; + */ +void x87_fyl2x( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xf1); +} + +/* st1 = st1 * log2(st0 + 1.0); + * pop_stack; + * + * A fast operation, with restrictions: -.29 < st0 < .29 + */ +void x87_fyl2xp1( struct x86_function *p ) +{ + emit_2ub(p, 0xd9, 0xf9); +} + + +void x87_fld( struct x86_function *p, struct x86_reg arg ) +{ + if (arg.file == file_x87) + emit_2ub(p, 0xd9, 0xc0 + arg.idx); + else { + emit_1ub(p, 0xd9); + emit_modrm_noreg(p, 0, arg); + } +} + +void x87_fst( struct x86_function *p, struct x86_reg dst ) +{ + if (dst.file == file_x87) + emit_2ub(p, 0xdd, 0xd0 + dst.idx); + else { + emit_1ub(p, 0xd9); + emit_modrm_noreg(p, 2, dst); + } +} + +void x87_fstp( struct x86_function *p, struct x86_reg dst ) +{ + if (dst.file == file_x87) + emit_2ub(p, 0xdd, 0xd8 + dst.idx); + else { + emit_1ub(p, 0xd9); + emit_modrm_noreg(p, 3, dst); + } +} + +void x87_fcom( struct x86_function *p, struct x86_reg dst ) +{ + if (dst.file == file_x87) + emit_2ub(p, 0xd8, 0xd0 + dst.idx); + else { + emit_1ub(p, 0xd8); + emit_modrm_noreg(p, 2, dst); + } +} + +void x87_fcomp( struct x86_function *p, struct x86_reg dst ) +{ + if (dst.file == file_x87) + emit_2ub(p, 0xd8, 0xd8 + dst.idx); + else { + emit_1ub(p, 0xd8); + emit_modrm_noreg(p, 3, dst); + } +} + + +void x87_fnstsw( struct x86_function *p, struct x86_reg dst ) +{ + assert(dst.file == file_REG32); + + if (dst.idx == reg_AX && + dst.mod == mod_REG) + emit_2ub(p, 0xdf, 0xe0); + else { + emit_1ub(p, 0xdd); + emit_modrm_noreg(p, 7, dst); + } +} + + + + +/*********************************************************************** + * MMX instructions + */ + +void mmx_emms( struct x86_function *p ) +{ + assert(p->need_emms); + emit_2ub(p, 0x0f, 0x77); + p->need_emms = 0; +} + +void mmx_packssdw( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + assert(dst.file == file_MMX && + (src.file == file_MMX || src.mod != mod_REG)); + + p->need_emms = 1; + + emit_2ub(p, X86_TWOB, 0x6b); + emit_modrm( p, dst, src ); +} + +void mmx_packuswb( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + assert(dst.file == file_MMX && + (src.file == file_MMX || src.mod != mod_REG)); + + p->need_emms = 1; + + emit_2ub(p, X86_TWOB, 0x67); + emit_modrm( p, dst, src ); +} + +void mmx_movd( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + p->need_emms = 1; + emit_1ub(p, X86_TWOB); + emit_op_modrm( p, 0x6e, 0x7e, dst, src ); +} + +void mmx_movq( struct x86_function *p, + struct x86_reg dst, + struct x86_reg src ) +{ + p->need_emms = 1; + emit_1ub(p, X86_TWOB); + emit_op_modrm( p, 0x6f, 0x7f, dst, src ); +} + + +/*********************************************************************** + * Helper functions + */ + + +/* Retreive a reference to one of the function arguments, taking into + * account any push/pop activity: + */ +struct x86_reg x86_fn_arg( struct x86_function *p, + unsigned arg ) +{ + return x86_make_disp(x86_make_reg(file_REG32, reg_SP), + p->stack_offset + arg * 4); /* ??? */ +} + + +void x86_init_func( struct x86_function *p ) +{ + p->size = 0; + p->store = NULL; + p->csr = p->store; +} + +int x86_init_func_size( struct x86_function *p, unsigned code_size ) +{ + p->size = code_size; + p->store = _mesa_exec_malloc(code_size); + p->csr = p->store; + return p->store != NULL; +} + +void x86_release_func( struct x86_function *p ) +{ + _mesa_exec_free(p->store); + p->store = NULL; + p->csr = NULL; + p->size = 0; +} + + +void (*x86_get_func( struct x86_function *p ))(void) +{ + if (DISASSEM && p->store) + printf("disassemble %p %p\n", p->store, p->csr); + return (void (*)(void)) (unsigned long) p->store; +} + +#else + +void x86sse_dummy( void ) +{ +} + +#endif + +#else /* USE_X86_ASM */ + +int x86sse_c_dummy_var; /* silence warning */ + +#endif /* USE_X86_ASM */ diff --git a/mesalib/src/mesa/x86/rtasm/x86sse.h b/mesalib/src/mesa/x86/rtasm/x86sse.h index 399a694bc..f6282f5bd 100644 --- a/mesalib/src/mesa/x86/rtasm/x86sse.h +++ b/mesalib/src/mesa/x86/rtasm/x86sse.h @@ -1,256 +1,256 @@ - -#ifndef _X86SSE_H_ -#define _X86SSE_H_ - -#if defined(__i386__) || defined(__386__) - -/* It is up to the caller to ensure that instructions issued are - * suitable for the host cpu. There are no checks made in this module - * for mmx/sse/sse2 support on the cpu. - */ -struct x86_reg { - unsigned file:3; - unsigned idx:3; - unsigned mod:2; /* mod_REG if this is just a register */ - int disp:24; /* only +/- 23bits of offset - should be enough... */ -}; - -struct x86_function { - unsigned size; - unsigned char *store; - unsigned char *csr; - unsigned stack_offset; - int need_emms; - const char *fn; -}; - -enum x86_reg_file { - file_REG32, - file_MMX, - file_XMM, - file_x87 -}; - -/* Values for mod field of modr/m byte - */ -enum x86_reg_mod { - mod_INDIRECT, - mod_DISP8, - mod_DISP32, - mod_REG -}; - -enum x86_reg_name { - reg_AX, - reg_CX, - reg_DX, - reg_BX, - reg_SP, - reg_BP, - reg_SI, - reg_DI -}; - - -enum x86_cc { - cc_O, /* overflow */ - cc_NO, /* not overflow */ - cc_NAE, /* not above or equal / carry */ - cc_AE, /* above or equal / not carry */ - cc_E, /* equal / zero */ - cc_NE /* not equal / not zero */ -}; - -enum sse_cc { - cc_Equal, - cc_LessThan, - cc_LessThanEqual, - cc_Unordered, - cc_NotEqual, - cc_NotLessThan, - cc_NotLessThanEqual, - cc_Ordered -}; - -#define cc_Z cc_E -#define cc_NZ cc_NE - -/* Begin/end/retreive function creation: - */ - - -void x86_init_func( struct x86_function *p ); -int x86_init_func_size( struct x86_function *p, unsigned code_size ); -void x86_release_func( struct x86_function *p ); -void (*x86_get_func( struct x86_function *p ))( void ); - - - -/* Create and manipulate registers and regmem values: - */ -struct x86_reg x86_make_reg( enum x86_reg_file file, - enum x86_reg_name idx ); - -struct x86_reg x86_make_disp( struct x86_reg reg, - int disp ); - -struct x86_reg x86_deref( struct x86_reg reg ); - -struct x86_reg x86_get_base_reg( struct x86_reg reg ); - - -/* Labels, jumps and fixup: - */ -unsigned char *x86_get_label( struct x86_function *p ); - -void x86_jcc( struct x86_function *p, - enum x86_cc cc, - unsigned char *label ); - -unsigned char *x86_jcc_forward( struct x86_function *p, - enum x86_cc cc ); - -unsigned char *x86_jmp_forward( struct x86_function *p); - -unsigned char *x86_call_forward( struct x86_function *p); - -void x86_fixup_fwd_jump( struct x86_function *p, - unsigned char *fixup ); - -void x86_jmp( struct x86_function *p, unsigned char *label ); - -/* void x86_call( struct x86_function *p, void (*label)() ); */ -void x86_call( struct x86_function *p, struct x86_reg reg); - -/* michal: - * Temporary. As I need immediate operands, and dont want to mess with the codegen, - * I load the immediate into general purpose register and use it. - */ -void x86_mov_reg_imm( struct x86_function *p, struct x86_reg dst, int imm ); - - -/* Macro for sse_shufps() and sse2_pshufd(): - */ -#define SHUF(_x,_y,_z,_w) (((_x)<<0) | ((_y)<<2) | ((_z)<<4) | ((_w)<<6)) -#define SHUF_NOOP RSW(0,1,2,3) -#define GET_SHUF(swz, idx) (((swz) >> ((idx)*2)) & 0x3) - -void mmx_emms( struct x86_function *p ); -void mmx_movd( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void mmx_movq( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void mmx_packssdw( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void mmx_packuswb( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); - -void sse2_cvtps2dq( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse2_cvttps2dq( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse2_movd( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse2_packssdw( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse2_packsswb( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse2_packuswb( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse2_pshufd( struct x86_function *p, struct x86_reg dest, struct x86_reg arg0, - unsigned char shuf ); -void sse2_rcpps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse2_rcpss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); - -void sse_addps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_addss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_cvtps2pi( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_divss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_andnps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_andps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_cmpps( struct x86_function *p, struct x86_reg dst, struct x86_reg src, - unsigned char cc ); -void sse_maxps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_maxss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_minps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_movaps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_movhlps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_movhps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_movlhps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_movlps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_movss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_movups( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_mulps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_mulss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_orps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_xorps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_subps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_rsqrtps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_rsqrtss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void sse_shufps( struct x86_function *p, struct x86_reg dest, struct x86_reg arg0, - unsigned char shuf ); -void sse_pmovmskb( struct x86_function *p, struct x86_reg dest, struct x86_reg src ); - -void x86_add( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_and( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_cmp( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_dec( struct x86_function *p, struct x86_reg reg ); -void x86_inc( struct x86_function *p, struct x86_reg reg ); -void x86_lea( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_mov( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_mul( struct x86_function *p, struct x86_reg src ); -void x86_or( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_pop( struct x86_function *p, struct x86_reg reg ); -void x86_push( struct x86_function *p, struct x86_reg reg ); -void x86_ret( struct x86_function *p ); -void x86_sub( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_test( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_xor( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); -void x86_sahf( struct x86_function *p ); - -void x87_f2xm1( struct x86_function *p ); -void x87_fabs( struct x86_function *p ); -void x87_fadd( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); -void x87_faddp( struct x86_function *p, struct x86_reg dst ); -void x87_fchs( struct x86_function *p ); -void x87_fclex( struct x86_function *p ); -void x87_fcom( struct x86_function *p, struct x86_reg dst ); -void x87_fcomp( struct x86_function *p, struct x86_reg dst ); -void x87_fcos( struct x86_function *p ); -void x87_fdiv( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); -void x87_fdivp( struct x86_function *p, struct x86_reg dst ); -void x87_fdivr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); -void x87_fdivrp( struct x86_function *p, struct x86_reg dst ); -void x87_fild( struct x86_function *p, struct x86_reg arg ); -void x87_fist( struct x86_function *p, struct x86_reg dst ); -void x87_fistp( struct x86_function *p, struct x86_reg dst ); -void x87_fld( struct x86_function *p, struct x86_reg arg ); -void x87_fld1( struct x86_function *p ); -void x87_fldcw( struct x86_function *p, struct x86_reg arg ); -void x87_fldl2e( struct x86_function *p ); -void x87_fldln2( struct x86_function *p ); -void x87_fldz( struct x86_function *p ); -void x87_fmul( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); -void x87_fmulp( struct x86_function *p, struct x86_reg dst ); -void x87_fnclex( struct x86_function *p ); -void x87_fprndint( struct x86_function *p ); -void x87_fscale( struct x86_function *p ); -void x87_fsin( struct x86_function *p ); -void x87_fsincos( struct x86_function *p ); -void x87_fsqrt( struct x86_function *p ); -void x87_fst( struct x86_function *p, struct x86_reg dst ); -void x87_fstp( struct x86_function *p, struct x86_reg dst ); -void x87_fsub( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); -void x87_fsubp( struct x86_function *p, struct x86_reg dst ); -void x87_fsubr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); -void x87_fsubrp( struct x86_function *p, struct x86_reg dst ); -void x87_fxch( struct x86_function *p, struct x86_reg dst ); -void x87_fxtract( struct x86_function *p ); -void x87_fyl2x( struct x86_function *p ); -void x87_fyl2xp1( struct x86_function *p ); -void x87_fwait( struct x86_function *p ); -void x87_fnstsw( struct x86_function *p, struct x86_reg dst ); -void x87_fucompp( struct x86_function *p ); -void x87_fucomp( struct x86_function *p, struct x86_reg arg ); -void x87_fucom( struct x86_function *p, struct x86_reg arg ); - - - -/* Retreive a reference to one of the function arguments, taking into - * account any push/pop activity. Note - doesn't track explict - * manipulation of ESP by other instructions. - */ -struct x86_reg x86_fn_arg( struct x86_function *p, unsigned arg ); - -#endif -#endif + +#ifndef _X86SSE_H_ +#define _X86SSE_H_ + +#if defined(__i386__) || defined(__386__) + +/* It is up to the caller to ensure that instructions issued are + * suitable for the host cpu. There are no checks made in this module + * for mmx/sse/sse2 support on the cpu. + */ +struct x86_reg { + unsigned file:3; + unsigned idx:3; + unsigned mod:2; /* mod_REG if this is just a register */ + int disp:24; /* only +/- 23bits of offset - should be enough... */ +}; + +struct x86_function { + unsigned size; + unsigned char *store; + unsigned char *csr; + unsigned stack_offset; + int need_emms; + const char *fn; +}; + +enum x86_reg_file { + file_REG32, + file_MMX, + file_XMM, + file_x87 +}; + +/* Values for mod field of modr/m byte + */ +enum x86_reg_mod { + mod_INDIRECT, + mod_DISP8, + mod_DISP32, + mod_REG +}; + +enum x86_reg_name { + reg_AX, + reg_CX, + reg_DX, + reg_BX, + reg_SP, + reg_BP, + reg_SI, + reg_DI +}; + + +enum x86_cc { + cc_O, /* overflow */ + cc_NO, /* not overflow */ + cc_NAE, /* not above or equal / carry */ + cc_AE, /* above or equal / not carry */ + cc_E, /* equal / zero */ + cc_NE /* not equal / not zero */ +}; + +enum sse_cc { + cc_Equal, + cc_LessThan, + cc_LessThanEqual, + cc_Unordered, + cc_NotEqual, + cc_NotLessThan, + cc_NotLessThanEqual, + cc_Ordered +}; + +#define cc_Z cc_E +#define cc_NZ cc_NE + +/* Begin/end/retreive function creation: + */ + + +void x86_init_func( struct x86_function *p ); +int x86_init_func_size( struct x86_function *p, unsigned code_size ); +void x86_release_func( struct x86_function *p ); +void (*x86_get_func( struct x86_function *p ))( void ); + + + +/* Create and manipulate registers and regmem values: + */ +struct x86_reg x86_make_reg( enum x86_reg_file file, + enum x86_reg_name idx ); + +struct x86_reg x86_make_disp( struct x86_reg reg, + int disp ); + +struct x86_reg x86_deref( struct x86_reg reg ); + +struct x86_reg x86_get_base_reg( struct x86_reg reg ); + + +/* Labels, jumps and fixup: + */ +unsigned char *x86_get_label( struct x86_function *p ); + +void x86_jcc( struct x86_function *p, + enum x86_cc cc, + unsigned char *label ); + +unsigned char *x86_jcc_forward( struct x86_function *p, + enum x86_cc cc ); + +unsigned char *x86_jmp_forward( struct x86_function *p); + +unsigned char *x86_call_forward( struct x86_function *p); + +void x86_fixup_fwd_jump( struct x86_function *p, + unsigned char *fixup ); + +void x86_jmp( struct x86_function *p, unsigned char *label ); + +/* void x86_call( struct x86_function *p, void (*label)() ); */ +void x86_call( struct x86_function *p, struct x86_reg reg); + +/* michal: + * Temporary. As I need immediate operands, and dont want to mess with the codegen, + * I load the immediate into general purpose register and use it. + */ +void x86_mov_reg_imm( struct x86_function *p, struct x86_reg dst, int imm ); + + +/* Macro for sse_shufps() and sse2_pshufd(): + */ +#define SHUF(_x,_y,_z,_w) (((_x)<<0) | ((_y)<<2) | ((_z)<<4) | ((_w)<<6)) +#define SHUF_NOOP RSW(0,1,2,3) +#define GET_SHUF(swz, idx) (((swz) >> ((idx)*2)) & 0x3) + +void mmx_emms( struct x86_function *p ); +void mmx_movd( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void mmx_movq( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void mmx_packssdw( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void mmx_packuswb( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); + +void sse2_cvtps2dq( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse2_cvttps2dq( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse2_movd( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse2_packssdw( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse2_packsswb( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse2_packuswb( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse2_pshufd( struct x86_function *p, struct x86_reg dest, struct x86_reg arg0, + unsigned char shuf ); +void sse2_rcpps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse2_rcpss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); + +void sse_addps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_addss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_cvtps2pi( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_divss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_andnps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_andps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_cmpps( struct x86_function *p, struct x86_reg dst, struct x86_reg src, + unsigned char cc ); +void sse_maxps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_maxss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_minps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_movaps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_movhlps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_movhps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_movlhps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_movlps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_movss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_movups( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_mulps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_mulss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_orps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_xorps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_subps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_rsqrtps( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_rsqrtss( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void sse_shufps( struct x86_function *p, struct x86_reg dest, struct x86_reg arg0, + unsigned char shuf ); +void sse_pmovmskb( struct x86_function *p, struct x86_reg dest, struct x86_reg src ); + +void x86_add( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_and( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_cmp( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_dec( struct x86_function *p, struct x86_reg reg ); +void x86_inc( struct x86_function *p, struct x86_reg reg ); +void x86_lea( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_mov( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_mul( struct x86_function *p, struct x86_reg src ); +void x86_or( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_pop( struct x86_function *p, struct x86_reg reg ); +void x86_push( struct x86_function *p, struct x86_reg reg ); +void x86_ret( struct x86_function *p ); +void x86_sub( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_test( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_xor( struct x86_function *p, struct x86_reg dst, struct x86_reg src ); +void x86_sahf( struct x86_function *p ); + +void x87_f2xm1( struct x86_function *p ); +void x87_fabs( struct x86_function *p ); +void x87_fadd( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); +void x87_faddp( struct x86_function *p, struct x86_reg dst ); +void x87_fchs( struct x86_function *p ); +void x87_fclex( struct x86_function *p ); +void x87_fcom( struct x86_function *p, struct x86_reg dst ); +void x87_fcomp( struct x86_function *p, struct x86_reg dst ); +void x87_fcos( struct x86_function *p ); +void x87_fdiv( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); +void x87_fdivp( struct x86_function *p, struct x86_reg dst ); +void x87_fdivr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); +void x87_fdivrp( struct x86_function *p, struct x86_reg dst ); +void x87_fild( struct x86_function *p, struct x86_reg arg ); +void x87_fist( struct x86_function *p, struct x86_reg dst ); +void x87_fistp( struct x86_function *p, struct x86_reg dst ); +void x87_fld( struct x86_function *p, struct x86_reg arg ); +void x87_fld1( struct x86_function *p ); +void x87_fldcw( struct x86_function *p, struct x86_reg arg ); +void x87_fldl2e( struct x86_function *p ); +void x87_fldln2( struct x86_function *p ); +void x87_fldz( struct x86_function *p ); +void x87_fmul( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); +void x87_fmulp( struct x86_function *p, struct x86_reg dst ); +void x87_fnclex( struct x86_function *p ); +void x87_fprndint( struct x86_function *p ); +void x87_fscale( struct x86_function *p ); +void x87_fsin( struct x86_function *p ); +void x87_fsincos( struct x86_function *p ); +void x87_fsqrt( struct x86_function *p ); +void x87_fst( struct x86_function *p, struct x86_reg dst ); +void x87_fstp( struct x86_function *p, struct x86_reg dst ); +void x87_fsub( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); +void x87_fsubp( struct x86_function *p, struct x86_reg dst ); +void x87_fsubr( struct x86_function *p, struct x86_reg dst, struct x86_reg arg ); +void x87_fsubrp( struct x86_function *p, struct x86_reg dst ); +void x87_fxch( struct x86_function *p, struct x86_reg dst ); +void x87_fxtract( struct x86_function *p ); +void x87_fyl2x( struct x86_function *p ); +void x87_fyl2xp1( struct x86_function *p ); +void x87_fwait( struct x86_function *p ); +void x87_fnstsw( struct x86_function *p, struct x86_reg dst ); +void x87_fucompp( struct x86_function *p ); +void x87_fucomp( struct x86_function *p, struct x86_reg arg ); +void x87_fucom( struct x86_function *p, struct x86_reg arg ); + + + +/* Retreive a reference to one of the function arguments, taking into + * account any push/pop activity. Note - doesn't track explict + * manipulation of ESP by other instructions. + */ +struct x86_reg x86_fn_arg( struct x86_function *p, unsigned arg ); + +#endif +#endif diff --git a/mesalib/src/mesa/x86/sse.c b/mesalib/src/mesa/x86/sse.c index b249811c7..aef15b531 100644 --- a/mesalib/src/mesa/x86/sse.c +++ b/mesalib/src/mesa/x86/sse.c @@ -1,123 +1,123 @@ -/* - * Mesa 3-D graphics library - * Version: 6.0 - * - * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * PentiumIII-SIMD (SSE) optimizations contributed by - * Andre Werthmann - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "math/m_xform.h" -#include "tnl/t_context.h" - -#include "sse.h" -#include "x86_xform.h" - -#ifdef DEBUG_MATH -#include "math/m_debug.h" -#endif - - -#ifdef USE_SSE_ASM -DECLARE_XFORM_GROUP( sse, 2 ) -DECLARE_XFORM_GROUP( sse, 3 ) - -#if 1 -/* Some functions are not written in SSE-assembly, because the fpu ones are faster */ -extern void _ASMAPI _mesa_sse_transform_normals_no_rot( NORM_ARGS ); -extern void _ASMAPI _mesa_sse_transform_rescale_normals( NORM_ARGS ); -extern void _ASMAPI _mesa_sse_transform_rescale_normals_no_rot( NORM_ARGS ); - -extern void _ASMAPI _mesa_sse_transform_points4_general( XFORM_ARGS ); -extern void _ASMAPI _mesa_sse_transform_points4_3d( XFORM_ARGS ); -/* XXX this function segfaults, see below */ -extern void _ASMAPI _mesa_sse_transform_points4_identity( XFORM_ARGS ); -/* XXX this one works, see below */ -extern void _ASMAPI _mesa_x86_transform_points4_identity( XFORM_ARGS ); -#else -DECLARE_NORM_GROUP( sse ) -#endif - - -extern void _ASMAPI -_mesa_v16_sse_general_xform( GLfloat *first_vert, - const GLfloat *m, - const GLfloat *src, - GLuint src_stride, - GLuint count ); - -extern void _ASMAPI -_mesa_sse_project_vertices( GLfloat *first, - GLfloat *last, - const GLfloat *m, - GLuint stride ); - -extern void _ASMAPI -_mesa_sse_project_clipped_vertices( GLfloat *first, - GLfloat *last, - const GLfloat *m, - GLuint stride, - const GLubyte *clipmask ); -#endif - - -void _mesa_init_sse_transform_asm( void ) -{ -#ifdef USE_SSE_ASM - ASSIGN_XFORM_GROUP( sse, 2 ); - ASSIGN_XFORM_GROUP( sse, 3 ); - -#if 1 - /* TODO: Finish these off. - */ - _mesa_transform_tab[4][MATRIX_GENERAL] = - _mesa_sse_transform_points4_general; - _mesa_transform_tab[4][MATRIX_3D] = - _mesa_sse_transform_points4_3d; - /* XXX NOTE: _mesa_sse_transform_points4_identity segfaults with the - conformance tests, so use the x86 version. - */ - _mesa_transform_tab[4][MATRIX_IDENTITY] = - _mesa_x86_transform_points4_identity;/*_mesa_sse_transform_points4_identity;*/ - - _mesa_normal_tab[NORM_TRANSFORM_NO_ROT] = - _mesa_sse_transform_normals_no_rot; - _mesa_normal_tab[NORM_TRANSFORM | NORM_RESCALE] = - _mesa_sse_transform_rescale_normals; - _mesa_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE] = - _mesa_sse_transform_rescale_normals_no_rot; -#else - ASSIGN_XFORM_GROUP( sse, 4 ); - - ASSIGN_NORM_GROUP( sse ); -#endif - -#ifdef DEBUG_MATH - _math_test_all_transform_functions( "SSE" ); - _math_test_all_normal_transform_functions( "SSE" ); -#endif -#endif -} - +/* + * Mesa 3-D graphics library + * Version: 6.0 + * + * Copyright (C) 1999-2004 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * PentiumIII-SIMD (SSE) optimizations contributed by + * Andre Werthmann + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "math/m_xform.h" +#include "tnl/t_context.h" + +#include "sse.h" +#include "x86_xform.h" + +#ifdef DEBUG_MATH +#include "math/m_debug.h" +#endif + + +#ifdef USE_SSE_ASM +DECLARE_XFORM_GROUP( sse, 2 ) +DECLARE_XFORM_GROUP( sse, 3 ) + +#if 1 +/* Some functions are not written in SSE-assembly, because the fpu ones are faster */ +extern void _ASMAPI _mesa_sse_transform_normals_no_rot( NORM_ARGS ); +extern void _ASMAPI _mesa_sse_transform_rescale_normals( NORM_ARGS ); +extern void _ASMAPI _mesa_sse_transform_rescale_normals_no_rot( NORM_ARGS ); + +extern void _ASMAPI _mesa_sse_transform_points4_general( XFORM_ARGS ); +extern void _ASMAPI _mesa_sse_transform_points4_3d( XFORM_ARGS ); +/* XXX this function segfaults, see below */ +extern void _ASMAPI _mesa_sse_transform_points4_identity( XFORM_ARGS ); +/* XXX this one works, see below */ +extern void _ASMAPI _mesa_x86_transform_points4_identity( XFORM_ARGS ); +#else +DECLARE_NORM_GROUP( sse ) +#endif + + +extern void _ASMAPI +_mesa_v16_sse_general_xform( GLfloat *first_vert, + const GLfloat *m, + const GLfloat *src, + GLuint src_stride, + GLuint count ); + +extern void _ASMAPI +_mesa_sse_project_vertices( GLfloat *first, + GLfloat *last, + const GLfloat *m, + GLuint stride ); + +extern void _ASMAPI +_mesa_sse_project_clipped_vertices( GLfloat *first, + GLfloat *last, + const GLfloat *m, + GLuint stride, + const GLubyte *clipmask ); +#endif + + +void _mesa_init_sse_transform_asm( void ) +{ +#ifdef USE_SSE_ASM + ASSIGN_XFORM_GROUP( sse, 2 ); + ASSIGN_XFORM_GROUP( sse, 3 ); + +#if 1 + /* TODO: Finish these off. + */ + _mesa_transform_tab[4][MATRIX_GENERAL] = + _mesa_sse_transform_points4_general; + _mesa_transform_tab[4][MATRIX_3D] = + _mesa_sse_transform_points4_3d; + /* XXX NOTE: _mesa_sse_transform_points4_identity segfaults with the + conformance tests, so use the x86 version. + */ + _mesa_transform_tab[4][MATRIX_IDENTITY] = + _mesa_x86_transform_points4_identity;/*_mesa_sse_transform_points4_identity;*/ + + _mesa_normal_tab[NORM_TRANSFORM_NO_ROT] = + _mesa_sse_transform_normals_no_rot; + _mesa_normal_tab[NORM_TRANSFORM | NORM_RESCALE] = + _mesa_sse_transform_rescale_normals; + _mesa_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE] = + _mesa_sse_transform_rescale_normals_no_rot; +#else + ASSIGN_XFORM_GROUP( sse, 4 ); + + ASSIGN_NORM_GROUP( sse ); +#endif + +#ifdef DEBUG_MATH + _math_test_all_transform_functions( "SSE" ); + _math_test_all_normal_transform_functions( "SSE" ); +#endif +#endif +} + diff --git a/mesalib/src/mesa/x86/sse.h b/mesalib/src/mesa/x86/sse.h index d240bf38b..e92ddc139 100644 --- a/mesalib/src/mesa/x86/sse.h +++ b/mesalib/src/mesa/x86/sse.h @@ -1,36 +1,36 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * PentiumIII-SIMD (SSE) optimizations contributed by - * Andre Werthmann - */ - -#ifndef __SSE_H__ -#define __SSE_H__ - -void _mesa_init_sse_transform_asm( void ); - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * PentiumIII-SIMD (SSE) optimizations contributed by + * Andre Werthmann + */ + +#ifndef __SSE_H__ +#define __SSE_H__ + +void _mesa_init_sse_transform_asm( void ); + +#endif diff --git a/mesalib/src/mesa/x86/sse_normal.S b/mesalib/src/mesa/x86/sse_normal.S index b6d5ef965..a8c0d38c7 100644 --- a/mesalib/src/mesa/x86/sse_normal.S +++ b/mesalib/src/mesa/x86/sse_normal.S @@ -1,261 +1,261 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** TODO: - * - insert PREFETCH instructions to avoid cache-misses ! - * - some more optimizations are possible... - * - for 40-50% more performance in the SSE-functions, the - * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! - */ - -#ifdef USE_SSE_ASM -#include "assyntax.h" -#include "matypes.h" -#include "norm_args.h" - - SEG_TEXT - -#define M(i) REGOFF(i * 4, EDX) -#define S(i) REGOFF(i * 4, ESI) -#define D(i) REGOFF(i * 4, EDI) -#define STRIDE REGOFF(12, ESI) - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_sse_transform_rescale_normals_no_rot) -HIDDEN(_mesa_sse_transform_rescale_normals_no_rot) -GLNAME(_mesa_sse_transform_rescale_normals_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L ( ARG_IN, ESI ) /* ptr to source GLvector3f */ - MOV_L ( ARG_DEST, EDI ) /* ptr to dest GLvector3f */ - - MOV_L ( ARG_MAT, EDX ) /* ptr to matrix */ - MOV_L ( REGOFF(MATRIX_INV, EDX), EDX) /* matrix->inv */ - - MOV_L ( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L ( ECX, ECX ) - JZ( LLBL(K_G3TRNNRR_finish) ) /* count was zero; go to finish */ - - MOV_L ( STRIDE, EAX ) /* stride */ - MOV_L ( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest-count */ - - IMUL_L( CONST(16), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM1 ) /* m0 */ - MOVSS ( M(5), XMM2 ) /* m5 */ - UNPCKLPS( XMM2, XMM1 ) /* m5 | m0 */ - MOVSS ( ARG_SCALE, XMM0 ) /* scale */ - SHUFPS ( CONST(0x0), XMM0, XMM0 ) /* scale | scale */ - MULPS ( XMM0, XMM1 ) /* m5*scale | m0*scale */ - MULSS ( M(10), XMM0 ) /* m10*scale */ - -ALIGNTEXT32 -LLBL(K_G3TRNNRR_top): - MOVLPS ( S(0), XMM2 ) /* uy | ux */ - MULPS ( XMM1, XMM2 ) /* uy*m5*scale | ux*m0*scale */ - MOVLPS ( XMM2, D(0) ) /* ->D(1) | D(0) */ - - MOVSS ( S(2), XMM2 ) /* uz */ - MULSS ( XMM0, XMM2 ) /* uz*m10*scale */ - MOVSS ( XMM2, D(2) ) /* ->D(2) */ - -LLBL(K_G3TRNNRR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_G3TRNNRR_top) ) - -LLBL(K_G3TRNNRR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_sse_transform_rescale_normals) -HIDDEN(_mesa_sse_transform_rescale_normals) -GLNAME(_mesa_sse_transform_rescale_normals): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L ( ARG_IN, ESI ) /* ptr to source GLvector3f */ - MOV_L ( ARG_DEST, EDI ) /* ptr to dest GLvector3f */ - - MOV_L ( ARG_MAT, EDX ) /* ptr to matrix */ - MOV_L ( REGOFF(MATRIX_INV, EDX), EDX) /* matrix->inv */ - - MOV_L ( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L ( ECX, ECX ) - JZ( LLBL(K_G3TRNR_finish) ) /* count was zero; go to finish */ - - MOV_L ( STRIDE, EAX ) /* stride */ - MOV_L ( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest-count */ - - IMUL_L( CONST(16), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM0 ) /* m0 */ - MOVSS ( M(4), XMM1 ) /* m4 */ - UNPCKLPS( XMM1, XMM0 ) /* m4 | m0 */ - - MOVSS ( ARG_SCALE, XMM4 ) /* scale */ - SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* scale | scale */ - - MULPS ( XMM4, XMM0 ) /* m4*scale | m0*scale */ - MOVSS ( M(1), XMM1 ) /* m1 */ - MOVSS ( M(5), XMM2 ) /* m5 */ - UNPCKLPS( XMM2, XMM1 ) /* m5 | m1 */ - MULPS ( XMM4, XMM1 ) /* m5*scale | m1*scale */ - MOVSS ( M(2), XMM2 ) /* m2 */ - MOVSS ( M(6), XMM3 ) /* m6 */ - UNPCKLPS( XMM3, XMM2 ) /* m6 | m2 */ - MULPS ( XMM4, XMM2 ) /* m6*scale | m2*scale */ - - MOVSS ( M(8), XMM6 ) /* m8 */ - MULSS ( ARG_SCALE, XMM6 ) /* m8*scale */ - MOVSS ( M(9), XMM7 ) /* m9 */ - MULSS ( ARG_SCALE, XMM7 ) /* m9*scale */ - -ALIGNTEXT32 -LLBL(K_G3TRNR_top): - MOVSS ( S(0), XMM3 ) /* ux */ - SHUFPS ( CONST(0x0), XMM3, XMM3 ) /* ux | ux */ - MULPS ( XMM0, XMM3 ) /* ux*m4 | ux*m0 */ - MOVSS ( S(1), XMM4 ) /* uy */ - SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* uy | uy */ - MULPS ( XMM1, XMM4 ) /* uy*m5 | uy*m1 */ - MOVSS ( S(2), XMM5 ) /* uz */ - SHUFPS ( CONST(0x0), XMM5, XMM5 ) /* uz | uz */ - MULPS ( XMM2, XMM5 ) /* uz*m6 | uz*m2 */ - - ADDPS ( XMM4, XMM3 ) - ADDPS ( XMM5, XMM3 ) - MOVLPS ( XMM3, D(0) ) - - MOVSS ( M(10), XMM3 ) /* m10 */ - MULSS ( ARG_SCALE, XMM3 ) /* m10*scale */ - MULSS ( S(2), XMM3 ) /* m10*scale*uz */ - MOVSS ( S(1), XMM4 ) /* uy */ - MULSS ( XMM7, XMM4 ) /* uy*m9*scale */ - MOVSS ( S(0), XMM5 ) /* ux */ - MULSS ( XMM6, XMM5 ) /* ux*m8*scale */ - - ADDSS ( XMM4, XMM3 ) - ADDSS ( XMM5, XMM3 ) - MOVSS ( XMM3, D(2) ) - -LLBL(K_G3TRNR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_G3TRNR_top) ) - -LLBL(K_G3TRNR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_sse_transform_normals_no_rot) -HIDDEN(_mesa_sse_transform_normals_no_rot) -GLNAME(_mesa_sse_transform_normals_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L ( ARG_IN, ESI ) /* ptr to source GLvector3f */ - MOV_L ( ARG_DEST, EDI ) /* ptr to dest GLvector3f */ - - MOV_L ( ARG_MAT, EDX ) /* ptr to matrix */ - MOV_L ( REGOFF(MATRIX_INV, EDX), EDX) /* matrix->inv */ - - MOV_L ( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L ( ECX, ECX ) - JZ( LLBL(K_G3TNNRR_finish) ) /* count was zero; go to finish */ - - MOV_L ( STRIDE, EAX ) /* stride */ - MOV_L ( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest-count */ - - IMUL_L( CONST(16), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS( M(0), XMM0 ) /* m0 */ - MOVSS( M(5), XMM1 ) /* m5 */ - UNPCKLPS( XMM1, XMM0 ) /* m5 | m0 */ - MOVSS( M(10), XMM1 ) /* m10 */ - -ALIGNTEXT32 -LLBL(K_G3TNNRR_top): - MOVLPS( S(0), XMM2 ) /* uy | ux */ - MULPS( XMM0, XMM2 ) /* uy*m5 | ux*m0 */ - MOVLPS( XMM2, D(0) ) - - MOVSS( S(2), XMM2 ) /* uz */ - MULSS( XMM1, XMM2 ) /* uz*m10 */ - MOVSS( XMM2, D(2) ) - -LLBL(K_G3TNNRR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_G3TNNRR_top) ) - -LLBL(K_G3TNNRR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** TODO: + * - insert PREFETCH instructions to avoid cache-misses ! + * - some more optimizations are possible... + * - for 40-50% more performance in the SSE-functions, the + * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! + */ + +#ifdef USE_SSE_ASM +#include "assyntax.h" +#include "matypes.h" +#include "norm_args.h" + + SEG_TEXT + +#define M(i) REGOFF(i * 4, EDX) +#define S(i) REGOFF(i * 4, ESI) +#define D(i) REGOFF(i * 4, EDI) +#define STRIDE REGOFF(12, ESI) + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_sse_transform_rescale_normals_no_rot) +HIDDEN(_mesa_sse_transform_rescale_normals_no_rot) +GLNAME(_mesa_sse_transform_rescale_normals_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L ( ARG_IN, ESI ) /* ptr to source GLvector3f */ + MOV_L ( ARG_DEST, EDI ) /* ptr to dest GLvector3f */ + + MOV_L ( ARG_MAT, EDX ) /* ptr to matrix */ + MOV_L ( REGOFF(MATRIX_INV, EDX), EDX) /* matrix->inv */ + + MOV_L ( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L ( ECX, ECX ) + JZ( LLBL(K_G3TRNNRR_finish) ) /* count was zero; go to finish */ + + MOV_L ( STRIDE, EAX ) /* stride */ + MOV_L ( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest-count */ + + IMUL_L( CONST(16), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM1 ) /* m0 */ + MOVSS ( M(5), XMM2 ) /* m5 */ + UNPCKLPS( XMM2, XMM1 ) /* m5 | m0 */ + MOVSS ( ARG_SCALE, XMM0 ) /* scale */ + SHUFPS ( CONST(0x0), XMM0, XMM0 ) /* scale | scale */ + MULPS ( XMM0, XMM1 ) /* m5*scale | m0*scale */ + MULSS ( M(10), XMM0 ) /* m10*scale */ + +ALIGNTEXT32 +LLBL(K_G3TRNNRR_top): + MOVLPS ( S(0), XMM2 ) /* uy | ux */ + MULPS ( XMM1, XMM2 ) /* uy*m5*scale | ux*m0*scale */ + MOVLPS ( XMM2, D(0) ) /* ->D(1) | D(0) */ + + MOVSS ( S(2), XMM2 ) /* uz */ + MULSS ( XMM0, XMM2 ) /* uz*m10*scale */ + MOVSS ( XMM2, D(2) ) /* ->D(2) */ + +LLBL(K_G3TRNNRR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_G3TRNNRR_top) ) + +LLBL(K_G3TRNNRR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_sse_transform_rescale_normals) +HIDDEN(_mesa_sse_transform_rescale_normals) +GLNAME(_mesa_sse_transform_rescale_normals): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L ( ARG_IN, ESI ) /* ptr to source GLvector3f */ + MOV_L ( ARG_DEST, EDI ) /* ptr to dest GLvector3f */ + + MOV_L ( ARG_MAT, EDX ) /* ptr to matrix */ + MOV_L ( REGOFF(MATRIX_INV, EDX), EDX) /* matrix->inv */ + + MOV_L ( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L ( ECX, ECX ) + JZ( LLBL(K_G3TRNR_finish) ) /* count was zero; go to finish */ + + MOV_L ( STRIDE, EAX ) /* stride */ + MOV_L ( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest-count */ + + IMUL_L( CONST(16), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM0 ) /* m0 */ + MOVSS ( M(4), XMM1 ) /* m4 */ + UNPCKLPS( XMM1, XMM0 ) /* m4 | m0 */ + + MOVSS ( ARG_SCALE, XMM4 ) /* scale */ + SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* scale | scale */ + + MULPS ( XMM4, XMM0 ) /* m4*scale | m0*scale */ + MOVSS ( M(1), XMM1 ) /* m1 */ + MOVSS ( M(5), XMM2 ) /* m5 */ + UNPCKLPS( XMM2, XMM1 ) /* m5 | m1 */ + MULPS ( XMM4, XMM1 ) /* m5*scale | m1*scale */ + MOVSS ( M(2), XMM2 ) /* m2 */ + MOVSS ( M(6), XMM3 ) /* m6 */ + UNPCKLPS( XMM3, XMM2 ) /* m6 | m2 */ + MULPS ( XMM4, XMM2 ) /* m6*scale | m2*scale */ + + MOVSS ( M(8), XMM6 ) /* m8 */ + MULSS ( ARG_SCALE, XMM6 ) /* m8*scale */ + MOVSS ( M(9), XMM7 ) /* m9 */ + MULSS ( ARG_SCALE, XMM7 ) /* m9*scale */ + +ALIGNTEXT32 +LLBL(K_G3TRNR_top): + MOVSS ( S(0), XMM3 ) /* ux */ + SHUFPS ( CONST(0x0), XMM3, XMM3 ) /* ux | ux */ + MULPS ( XMM0, XMM3 ) /* ux*m4 | ux*m0 */ + MOVSS ( S(1), XMM4 ) /* uy */ + SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* uy | uy */ + MULPS ( XMM1, XMM4 ) /* uy*m5 | uy*m1 */ + MOVSS ( S(2), XMM5 ) /* uz */ + SHUFPS ( CONST(0x0), XMM5, XMM5 ) /* uz | uz */ + MULPS ( XMM2, XMM5 ) /* uz*m6 | uz*m2 */ + + ADDPS ( XMM4, XMM3 ) + ADDPS ( XMM5, XMM3 ) + MOVLPS ( XMM3, D(0) ) + + MOVSS ( M(10), XMM3 ) /* m10 */ + MULSS ( ARG_SCALE, XMM3 ) /* m10*scale */ + MULSS ( S(2), XMM3 ) /* m10*scale*uz */ + MOVSS ( S(1), XMM4 ) /* uy */ + MULSS ( XMM7, XMM4 ) /* uy*m9*scale */ + MOVSS ( S(0), XMM5 ) /* ux */ + MULSS ( XMM6, XMM5 ) /* ux*m8*scale */ + + ADDSS ( XMM4, XMM3 ) + ADDSS ( XMM5, XMM3 ) + MOVSS ( XMM3, D(2) ) + +LLBL(K_G3TRNR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_G3TRNR_top) ) + +LLBL(K_G3TRNR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_sse_transform_normals_no_rot) +HIDDEN(_mesa_sse_transform_normals_no_rot) +GLNAME(_mesa_sse_transform_normals_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L ( ARG_IN, ESI ) /* ptr to source GLvector3f */ + MOV_L ( ARG_DEST, EDI ) /* ptr to dest GLvector3f */ + + MOV_L ( ARG_MAT, EDX ) /* ptr to matrix */ + MOV_L ( REGOFF(MATRIX_INV, EDX), EDX) /* matrix->inv */ + + MOV_L ( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L ( ECX, ECX ) + JZ( LLBL(K_G3TNNRR_finish) ) /* count was zero; go to finish */ + + MOV_L ( STRIDE, EAX ) /* stride */ + MOV_L ( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest-count */ + + IMUL_L( CONST(16), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS( M(0), XMM0 ) /* m0 */ + MOVSS( M(5), XMM1 ) /* m5 */ + UNPCKLPS( XMM1, XMM0 ) /* m5 | m0 */ + MOVSS( M(10), XMM1 ) /* m10 */ + +ALIGNTEXT32 +LLBL(K_G3TNNRR_top): + MOVLPS( S(0), XMM2 ) /* uy | ux */ + MULPS( XMM0, XMM2 ) /* uy*m5 | ux*m0 */ + MOVLPS( XMM2, D(0) ) + + MOVSS( S(2), XMM2 ) /* uz */ + MULSS( XMM1, XMM2 ) /* uz*m10 */ + MOVSS( XMM2, D(2) ) + +LLBL(K_G3TNNRR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_G3TNNRR_top) ) + +LLBL(K_G3TNNRR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/sse_xform1.S b/mesalib/src/mesa/x86/sse_xform1.S index 45a6dd7ec..4aa9de607 100644 --- a/mesalib/src/mesa/x86/sse_xform1.S +++ b/mesalib/src/mesa/x86/sse_xform1.S @@ -1,446 +1,446 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** TODO: - * - insert PREFETCH instructions to avoid cache-misses ! - * - some more optimizations are possible... - * - for 40-50% more performance in the SSE-functions, the - * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! - */ - -#ifdef USE_SSE_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define S(i) REGOFF(i * 4, ESI) -#define D(i) REGOFF(i * 4, EDI) -#define M(i) REGOFF(i * 4, EDX) - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points1_general) -HIDDEN( _mesa_sse_transform_points1_general ) -GLNAME( _mesa_sse_transform_points1_general ): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - CMP_L( CONST(0), ECX ) /* count == 0 ? */ - JE( LLBL(K_GTP1GR_finish) ) /* yes -> nothing to do. */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - -ALIGNTEXT32 - MOVAPS( M(0), XMM0 ) /* m3 | m2 | m1 | m0 */ - MOVAPS( M(12), XMM1 ) /* m15 | m14 | m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP1GR_top): - MOVSS( S(0), XMM2 ) /* ox */ - SHUFPS( CONST(0x0), XMM2, XMM2 ) /* ox | ox | ox | ox */ - MULPS( XMM0, XMM2 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ - ADDPS( XMM1, XMM2 ) /* + | + | + | + */ - MOVUPS( XMM2, D(0) ) - -LLBL(K_GTP1GR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP1GR_top) ) - -LLBL(K_GTP1GR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points1_identity) -HIDDEN(_mesa_sse_transform_points1_identity) -GLNAME( _mesa_sse_transform_points1_identity ): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP1IR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_1), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(1), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - CMP_L( ESI, EDI ) - JE( LLBL(K_GTP1IR_finish) ) - - -ALIGNTEXT32 -LLBL(K_GTP1IR_top): - MOV_L( S(0), EDX ) - MOV_L( EDX, D(0) ) - -LLBL(K_GTP1IR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP1IR_top) ) - -LLBL(K_GTP1IR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points1_3d_no_rot) -HIDDEN(_mesa_sse_transform_points1_3d_no_rot) -GLNAME(_mesa_sse_transform_points1_3d_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP13DNRR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - -ALIGNTEXT32 - MOVSS( M(0), XMM0 ) /* m0 */ - MOVSS( M(12), XMM1 ) /* m12 */ - MOVSS( M(13), XMM2 ) /* m13 */ - MOVSS( M(14), XMM3 ) /* m14 */ - -ALIGNTEXT32 -LLBL(K_GTP13DNRR_top): - MOVSS( S(0), XMM4 ) /* ox */ - MULSS( XMM0, XMM4 ) /* ox*m0 */ - ADDSS( XMM1, XMM4 ) /* ox*m0+m12 */ - MOVSS( XMM4, D(0) ) - - MOVSS( XMM2, D(1) ) - MOVSS( XMM3, D(2) ) - -LLBL(K_GTP13DNRR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP13DNRR_top) ) - -LLBL(K_GTP13DNRR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points1_perspective) -HIDDEN(_mesa_sse_transform_points1_perspective) -GLNAME(_mesa_sse_transform_points1_perspective): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP13PR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - -ALIGNTEXT32 - XORPS( XMM0, XMM0 ) /* 0 | 0 | 0 | 0 */ - MOVSS( M(0), XMM1 ) /* m0 */ - MOVSS( M(14), XMM2 ) /* m14 */ - -ALIGNTEXT32 -LLBL(K_GTP13PR_top): - MOVSS( S(0), XMM3 ) /* ox */ - MULSS( XMM1, XMM3 ) /* ox*m0 */ - MOVSS( XMM3, D(0) ) /* ox*m0->D(0) */ - MOVSS( XMM2, D(2) ) /* m14->D(2) */ - - MOVSS( XMM0, D(1) ) - MOVSS( XMM0, D(3) ) - -LLBL(K_GTP13PR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP13PR_top) ) - -LLBL(K_GTP13PR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points1_2d) -HIDDEN(_mesa_sse_transform_points1_2d) -GLNAME(_mesa_sse_transform_points1_2d): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP13P2DR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVLPS( M(0), XMM0 ) /* m1 | m0 */ - MOVLPS( M(12), XMM1 ) /* m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP13P2DR_top): - MOVSS( S(0), XMM2 ) /* ox */ - SHUFPS( CONST(0x0), XMM2, XMM2 ) /* ox | ox | ox | ox */ - MULPS( XMM0, XMM2 ) /* - | - | ox*m1 | ox*m0 */ - ADDPS( XMM1, XMM2 ) /* - | - | ox*m1+m13 | ox*m0+m12 */ - MOVLPS( XMM2, D(0) ) - -LLBL(K_GTP13P2DR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP13P2DR_top) ) - -LLBL(K_GTP13P2DR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points1_2d_no_rot) -HIDDEN(_mesa_sse_transform_points1_2d_no_rot) -GLNAME(_mesa_sse_transform_points1_2d_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP13P2DNRR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS( M(0), XMM0 ) /* m0 */ - MOVSS( M(12), XMM1 ) /* m12 */ - MOVSS( M(13), XMM2 ) /* m13 */ - -ALIGNTEXT32 -LLBL(K_GTP13P2DNRR_top): - MOVSS( S(0), XMM3 ) /* ox */ - MULSS( XMM0, XMM3 ) /* ox*m0 */ - ADDSS( XMM1, XMM3 ) /* ox*m0+m12 */ - MOVSS( XMM3, D(0) ) - MOVSS( XMM2, D(1) ) - -LLBL(K_GTP13P2DNRR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP13P2DNRR_top) ) - -LLBL(K_GTP13P2DNRR_finish): - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points1_3d) -HIDDEN(_mesa_sse_transform_points1_3d) -GLNAME(_mesa_sse_transform_points1_3d): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP13P3DR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - -ALIGNTEXT32 - MOVAPS( M(0), XMM0 ) /* m3 | m2 | m1 | m0 */ - MOVAPS( M(12), XMM1 ) /* m15 | m14 | m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP13P3DR_top): - MOVSS( S(0), XMM2 ) /* ox */ - SHUFPS( CONST(0x0), XMM2, XMM2 ) /* ox | ox | ox | ox */ - MULPS( XMM0, XMM2 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ - ADDPS( XMM1, XMM2 ) /* +m15 | +m14 | +m13 | +m12 */ - MOVLPS( XMM2, D(0) ) /* - | - | ->D(1)| ->D(0)*/ - UNPCKHPS( XMM2, XMM2 ) /* ox*m3+m15 | ox*m3+m15 | ox*m2+m14 | ox*m2+m14 */ - MOVSS( XMM2, D(2) ) - -LLBL(K_GTP13P3DR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP13P3DR_top) ) - -LLBL(K_GTP13P3DR_finish): - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** TODO: + * - insert PREFETCH instructions to avoid cache-misses ! + * - some more optimizations are possible... + * - for 40-50% more performance in the SSE-functions, the + * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! + */ + +#ifdef USE_SSE_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define S(i) REGOFF(i * 4, ESI) +#define D(i) REGOFF(i * 4, EDI) +#define M(i) REGOFF(i * 4, EDX) + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points1_general) +HIDDEN( _mesa_sse_transform_points1_general ) +GLNAME( _mesa_sse_transform_points1_general ): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + CMP_L( CONST(0), ECX ) /* count == 0 ? */ + JE( LLBL(K_GTP1GR_finish) ) /* yes -> nothing to do. */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + +ALIGNTEXT32 + MOVAPS( M(0), XMM0 ) /* m3 | m2 | m1 | m0 */ + MOVAPS( M(12), XMM1 ) /* m15 | m14 | m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP1GR_top): + MOVSS( S(0), XMM2 ) /* ox */ + SHUFPS( CONST(0x0), XMM2, XMM2 ) /* ox | ox | ox | ox */ + MULPS( XMM0, XMM2 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ + ADDPS( XMM1, XMM2 ) /* + | + | + | + */ + MOVUPS( XMM2, D(0) ) + +LLBL(K_GTP1GR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP1GR_top) ) + +LLBL(K_GTP1GR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points1_identity) +HIDDEN(_mesa_sse_transform_points1_identity) +GLNAME( _mesa_sse_transform_points1_identity ): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP1IR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_1), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(1), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + CMP_L( ESI, EDI ) + JE( LLBL(K_GTP1IR_finish) ) + + +ALIGNTEXT32 +LLBL(K_GTP1IR_top): + MOV_L( S(0), EDX ) + MOV_L( EDX, D(0) ) + +LLBL(K_GTP1IR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP1IR_top) ) + +LLBL(K_GTP1IR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points1_3d_no_rot) +HIDDEN(_mesa_sse_transform_points1_3d_no_rot) +GLNAME(_mesa_sse_transform_points1_3d_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP13DNRR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + +ALIGNTEXT32 + MOVSS( M(0), XMM0 ) /* m0 */ + MOVSS( M(12), XMM1 ) /* m12 */ + MOVSS( M(13), XMM2 ) /* m13 */ + MOVSS( M(14), XMM3 ) /* m14 */ + +ALIGNTEXT32 +LLBL(K_GTP13DNRR_top): + MOVSS( S(0), XMM4 ) /* ox */ + MULSS( XMM0, XMM4 ) /* ox*m0 */ + ADDSS( XMM1, XMM4 ) /* ox*m0+m12 */ + MOVSS( XMM4, D(0) ) + + MOVSS( XMM2, D(1) ) + MOVSS( XMM3, D(2) ) + +LLBL(K_GTP13DNRR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP13DNRR_top) ) + +LLBL(K_GTP13DNRR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points1_perspective) +HIDDEN(_mesa_sse_transform_points1_perspective) +GLNAME(_mesa_sse_transform_points1_perspective): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP13PR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + +ALIGNTEXT32 + XORPS( XMM0, XMM0 ) /* 0 | 0 | 0 | 0 */ + MOVSS( M(0), XMM1 ) /* m0 */ + MOVSS( M(14), XMM2 ) /* m14 */ + +ALIGNTEXT32 +LLBL(K_GTP13PR_top): + MOVSS( S(0), XMM3 ) /* ox */ + MULSS( XMM1, XMM3 ) /* ox*m0 */ + MOVSS( XMM3, D(0) ) /* ox*m0->D(0) */ + MOVSS( XMM2, D(2) ) /* m14->D(2) */ + + MOVSS( XMM0, D(1) ) + MOVSS( XMM0, D(3) ) + +LLBL(K_GTP13PR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP13PR_top) ) + +LLBL(K_GTP13PR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points1_2d) +HIDDEN(_mesa_sse_transform_points1_2d) +GLNAME(_mesa_sse_transform_points1_2d): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP13P2DR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVLPS( M(0), XMM0 ) /* m1 | m0 */ + MOVLPS( M(12), XMM1 ) /* m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP13P2DR_top): + MOVSS( S(0), XMM2 ) /* ox */ + SHUFPS( CONST(0x0), XMM2, XMM2 ) /* ox | ox | ox | ox */ + MULPS( XMM0, XMM2 ) /* - | - | ox*m1 | ox*m0 */ + ADDPS( XMM1, XMM2 ) /* - | - | ox*m1+m13 | ox*m0+m12 */ + MOVLPS( XMM2, D(0) ) + +LLBL(K_GTP13P2DR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP13P2DR_top) ) + +LLBL(K_GTP13P2DR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points1_2d_no_rot) +HIDDEN(_mesa_sse_transform_points1_2d_no_rot) +GLNAME(_mesa_sse_transform_points1_2d_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP13P2DNRR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS( M(0), XMM0 ) /* m0 */ + MOVSS( M(12), XMM1 ) /* m12 */ + MOVSS( M(13), XMM2 ) /* m13 */ + +ALIGNTEXT32 +LLBL(K_GTP13P2DNRR_top): + MOVSS( S(0), XMM3 ) /* ox */ + MULSS( XMM0, XMM3 ) /* ox*m0 */ + ADDSS( XMM1, XMM3 ) /* ox*m0+m12 */ + MOVSS( XMM3, D(0) ) + MOVSS( XMM2, D(1) ) + +LLBL(K_GTP13P2DNRR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP13P2DNRR_top) ) + +LLBL(K_GTP13P2DNRR_finish): + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points1_3d) +HIDDEN(_mesa_sse_transform_points1_3d) +GLNAME(_mesa_sse_transform_points1_3d): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP13P3DR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + +ALIGNTEXT32 + MOVAPS( M(0), XMM0 ) /* m3 | m2 | m1 | m0 */ + MOVAPS( M(12), XMM1 ) /* m15 | m14 | m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP13P3DR_top): + MOVSS( S(0), XMM2 ) /* ox */ + SHUFPS( CONST(0x0), XMM2, XMM2 ) /* ox | ox | ox | ox */ + MULPS( XMM0, XMM2 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ + ADDPS( XMM1, XMM2 ) /* +m15 | +m14 | +m13 | +m12 */ + MOVLPS( XMM2, D(0) ) /* - | - | ->D(1)| ->D(0)*/ + UNPCKHPS( XMM2, XMM2 ) /* ox*m3+m15 | ox*m3+m15 | ox*m2+m14 | ox*m2+m14 */ + MOVSS( XMM2, D(2) ) + +LLBL(K_GTP13P3DR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP13P3DR_top) ) + +LLBL(K_GTP13P3DR_finish): + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/sse_xform2.S b/mesalib/src/mesa/x86/sse_xform2.S index 7fb52e2e7..a443dad35 100644 --- a/mesalib/src/mesa/x86/sse_xform2.S +++ b/mesalib/src/mesa/x86/sse_xform2.S @@ -1,466 +1,466 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** TODO: - * - insert PREFETCH instructions to avoid cache-misses ! - * - some more optimizations are possible... - * - for 40-50% more performance in the SSE-functions, the - * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! - */ - -#ifdef USE_SSE_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define S(i) REGOFF(i * 4, ESI) -#define D(i) REGOFF(i * 4, EDI) -#define M(i) REGOFF(i * 4, EDX) - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points2_general) -HIDDEN (_mesa_sse_transform_points2_general) -GLNAME( _mesa_sse_transform_points2_general ): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX ) - JZ( LLBL(K_GTP2GR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVAPS( M(0), XMM0 ) /* m3 | m2 | m1 | m0 */ - MOVAPS( M(4), XMM1 ) /* m7 | m6 | m5 | m4 */ - MOVAPS( M(12), XMM2 ) /* m15 | m14 | m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP2GR_top): - MOVSS( S(0), XMM3 ) /* ox */ - SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ox | ox | ox | ox */ - MULPS( XMM0, XMM3 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ - MOVSS( S(1), XMM4 ) /* oy */ - SHUFPS( CONST(0x0), XMM4, XMM4 ) /* oy | oy | oy | oy */ - MULPS( XMM1, XMM4 ) /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */ - - ADDPS( XMM4, XMM3 ) - ADDPS( XMM2, XMM3 ) - MOVAPS( XMM3, D(0) ) - -LLBL(K_GTP2GR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP2GR_top) ) - -LLBL(K_GTP2GR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points2_identity) -HIDDEN(_mesa_sse_transform_points2_identity) -GLNAME( _mesa_sse_transform_points2_identity ): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP2IR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - CMP_L( ESI, EDI ) - JE( LLBL(K_GTP2IR_finish) ) - - -ALIGNTEXT32 -LLBL(K_GTP2IR_top): - MOV_L ( S(0), EDX ) - MOV_L ( EDX, D(0) ) - MOV_L ( S(1), EDX ) - MOV_L ( EDX, D(1) ) - -LLBL(K_GTP2IR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP2IR_top) ) - -LLBL(K_GTP2IR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points2_3d_no_rot) -HIDDEN(_mesa_sse_transform_points2_3d_no_rot) -GLNAME(_mesa_sse_transform_points2_3d_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP23DNRR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - XORPS( XMM0, XMM0 ) /* clean the working register */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ - MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ - UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ - MOVLPS ( M(12), XMM2 ) /* - | - | m13 | m12 */ - MOVSS ( M(14), XMM3 ) /* - | - | - | m14 */ - -ALIGNTEXT32 -LLBL(K_GTP23DNRR_top): - MOVLPS ( S(0), XMM0 ) /* - | - | oy | ox */ - MULPS ( XMM1, XMM0 ) /* - | - | oy*m5 | ox*m0 */ - ADDPS ( XMM2, XMM0 ) /* - | - | +m13 | +m12 */ - MOVLPS ( XMM0, D(0) ) /* -> D(1) | -> D(0) */ - - MOVSS ( XMM3, D(2) ) /* -> D(2) */ - -LLBL(K_GTP23DNRR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP23DNRR_top) ) - -LLBL(K_GTP23DNRR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points2_perspective) -HIDDEN(_mesa_sse_transform_points2_perspective) -GLNAME(_mesa_sse_transform_points2_perspective): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP23PR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ - MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ - UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ - MOVSS ( M(14), XMM3 ) /* m14 */ - XORPS ( XMM0, XMM0 ) /* 0 | 0 | 0 | 0 */ - -ALIGNTEXT32 -LLBL(K_GTP23PR_top): - MOVLPS( S(0), XMM4 ) /* oy | ox */ - MULPS( XMM1, XMM4 ) /* oy*m5 | ox*m0 */ - MOVLPS( XMM4, D(0) ) /* ->D(1) | ->D(0) */ - MOVSS( XMM3, D(2) ) /* ->D(2) */ - MOVSS( XMM0, D(3) ) /* ->D(3) */ - -LLBL(K_GTP23PR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP23PR_top) ) - -LLBL(K_GTP23PR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points2_2d) -HIDDEN(_mesa_sse_transform_points2_2d) -GLNAME(_mesa_sse_transform_points2_2d): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP23P2DR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVLPS( M(0), XMM0 ) /* m1 | m0 */ - MOVLPS( M(4), XMM1 ) /* m5 | m4 */ - MOVLPS( M(12), XMM2 ) /* m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP23P2DR_top): - MOVSS( S(0), XMM3 ) /* ox */ - SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ox | ox */ - MULPS( XMM0, XMM3 ) /* ox*m1 | ox*m0 */ - - MOVSS( S(1), XMM4 ) /* oy */ - SHUFPS( CONST(0x0), XMM4, XMM4 ) /* oy | oy */ - MULPS( XMM1, XMM4 ) /* oy*m5 | oy*m4 */ - - ADDPS( XMM4, XMM3 ) - ADDPS( XMM2, XMM3 ) - MOVLPS( XMM3, D(0) ) /* ->D(1) | ->D(0) */ - -LLBL(K_GTP23P2DR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP23P2DR_top) ) - -LLBL(K_GTP23P2DR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points2_2d_no_rot) -HIDDEN(_mesa_sse_transform_points2_2d_no_rot) -GLNAME(_mesa_sse_transform_points2_2d_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP23P2DNRR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM1 ) /* m0 */ - MOVSS ( M(5), XMM2 ) /* m5 */ - UNPCKLPS ( XMM2, XMM1 ) /* m5 | m0 */ - MOVLPS ( M(12), XMM2 ) /* m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP23P2DNRR_top): - MOVLPS( S(0), XMM0 ) /* oy | ox */ - MULPS( XMM1, XMM0 ) /* oy*m5 | ox*m0 */ - ADDPS( XMM2, XMM0 ) /* +m13 | +m12 */ - MOVLPS( XMM0, D(0) ) /* ->D(1) | ->D(0) */ - -LLBL(K_GTP23P2DNRR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP23P2DNRR_top) ) - -LLBL(K_GTP23P2DNRR_finish): - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points2_3d) -HIDDEN(_mesa_sse_transform_points2_3d) -GLNAME(_mesa_sse_transform_points2_3d): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP23P3DR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVAPS( M(0), XMM0 ) /* m2 | m1 | m0 */ - MOVAPS( M(4), XMM1 ) /* m6 | m5 | m4 */ - MOVAPS( M(12), XMM2 ) /* m14 | m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP23P3DR_top): - MOVSS( S(0), XMM3 ) /* ox */ - SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ox | ox | ox */ - MULPS( XMM0, XMM3 ) /* ox*m2 | ox*m1 | ox*m0 */ - - MOVSS( S(1), XMM4 ) /* oy */ - SHUFPS( CONST(0x0), XMM4, XMM4 ) /* oy | oy | oy */ - MULPS( XMM1, XMM4 ) /* oy*m6 | oy*m5 | oy*m4 */ - - ADDPS( XMM4, XMM3 ) - ADDPS( XMM2, XMM3 ) - - MOVLPS( XMM3, D(0) ) /* ->D(1) | ->D(0) */ - UNPCKHPS( XMM3, XMM3 ) - MOVSS( XMM3, D(2) ) /* ->D(2) */ - -LLBL(K_GTP23P3DR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP23P3DR_top) ) - -LLBL(K_GTP23P3DR_finish): - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** TODO: + * - insert PREFETCH instructions to avoid cache-misses ! + * - some more optimizations are possible... + * - for 40-50% more performance in the SSE-functions, the + * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! + */ + +#ifdef USE_SSE_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define S(i) REGOFF(i * 4, ESI) +#define D(i) REGOFF(i * 4, EDI) +#define M(i) REGOFF(i * 4, EDX) + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points2_general) +HIDDEN (_mesa_sse_transform_points2_general) +GLNAME( _mesa_sse_transform_points2_general ): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX ) + JZ( LLBL(K_GTP2GR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVAPS( M(0), XMM0 ) /* m3 | m2 | m1 | m0 */ + MOVAPS( M(4), XMM1 ) /* m7 | m6 | m5 | m4 */ + MOVAPS( M(12), XMM2 ) /* m15 | m14 | m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP2GR_top): + MOVSS( S(0), XMM3 ) /* ox */ + SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ox | ox | ox | ox */ + MULPS( XMM0, XMM3 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ + MOVSS( S(1), XMM4 ) /* oy */ + SHUFPS( CONST(0x0), XMM4, XMM4 ) /* oy | oy | oy | oy */ + MULPS( XMM1, XMM4 ) /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */ + + ADDPS( XMM4, XMM3 ) + ADDPS( XMM2, XMM3 ) + MOVAPS( XMM3, D(0) ) + +LLBL(K_GTP2GR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP2GR_top) ) + +LLBL(K_GTP2GR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points2_identity) +HIDDEN(_mesa_sse_transform_points2_identity) +GLNAME( _mesa_sse_transform_points2_identity ): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP2IR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + CMP_L( ESI, EDI ) + JE( LLBL(K_GTP2IR_finish) ) + + +ALIGNTEXT32 +LLBL(K_GTP2IR_top): + MOV_L ( S(0), EDX ) + MOV_L ( EDX, D(0) ) + MOV_L ( S(1), EDX ) + MOV_L ( EDX, D(1) ) + +LLBL(K_GTP2IR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP2IR_top) ) + +LLBL(K_GTP2IR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points2_3d_no_rot) +HIDDEN(_mesa_sse_transform_points2_3d_no_rot) +GLNAME(_mesa_sse_transform_points2_3d_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP23DNRR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + XORPS( XMM0, XMM0 ) /* clean the working register */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ + MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ + UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ + MOVLPS ( M(12), XMM2 ) /* - | - | m13 | m12 */ + MOVSS ( M(14), XMM3 ) /* - | - | - | m14 */ + +ALIGNTEXT32 +LLBL(K_GTP23DNRR_top): + MOVLPS ( S(0), XMM0 ) /* - | - | oy | ox */ + MULPS ( XMM1, XMM0 ) /* - | - | oy*m5 | ox*m0 */ + ADDPS ( XMM2, XMM0 ) /* - | - | +m13 | +m12 */ + MOVLPS ( XMM0, D(0) ) /* -> D(1) | -> D(0) */ + + MOVSS ( XMM3, D(2) ) /* -> D(2) */ + +LLBL(K_GTP23DNRR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP23DNRR_top) ) + +LLBL(K_GTP23DNRR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points2_perspective) +HIDDEN(_mesa_sse_transform_points2_perspective) +GLNAME(_mesa_sse_transform_points2_perspective): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP23PR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ + MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ + UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ + MOVSS ( M(14), XMM3 ) /* m14 */ + XORPS ( XMM0, XMM0 ) /* 0 | 0 | 0 | 0 */ + +ALIGNTEXT32 +LLBL(K_GTP23PR_top): + MOVLPS( S(0), XMM4 ) /* oy | ox */ + MULPS( XMM1, XMM4 ) /* oy*m5 | ox*m0 */ + MOVLPS( XMM4, D(0) ) /* ->D(1) | ->D(0) */ + MOVSS( XMM3, D(2) ) /* ->D(2) */ + MOVSS( XMM0, D(3) ) /* ->D(3) */ + +LLBL(K_GTP23PR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP23PR_top) ) + +LLBL(K_GTP23PR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points2_2d) +HIDDEN(_mesa_sse_transform_points2_2d) +GLNAME(_mesa_sse_transform_points2_2d): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP23P2DR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVLPS( M(0), XMM0 ) /* m1 | m0 */ + MOVLPS( M(4), XMM1 ) /* m5 | m4 */ + MOVLPS( M(12), XMM2 ) /* m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP23P2DR_top): + MOVSS( S(0), XMM3 ) /* ox */ + SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ox | ox */ + MULPS( XMM0, XMM3 ) /* ox*m1 | ox*m0 */ + + MOVSS( S(1), XMM4 ) /* oy */ + SHUFPS( CONST(0x0), XMM4, XMM4 ) /* oy | oy */ + MULPS( XMM1, XMM4 ) /* oy*m5 | oy*m4 */ + + ADDPS( XMM4, XMM3 ) + ADDPS( XMM2, XMM3 ) + MOVLPS( XMM3, D(0) ) /* ->D(1) | ->D(0) */ + +LLBL(K_GTP23P2DR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP23P2DR_top) ) + +LLBL(K_GTP23P2DR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points2_2d_no_rot) +HIDDEN(_mesa_sse_transform_points2_2d_no_rot) +GLNAME(_mesa_sse_transform_points2_2d_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP23P2DNRR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM1 ) /* m0 */ + MOVSS ( M(5), XMM2 ) /* m5 */ + UNPCKLPS ( XMM2, XMM1 ) /* m5 | m0 */ + MOVLPS ( M(12), XMM2 ) /* m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP23P2DNRR_top): + MOVLPS( S(0), XMM0 ) /* oy | ox */ + MULPS( XMM1, XMM0 ) /* oy*m5 | ox*m0 */ + ADDPS( XMM2, XMM0 ) /* +m13 | +m12 */ + MOVLPS( XMM0, D(0) ) /* ->D(1) | ->D(0) */ + +LLBL(K_GTP23P2DNRR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP23P2DNRR_top) ) + +LLBL(K_GTP23P2DNRR_finish): + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points2_3d) +HIDDEN(_mesa_sse_transform_points2_3d) +GLNAME(_mesa_sse_transform_points2_3d): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP23P3DR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVAPS( M(0), XMM0 ) /* m2 | m1 | m0 */ + MOVAPS( M(4), XMM1 ) /* m6 | m5 | m4 */ + MOVAPS( M(12), XMM2 ) /* m14 | m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP23P3DR_top): + MOVSS( S(0), XMM3 ) /* ox */ + SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ox | ox | ox */ + MULPS( XMM0, XMM3 ) /* ox*m2 | ox*m1 | ox*m0 */ + + MOVSS( S(1), XMM4 ) /* oy */ + SHUFPS( CONST(0x0), XMM4, XMM4 ) /* oy | oy | oy */ + MULPS( XMM1, XMM4 ) /* oy*m6 | oy*m5 | oy*m4 */ + + ADDPS( XMM4, XMM3 ) + ADDPS( XMM2, XMM3 ) + + MOVLPS( XMM3, D(0) ) /* ->D(1) | ->D(0) */ + UNPCKHPS( XMM3, XMM3 ) + MOVSS( XMM3, D(2) ) /* ->D(2) */ + +LLBL(K_GTP23P3DR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP23P3DR_top) ) + +LLBL(K_GTP23P3DR_finish): + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/sse_xform3.S b/mesalib/src/mesa/x86/sse_xform3.S index 4e39312bf..4bc22d8a5 100644 --- a/mesalib/src/mesa/x86/sse_xform3.S +++ b/mesalib/src/mesa/x86/sse_xform3.S @@ -1,512 +1,512 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/** TODO: - * - insert PREFETCH instructions to avoid cache-misses ! - * - some more optimizations are possible... - * - for 40-50% more performance in the SSE-functions, the - * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! - */ - -#ifdef USE_SSE_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define S(i) REGOFF(i * 4, ESI) -#define D(i) REGOFF(i * 4, EDI) -#define M(i) REGOFF(i * 4, EDX) - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points3_general) -HIDDEN(_mesa_sse_transform_points3_general) -GLNAME( _mesa_sse_transform_points3_general ): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - CMP_L ( CONST(0), ECX ) /* count == 0 ? */ - JE ( LLBL(K_GTPGR_finish) ) /* yes -> nothing to do. */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - -ALIGNTEXT32 - MOVAPS ( REGOFF(0, EDX), XMM0 ) /* m0 | m1 | m2 | m3 */ - MOVAPS ( REGOFF(16, EDX), XMM1 ) /* m4 | m5 | m6 | m7 */ - MOVAPS ( REGOFF(32, EDX), XMM2 ) /* m8 | m9 | m10 | m11 */ - MOVAPS ( REGOFF(48, EDX), XMM3 ) /* m12 | m13 | m14 | m15 */ - - -ALIGNTEXT32 -LLBL(K_GTPGR_top): - MOVSS ( REGOFF(0, ESI), XMM4 ) /* | | | ox */ - SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* ox | ox | ox | ox */ - MOVSS ( REGOFF(4, ESI), XMM5 ) /* | | | oy */ - SHUFPS ( CONST(0x0), XMM5, XMM5 ) /* oy | oy | oy | oy */ - MOVSS ( REGOFF(8, ESI), XMM6 ) /* | | | oz */ - SHUFPS ( CONST(0x0), XMM6, XMM6 ) /* oz | oz | oz | oz */ - - MULPS ( XMM0, XMM4 ) /* m3*ox | m2*ox | m1*ox | m0*ox */ - MULPS ( XMM1, XMM5 ) /* m7*oy | m6*oy | m5*oy | m4*oy */ - MULPS ( XMM2, XMM6 ) /* m11*oz | m10*oz | m9*oz | m8*oz */ - - ADDPS ( XMM5, XMM4 ) - ADDPS ( XMM6, XMM4 ) - ADDPS ( XMM3, XMM4 ) - - MOVAPS ( XMM4, REGOFF(0, EDI) ) - -LLBL(K_GTPGR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTPGR_top) ) - -LLBL(K_GTPGR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points3_identity) -HIDDEN(_mesa_sse_transform_points3_identity) -GLNAME( _mesa_sse_transform_points3_identity ): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTPIR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - CMP_L( ESI, EDI ) - JE( LLBL(K_GTPIR_finish) ) - - -ALIGNTEXT32 -LLBL(K_GTPIR_top): - MOVLPS ( S(0), XMM0 ) - MOVLPS ( XMM0, D(0) ) - MOVSS ( S(2), XMM0 ) - MOVSS ( XMM0, D(2) ) - -LLBL(K_GTPIR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTPIR_top) ) - -LLBL(K_GTPIR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points3_3d_no_rot) -HIDDEN(_mesa_sse_transform_points3_3d_no_rot) -GLNAME(_mesa_sse_transform_points3_3d_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP3DNRR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - XORPS( XMM0, XMM0 ) /* clean the working register */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ - MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ - UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ - MOVLPS ( M(12), XMM2 ) /* - | - | m13 | m12 */ - MOVSS ( M(10), XMM3 ) /* - | - | - | m10 */ - MOVSS ( M(14), XMM4 ) /* - | - | - | m14 */ - -ALIGNTEXT32 -LLBL(K_GTP3DNRR_top): - - MOVLPS ( S(0), XMM0 ) /* - | - | s1 | s0 */ - MULPS ( XMM1, XMM0 ) /* - | - | s1*m5 | s0*m0 */ - ADDPS ( XMM2, XMM0 ) /* - | - | +m13 | +m12 */ - MOVLPS ( XMM0, D(0) ) /* -> D(1) | -> D(0) */ - - MOVSS ( S(2), XMM0 ) /* sz */ - MULSS ( XMM3, XMM0 ) /* sz*m10 */ - ADDSS ( XMM4, XMM0 ) /* +m14 */ - MOVSS ( XMM0, D(2) ) /* -> D(2) */ - -LLBL(K_GTP3DNRR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP3DNRR_top) ) - -LLBL(K_GTP3DNRR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points3_perspective) -HIDDEN(_mesa_sse_transform_points3_perspective) -GLNAME(_mesa_sse_transform_points3_perspective): - -#define FRAME_OFFSET 8 - PUSH_L ( ESI ) - PUSH_L ( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP3PR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ - MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ - UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ - MOVLPS ( M(8), XMM2 ) /* - | - | m9 | m8 */ - MOVSS ( M(10), XMM3 ) /* m10 */ - MOVSS ( M(14), XMM4 ) /* m14 */ - XORPS ( XMM6, XMM6 ) /* 0 */ - -ALIGNTEXT32 -LLBL(K_GTP3PR_top): - MOVLPS ( S(0), XMM0 ) /* oy | ox */ - MULPS ( XMM1, XMM0 ) /* oy*m5 | ox*m0 */ - MOVSS ( S(2), XMM5 ) /* oz */ - SHUFPS ( CONST(0x0), XMM5, XMM5 ) /* oz | oz */ - MULPS ( XMM2, XMM5 ) /* oz*m9 | oz*m8 */ - ADDPS ( XMM5, XMM0 ) /* +oy*m5 | +ox*m0 */ - MOVLPS ( XMM0, D(0) ) /* ->D(1) | ->D(0) */ - - MOVSS ( S(2), XMM0 ) /* oz */ - MULSS ( XMM3, XMM0 ) /* oz*m10 */ - ADDSS ( XMM4, XMM0 ) /* +m14 */ - MOVSS ( XMM0, D(2) ) /* ->D(2) */ - - MOVSS ( S(2), XMM0 ) /* oz */ - MOVSS ( XMM6, XMM5 ) /* 0 */ - SUBPS ( XMM0, XMM5 ) /* -oz */ - MOVSS ( XMM5, D(3) ) /* ->D(3) */ - -LLBL(K_GTP3PR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP3PR_top) ) - -LLBL(K_GTP3PR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points3_2d) -HIDDEN(_mesa_sse_transform_points3_2d) -GLNAME(_mesa_sse_transform_points3_2d): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP3P2DR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVLPS( M(0), XMM0 ) /* m1 | m0 */ - MOVLPS( M(4), XMM1 ) /* m5 | m4 */ - MOVLPS( M(12), XMM2 ) /* m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP3P2DR_top): - MOVSS ( S(0), XMM3 ) /* ox */ - SHUFPS ( CONST(0x0), XMM3, XMM3 ) /* ox | ox */ - MULPS ( XMM0, XMM3 ) /* ox*m1 | ox*m0 */ - MOVSS ( S(1), XMM4 ) /* oy */ - SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* oy | oy */ - MULPS ( XMM1, XMM4 ) /* oy*m5 | oy*m4 */ - - ADDPS ( XMM4, XMM3 ) - ADDPS ( XMM2, XMM3 ) - MOVLPS ( XMM3, D(0) ) - - MOVSS ( S(2), XMM3 ) - MOVSS ( XMM3, D(2) ) - -LLBL(K_GTP3P2DR_skip): - ADD_L ( CONST(16), EDI ) - ADD_L ( EAX, ESI ) - CMP_L ( ECX, EDI ) - JNE ( LLBL(K_GTP3P2DR_top) ) - -LLBL(K_GTP3P2DR_finish): - POP_L ( EDI ) - POP_L ( ESI ) - RET -#undef FRAME_OFFSET - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points3_2d_no_rot) -HIDDEN(_mesa_sse_transform_points3_2d_no_rot) -GLNAME(_mesa_sse_transform_points3_2d_no_rot): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP3P2DNRR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - -ALIGNTEXT32 - MOVSS ( M(0), XMM1 ) /* m0 */ - MOVSS ( M(5), XMM2 ) /* m5 */ - UNPCKLPS ( XMM2, XMM1 ) /* m5 | m0 */ - MOVLPS ( M(12), XMM2 ) /* m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP3P2DNRR_top): - MOVLPS( S(0), XMM0 ) /* oy | ox */ - MULPS( XMM1, XMM0 ) /* oy*m5 | ox*m0 */ - ADDPS( XMM2, XMM0 ) /* +m13 | +m12 */ - MOVLPS( XMM0, D(0) ) /* ->D(1) | ->D(0) */ - - MOVSS( S(2), XMM0 ) - MOVSS( XMM0, D(2) ) - -LLBL(K_GTP3P2DNRR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP3P2DNRR_top) ) - -LLBL(K_GTP3P2DNRR_finish): - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT4 -GLOBL GLNAME(_mesa_sse_transform_points3_3d) -HIDDEN(_mesa_sse_transform_points3_3d) -GLNAME(_mesa_sse_transform_points3_3d): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ - MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ - - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP3P3DR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - -ALIGNTEXT32 - MOVAPS( M(0), XMM0 ) /* m2 | m1 | m0 */ - MOVAPS( M(4), XMM1 ) /* m6 | m5 | m4 */ - MOVAPS( M(8), XMM2 ) /* m10 | m9 | m8 */ - MOVAPS( M(12), XMM3 ) /* m14 | m13 | m12 */ - -ALIGNTEXT32 -LLBL(K_GTP3P3DR_top): - MOVSS( S(0), XMM4 ) - SHUFPS( CONST(0x0), XMM4, XMM4 ) /* ox | ox | ox */ - MULPS( XMM0, XMM4 ) /* ox*m2 | ox*m1 | ox*m0 */ - - MOVSS( S(1), XMM5 ) - SHUFPS( CONST(0x0), XMM5, XMM5 ) /* oy | oy | oy */ - MULPS( XMM1, XMM5 ) /* oy*m6 | oy*m5 | oy*m4 */ - - MOVSS( S(2), XMM6 ) - SHUFPS( CONST(0x0), XMM6, XMM6 ) /* oz | oz | oz */ - MULPS( XMM2, XMM6 ) /* oz*m10 | oz*m9 | oz*m8 */ - - ADDPS( XMM5, XMM4 ) /* + | + | + */ - ADDPS( XMM6, XMM4 ) /* + | + | + */ - ADDPS( XMM3, XMM4 ) /* + | + | + */ - - MOVLPS( XMM4, D(0) ) /* => D(1) | => D(0) */ - UNPCKHPS( XMM4, XMM4 ) - MOVSS( XMM4, D(2) ) - -LLBL(K_GTP3P3DR_skip): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP3P3DR_top) ) - -LLBL(K_GTP3P3DR_finish): - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/** TODO: + * - insert PREFETCH instructions to avoid cache-misses ! + * - some more optimizations are possible... + * - for 40-50% more performance in the SSE-functions, the + * data (trans-matrix, src_vert, dst_vert) needs to be 16byte aligned ! + */ + +#ifdef USE_SSE_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define S(i) REGOFF(i * 4, ESI) +#define D(i) REGOFF(i * 4, EDI) +#define M(i) REGOFF(i * 4, EDX) + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points3_general) +HIDDEN(_mesa_sse_transform_points3_general) +GLNAME( _mesa_sse_transform_points3_general ): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + CMP_L ( CONST(0), ECX ) /* count == 0 ? */ + JE ( LLBL(K_GTPGR_finish) ) /* yes -> nothing to do. */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + +ALIGNTEXT32 + MOVAPS ( REGOFF(0, EDX), XMM0 ) /* m0 | m1 | m2 | m3 */ + MOVAPS ( REGOFF(16, EDX), XMM1 ) /* m4 | m5 | m6 | m7 */ + MOVAPS ( REGOFF(32, EDX), XMM2 ) /* m8 | m9 | m10 | m11 */ + MOVAPS ( REGOFF(48, EDX), XMM3 ) /* m12 | m13 | m14 | m15 */ + + +ALIGNTEXT32 +LLBL(K_GTPGR_top): + MOVSS ( REGOFF(0, ESI), XMM4 ) /* | | | ox */ + SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* ox | ox | ox | ox */ + MOVSS ( REGOFF(4, ESI), XMM5 ) /* | | | oy */ + SHUFPS ( CONST(0x0), XMM5, XMM5 ) /* oy | oy | oy | oy */ + MOVSS ( REGOFF(8, ESI), XMM6 ) /* | | | oz */ + SHUFPS ( CONST(0x0), XMM6, XMM6 ) /* oz | oz | oz | oz */ + + MULPS ( XMM0, XMM4 ) /* m3*ox | m2*ox | m1*ox | m0*ox */ + MULPS ( XMM1, XMM5 ) /* m7*oy | m6*oy | m5*oy | m4*oy */ + MULPS ( XMM2, XMM6 ) /* m11*oz | m10*oz | m9*oz | m8*oz */ + + ADDPS ( XMM5, XMM4 ) + ADDPS ( XMM6, XMM4 ) + ADDPS ( XMM3, XMM4 ) + + MOVAPS ( XMM4, REGOFF(0, EDI) ) + +LLBL(K_GTPGR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTPGR_top) ) + +LLBL(K_GTPGR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points3_identity) +HIDDEN(_mesa_sse_transform_points3_identity) +GLNAME( _mesa_sse_transform_points3_identity ): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTPIR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + CMP_L( ESI, EDI ) + JE( LLBL(K_GTPIR_finish) ) + + +ALIGNTEXT32 +LLBL(K_GTPIR_top): + MOVLPS ( S(0), XMM0 ) + MOVLPS ( XMM0, D(0) ) + MOVSS ( S(2), XMM0 ) + MOVSS ( XMM0, D(2) ) + +LLBL(K_GTPIR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTPIR_top) ) + +LLBL(K_GTPIR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points3_3d_no_rot) +HIDDEN(_mesa_sse_transform_points3_3d_no_rot) +GLNAME(_mesa_sse_transform_points3_3d_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP3DNRR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + XORPS( XMM0, XMM0 ) /* clean the working register */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ + MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ + UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ + MOVLPS ( M(12), XMM2 ) /* - | - | m13 | m12 */ + MOVSS ( M(10), XMM3 ) /* - | - | - | m10 */ + MOVSS ( M(14), XMM4 ) /* - | - | - | m14 */ + +ALIGNTEXT32 +LLBL(K_GTP3DNRR_top): + + MOVLPS ( S(0), XMM0 ) /* - | - | s1 | s0 */ + MULPS ( XMM1, XMM0 ) /* - | - | s1*m5 | s0*m0 */ + ADDPS ( XMM2, XMM0 ) /* - | - | +m13 | +m12 */ + MOVLPS ( XMM0, D(0) ) /* -> D(1) | -> D(0) */ + + MOVSS ( S(2), XMM0 ) /* sz */ + MULSS ( XMM3, XMM0 ) /* sz*m10 */ + ADDSS ( XMM4, XMM0 ) /* +m14 */ + MOVSS ( XMM0, D(2) ) /* -> D(2) */ + +LLBL(K_GTP3DNRR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP3DNRR_top) ) + +LLBL(K_GTP3DNRR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points3_perspective) +HIDDEN(_mesa_sse_transform_points3_perspective) +GLNAME(_mesa_sse_transform_points3_perspective): + +#define FRAME_OFFSET 8 + PUSH_L ( ESI ) + PUSH_L ( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP3PR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM1 ) /* - | - | - | m0 */ + MOVSS ( M(5), XMM2 ) /* - | - | - | m5 */ + UNPCKLPS ( XMM2, XMM1 ) /* - | - | m5 | m0 */ + MOVLPS ( M(8), XMM2 ) /* - | - | m9 | m8 */ + MOVSS ( M(10), XMM3 ) /* m10 */ + MOVSS ( M(14), XMM4 ) /* m14 */ + XORPS ( XMM6, XMM6 ) /* 0 */ + +ALIGNTEXT32 +LLBL(K_GTP3PR_top): + MOVLPS ( S(0), XMM0 ) /* oy | ox */ + MULPS ( XMM1, XMM0 ) /* oy*m5 | ox*m0 */ + MOVSS ( S(2), XMM5 ) /* oz */ + SHUFPS ( CONST(0x0), XMM5, XMM5 ) /* oz | oz */ + MULPS ( XMM2, XMM5 ) /* oz*m9 | oz*m8 */ + ADDPS ( XMM5, XMM0 ) /* +oy*m5 | +ox*m0 */ + MOVLPS ( XMM0, D(0) ) /* ->D(1) | ->D(0) */ + + MOVSS ( S(2), XMM0 ) /* oz */ + MULSS ( XMM3, XMM0 ) /* oz*m10 */ + ADDSS ( XMM4, XMM0 ) /* +m14 */ + MOVSS ( XMM0, D(2) ) /* ->D(2) */ + + MOVSS ( S(2), XMM0 ) /* oz */ + MOVSS ( XMM6, XMM5 ) /* 0 */ + SUBPS ( XMM0, XMM5 ) /* -oz */ + MOVSS ( XMM5, D(3) ) /* ->D(3) */ + +LLBL(K_GTP3PR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP3PR_top) ) + +LLBL(K_GTP3PR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points3_2d) +HIDDEN(_mesa_sse_transform_points3_2d) +GLNAME(_mesa_sse_transform_points3_2d): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP3P2DR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVLPS( M(0), XMM0 ) /* m1 | m0 */ + MOVLPS( M(4), XMM1 ) /* m5 | m4 */ + MOVLPS( M(12), XMM2 ) /* m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP3P2DR_top): + MOVSS ( S(0), XMM3 ) /* ox */ + SHUFPS ( CONST(0x0), XMM3, XMM3 ) /* ox | ox */ + MULPS ( XMM0, XMM3 ) /* ox*m1 | ox*m0 */ + MOVSS ( S(1), XMM4 ) /* oy */ + SHUFPS ( CONST(0x0), XMM4, XMM4 ) /* oy | oy */ + MULPS ( XMM1, XMM4 ) /* oy*m5 | oy*m4 */ + + ADDPS ( XMM4, XMM3 ) + ADDPS ( XMM2, XMM3 ) + MOVLPS ( XMM3, D(0) ) + + MOVSS ( S(2), XMM3 ) + MOVSS ( XMM3, D(2) ) + +LLBL(K_GTP3P2DR_skip): + ADD_L ( CONST(16), EDI ) + ADD_L ( EAX, ESI ) + CMP_L ( ECX, EDI ) + JNE ( LLBL(K_GTP3P2DR_top) ) + +LLBL(K_GTP3P2DR_finish): + POP_L ( EDI ) + POP_L ( ESI ) + RET +#undef FRAME_OFFSET + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points3_2d_no_rot) +HIDDEN(_mesa_sse_transform_points3_2d_no_rot) +GLNAME(_mesa_sse_transform_points3_2d_no_rot): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP3P2DNRR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + +ALIGNTEXT32 + MOVSS ( M(0), XMM1 ) /* m0 */ + MOVSS ( M(5), XMM2 ) /* m5 */ + UNPCKLPS ( XMM2, XMM1 ) /* m5 | m0 */ + MOVLPS ( M(12), XMM2 ) /* m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP3P2DNRR_top): + MOVLPS( S(0), XMM0 ) /* oy | ox */ + MULPS( XMM1, XMM0 ) /* oy*m5 | ox*m0 */ + ADDPS( XMM2, XMM0 ) /* +m13 | +m12 */ + MOVLPS( XMM0, D(0) ) /* ->D(1) | ->D(0) */ + + MOVSS( S(2), XMM0 ) + MOVSS( XMM0, D(2) ) + +LLBL(K_GTP3P2DNRR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP3P2DNRR_top) ) + +LLBL(K_GTP3P2DNRR_finish): + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT4 +GLOBL GLNAME(_mesa_sse_transform_points3_3d) +HIDDEN(_mesa_sse_transform_points3_3d) +GLNAME(_mesa_sse_transform_points3_3d): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( REGOFF(OFFSET_SOURCE+8, ESP), ESI ) /* ptr to source GLvector4f */ + MOV_L( REGOFF(OFFSET_DEST+8, ESP), EDI ) /* ptr to dest GLvector4f */ + + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP3P3DR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) /* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + +ALIGNTEXT32 + MOVAPS( M(0), XMM0 ) /* m2 | m1 | m0 */ + MOVAPS( M(4), XMM1 ) /* m6 | m5 | m4 */ + MOVAPS( M(8), XMM2 ) /* m10 | m9 | m8 */ + MOVAPS( M(12), XMM3 ) /* m14 | m13 | m12 */ + +ALIGNTEXT32 +LLBL(K_GTP3P3DR_top): + MOVSS( S(0), XMM4 ) + SHUFPS( CONST(0x0), XMM4, XMM4 ) /* ox | ox | ox */ + MULPS( XMM0, XMM4 ) /* ox*m2 | ox*m1 | ox*m0 */ + + MOVSS( S(1), XMM5 ) + SHUFPS( CONST(0x0), XMM5, XMM5 ) /* oy | oy | oy */ + MULPS( XMM1, XMM5 ) /* oy*m6 | oy*m5 | oy*m4 */ + + MOVSS( S(2), XMM6 ) + SHUFPS( CONST(0x0), XMM6, XMM6 ) /* oz | oz | oz */ + MULPS( XMM2, XMM6 ) /* oz*m10 | oz*m9 | oz*m8 */ + + ADDPS( XMM5, XMM4 ) /* + | + | + */ + ADDPS( XMM6, XMM4 ) /* + | + | + */ + ADDPS( XMM3, XMM4 ) /* + | + | + */ + + MOVLPS( XMM4, D(0) ) /* => D(1) | => D(0) */ + UNPCKHPS( XMM4, XMM4 ) + MOVSS( XMM4, D(2) ) + +LLBL(K_GTP3P3DR_skip): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP3P3DR_top) ) + +LLBL(K_GTP3P3DR_finish): + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/sse_xform4.S b/mesalib/src/mesa/x86/sse_xform4.S index 75e0ff396..fb1fa741c 100644 --- a/mesalib/src/mesa/x86/sse_xform4.S +++ b/mesalib/src/mesa/x86/sse_xform4.S @@ -1,235 +1,235 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -#ifdef USE_SSE_ASM -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FRAME_OFFSET 8 - -#define SRC(i) REGOFF(i * 4, ESI) -#define DST(i) REGOFF(i * 4, EDI) -#define MAT(i) REGOFF(i * 4, EDX) - -#define SELECT(r0, r1, r2, r3) CONST( r0 * 64 + r1 * 16 + r2 * 4 + r3 ) - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_sse_transform_points4_general ) -HIDDEN(_mesa_sse_transform_points4_general) -GLNAME( _mesa_sse_transform_points4_general ): - - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) /* verify non-zero count */ - JE( LLBL( sse_general_done ) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )/* set dest size */ - - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - - PREFETCHT0( REGIND(ESI) ) - - MOVAPS( MAT(0), XMM4 ) /* m3 | m2 | m1 | m0 */ - MOVAPS( MAT(4), XMM5 ) /* m7 | m6 | m5 | m4 */ - MOVAPS( MAT(8), XMM6 ) /* m11 | m10 | m9 | m8 */ - MOVAPS( MAT(12), XMM7 ) /* m15 | m14 | m13 | m12 */ - -ALIGNTEXT16 -LLBL( sse_general_loop ): - - MOVSS( SRC(0), XMM0 ) /* ox */ - SHUFPS( CONST(0x0), XMM0, XMM0 ) /* ox | ox | ox | ox */ - MULPS( XMM4, XMM0 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ - - MOVSS( SRC(1), XMM1 ) /* oy */ - SHUFPS( CONST(0x0), XMM1, XMM1 ) /* oy | oy | oy | oy */ - MULPS( XMM5, XMM1 ) /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */ - - MOVSS( SRC(2), XMM2 ) /* oz */ - SHUFPS( CONST(0x0), XMM2, XMM2 ) /* oz | oz | oz | oz */ - MULPS( XMM6, XMM2 ) /* oz*m11 | oz*m10 | oz*m9 | oz*m8 */ - - MOVSS( SRC(3), XMM3 ) /* ow */ - SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ow | ow | ow | ow */ - MULPS( XMM7, XMM3 ) /* ow*m15 | ow*m14 | ow*m13 | ow*m12 */ - - ADDPS( XMM1, XMM0 ) /* ox*m3+oy*m7 | ... */ - ADDPS( XMM2, XMM0 ) /* ox*m3+oy*m7+oz*m11 | ... */ - ADDPS( XMM3, XMM0 ) /* ox*m3+oy*m7+oz*m11+ow*m15 | ... */ - MOVAPS( XMM0, DST(0) ) /* ->D(3) | ->D(2) | ->D(1) | ->D(0) */ - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - - DEC_L( ECX ) - JNZ( LLBL( sse_general_loop ) ) - -LLBL( sse_general_done ): - - POP_L( EDI ) - POP_L( ESI ) - RET - - - - -ALIGNTEXT4 -GLOBL GLNAME( _mesa_sse_transform_points4_3d ) -HIDDEN(_mesa_sse_transform_points4_3d) -GLNAME( _mesa_sse_transform_points4_3d ): - - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) /* ptr to source GLvector4f */ - MOV_L( ARG_DEST, EDI ) /* ptr to dest GLvector4f */ - - MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ - - TEST_L( ECX, ECX) - JZ( LLBL(K_GTP43P3DR_finish) ) /* count was zero; go to finish */ - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )/* set dest size */ - - SHL_L( CONST(4), ECX ) /* count *= 16 */ - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - ADD_L( EDI, ECX ) /* count += dest ptr */ - - MOVAPS( MAT(0), XMM0 ) /* m3 | m2 | m1 | m0 */ - MOVAPS( MAT(4), XMM1 ) /* m7 | m6 | m5 | m4 */ - MOVAPS( MAT(8), XMM2 ) /* m11 | m10 | m9 | m8 */ - MOVAPS( MAT(12), XMM3 ) /* m15 | m14 | m13 | m12 */ - -ALIGNTEXT32 -LLBL( K_GTP43P3DR_top ): - MOVSS( SRC(0), XMM4 ) /* ox */ - SHUFPS( CONST(0x0), XMM4, XMM4 ) /* ox | ox | ox | ox */ - MULPS( XMM0, XMM4 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ - - MOVSS( SRC(1), XMM5 ) /* oy */ - SHUFPS( CONST(0x0), XMM5, XMM5 ) /* oy | oy | oy | oy */ - MULPS( XMM1, XMM5 ) /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */ - - MOVSS( SRC(2), XMM6 ) /* oz */ - SHUFPS( CONST(0x0), XMM6, XMM6 ) /* oz | oz | oz | oz */ - MULPS( XMM2, XMM6 ) /* oz*m11 | oz*m10 | oz*m9 | oz*m8 */ - - MOVSS( SRC(3), XMM7 ) /* ow */ - SHUFPS( CONST(0x0), XMM7, XMM7 ) /* ow | ow | ow | ow */ - MULPS( XMM3, XMM7 ) /* ow*m15 | ow*m14 | ow*m13 | ow*m12 */ - - ADDPS( XMM5, XMM4 ) /* ox*m3+oy*m7 | ... */ - ADDPS( XMM6, XMM4 ) /* ox*m3+oy*m7+oz*m11 | ... */ - ADDPS( XMM7, XMM4 ) /* ox*m3+oy*m7+oz*m11+ow*m15 | ... */ - MOVAPS( XMM4, DST(0) ) /* ->D(3) | ->D(2) | ->D(1) | ->D(0) */ - - MOVSS( SRC(3), XMM4 ) /* ow */ - MOVSS( XMM4, DST(3) ) /* ->D(3) */ - -LLBL( K_GTP43P3DR_skip ): - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(K_GTP43P3DR_top) ) - -LLBL( K_GTP43P3DR_finish ): - POP_L( EDI ) - POP_L( ESI ) - RET - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_sse_transform_points4_identity ) -HIDDEN(_mesa_sse_transform_points4_identity) -GLNAME( _mesa_sse_transform_points4_identity ): - - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) /* verify non-zero count */ - JE( LLBL( sse_identity_done ) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )/* set dest size */ - - MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ - MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ - -ALIGNTEXT16 -LLBL( sse_identity_loop ): - - PREFETCHNTA( REGOFF(32, ESI) ) - - MOVAPS( REGIND(ESI), XMM0 ) - ADD_L( EAX, ESI ) - - MOVAPS( XMM0, REGIND(EDI) ) - ADD_L( CONST(16), EDI ) - - DEC_L( ECX ) - JNZ( LLBL( sse_identity_loop ) ) - -LLBL( sse_identity_done ): - - POP_L( EDI ) - POP_L( ESI ) - RET -#endif - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifdef USE_SSE_ASM +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FRAME_OFFSET 8 + +#define SRC(i) REGOFF(i * 4, ESI) +#define DST(i) REGOFF(i * 4, EDI) +#define MAT(i) REGOFF(i * 4, EDX) + +#define SELECT(r0, r1, r2, r3) CONST( r0 * 64 + r1 * 16 + r2 * 4 + r3 ) + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_sse_transform_points4_general ) +HIDDEN(_mesa_sse_transform_points4_general) +GLNAME( _mesa_sse_transform_points4_general ): + + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) /* verify non-zero count */ + JE( LLBL( sse_general_done ) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )/* set dest size */ + + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + + PREFETCHT0( REGIND(ESI) ) + + MOVAPS( MAT(0), XMM4 ) /* m3 | m2 | m1 | m0 */ + MOVAPS( MAT(4), XMM5 ) /* m7 | m6 | m5 | m4 */ + MOVAPS( MAT(8), XMM6 ) /* m11 | m10 | m9 | m8 */ + MOVAPS( MAT(12), XMM7 ) /* m15 | m14 | m13 | m12 */ + +ALIGNTEXT16 +LLBL( sse_general_loop ): + + MOVSS( SRC(0), XMM0 ) /* ox */ + SHUFPS( CONST(0x0), XMM0, XMM0 ) /* ox | ox | ox | ox */ + MULPS( XMM4, XMM0 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ + + MOVSS( SRC(1), XMM1 ) /* oy */ + SHUFPS( CONST(0x0), XMM1, XMM1 ) /* oy | oy | oy | oy */ + MULPS( XMM5, XMM1 ) /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */ + + MOVSS( SRC(2), XMM2 ) /* oz */ + SHUFPS( CONST(0x0), XMM2, XMM2 ) /* oz | oz | oz | oz */ + MULPS( XMM6, XMM2 ) /* oz*m11 | oz*m10 | oz*m9 | oz*m8 */ + + MOVSS( SRC(3), XMM3 ) /* ow */ + SHUFPS( CONST(0x0), XMM3, XMM3 ) /* ow | ow | ow | ow */ + MULPS( XMM7, XMM3 ) /* ow*m15 | ow*m14 | ow*m13 | ow*m12 */ + + ADDPS( XMM1, XMM0 ) /* ox*m3+oy*m7 | ... */ + ADDPS( XMM2, XMM0 ) /* ox*m3+oy*m7+oz*m11 | ... */ + ADDPS( XMM3, XMM0 ) /* ox*m3+oy*m7+oz*m11+ow*m15 | ... */ + MOVAPS( XMM0, DST(0) ) /* ->D(3) | ->D(2) | ->D(1) | ->D(0) */ + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + + DEC_L( ECX ) + JNZ( LLBL( sse_general_loop ) ) + +LLBL( sse_general_done ): + + POP_L( EDI ) + POP_L( ESI ) + RET + + + + +ALIGNTEXT4 +GLOBL GLNAME( _mesa_sse_transform_points4_3d ) +HIDDEN(_mesa_sse_transform_points4_3d) +GLNAME( _mesa_sse_transform_points4_3d ): + + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) /* ptr to source GLvector4f */ + MOV_L( ARG_DEST, EDI ) /* ptr to dest GLvector4f */ + + MOV_L( ARG_MATRIX, EDX ) /* ptr to matrix */ + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) /* source count */ + + TEST_L( ECX, ECX) + JZ( LLBL(K_GTP43P3DR_finish) ) /* count was zero; go to finish */ + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) )/* set dest size */ + + SHL_L( CONST(4), ECX ) /* count *= 16 */ + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + ADD_L( EDI, ECX ) /* count += dest ptr */ + + MOVAPS( MAT(0), XMM0 ) /* m3 | m2 | m1 | m0 */ + MOVAPS( MAT(4), XMM1 ) /* m7 | m6 | m5 | m4 */ + MOVAPS( MAT(8), XMM2 ) /* m11 | m10 | m9 | m8 */ + MOVAPS( MAT(12), XMM3 ) /* m15 | m14 | m13 | m12 */ + +ALIGNTEXT32 +LLBL( K_GTP43P3DR_top ): + MOVSS( SRC(0), XMM4 ) /* ox */ + SHUFPS( CONST(0x0), XMM4, XMM4 ) /* ox | ox | ox | ox */ + MULPS( XMM0, XMM4 ) /* ox*m3 | ox*m2 | ox*m1 | ox*m0 */ + + MOVSS( SRC(1), XMM5 ) /* oy */ + SHUFPS( CONST(0x0), XMM5, XMM5 ) /* oy | oy | oy | oy */ + MULPS( XMM1, XMM5 ) /* oy*m7 | oy*m6 | oy*m5 | oy*m4 */ + + MOVSS( SRC(2), XMM6 ) /* oz */ + SHUFPS( CONST(0x0), XMM6, XMM6 ) /* oz | oz | oz | oz */ + MULPS( XMM2, XMM6 ) /* oz*m11 | oz*m10 | oz*m9 | oz*m8 */ + + MOVSS( SRC(3), XMM7 ) /* ow */ + SHUFPS( CONST(0x0), XMM7, XMM7 ) /* ow | ow | ow | ow */ + MULPS( XMM3, XMM7 ) /* ow*m15 | ow*m14 | ow*m13 | ow*m12 */ + + ADDPS( XMM5, XMM4 ) /* ox*m3+oy*m7 | ... */ + ADDPS( XMM6, XMM4 ) /* ox*m3+oy*m7+oz*m11 | ... */ + ADDPS( XMM7, XMM4 ) /* ox*m3+oy*m7+oz*m11+ow*m15 | ... */ + MOVAPS( XMM4, DST(0) ) /* ->D(3) | ->D(2) | ->D(1) | ->D(0) */ + + MOVSS( SRC(3), XMM4 ) /* ow */ + MOVSS( XMM4, DST(3) ) /* ->D(3) */ + +LLBL( K_GTP43P3DR_skip ): + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(K_GTP43P3DR_top) ) + +LLBL( K_GTP43P3DR_finish ): + POP_L( EDI ) + POP_L( ESI ) + RET + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_sse_transform_points4_identity ) +HIDDEN(_mesa_sse_transform_points4_identity) +GLNAME( _mesa_sse_transform_points4_identity ): + + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) /* verify non-zero count */ + JE( LLBL( sse_identity_done ) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) /* stride */ + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) /* set dest flags */ + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) /* set dest count */ + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) )/* set dest size */ + + MOV_L( REGOFF(V4F_START, ESI), ESI ) /* ptr to first source vertex */ + MOV_L( REGOFF(V4F_START, EDI), EDI ) /* ptr to first dest vertex */ + +ALIGNTEXT16 +LLBL( sse_identity_loop ): + + PREFETCHNTA( REGOFF(32, ESI) ) + + MOVAPS( REGIND(ESI), XMM0 ) + ADD_L( EAX, ESI ) + + MOVAPS( XMM0, REGIND(EDI) ) + ADD_L( CONST(16), EDI ) + + DEC_L( ECX ) + JNZ( LLBL( sse_identity_loop ) ) + +LLBL( sse_identity_done ): + + POP_L( EDI ) + POP_L( ESI ) + RET +#endif + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/x86_cliptest.S b/mesalib/src/mesa/x86/x86_cliptest.S index 8556a887a..e413aee61 100644 --- a/mesalib/src/mesa/x86/x86_cliptest.S +++ b/mesalib/src/mesa/x86/x86_cliptest.S @@ -1,407 +1,407 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * NOTE: Avoid using spaces in between '(' ')' and arguments, especially - * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces - * in there will break the build on some platforms. - */ - -#include "assyntax.h" -#include "matypes.h" -#include "clip_args.h" - -#define SRC0 REGOFF(0, ESI) -#define SRC1 REGOFF(4, ESI) -#define SRC2 REGOFF(8, ESI) -#define SRC3 REGOFF(12, ESI) -#define DST0 REGOFF(0, EDI) -#define DST1 REGOFF(4, EDI) -#define DST2 REGOFF(8, EDI) -#define DST3 REGOFF(12, EDI) -#define MAT0 REGOFF(0, EDX) -#define MAT1 REGOFF(4, EDX) -#define MAT2 REGOFF(8, EDX) -#define MAT3 REGOFF(12, EDX) - - -/* - * Table for clip test. - * - * bit6 = SRC3 < 0 - * bit5 = SRC2 < 0 - * bit4 = abs(S(2)) > abs(S(3)) - * bit3 = SRC1 < 0 - * bit2 = abs(S(1)) > abs(S(3)) - * bit1 = SRC0 < 0 - * bit0 = abs(S(0)) > abs(S(3)) - */ - - SEG_DATA - -clip_table: - D_BYTE 0x00, 0x01, 0x00, 0x02, 0x04, 0x05, 0x04, 0x06 - D_BYTE 0x00, 0x01, 0x00, 0x02, 0x08, 0x09, 0x08, 0x0a - D_BYTE 0x20, 0x21, 0x20, 0x22, 0x24, 0x25, 0x24, 0x26 - D_BYTE 0x20, 0x21, 0x20, 0x22, 0x28, 0x29, 0x28, 0x2a - D_BYTE 0x00, 0x01, 0x00, 0x02, 0x04, 0x05, 0x04, 0x06 - D_BYTE 0x00, 0x01, 0x00, 0x02, 0x08, 0x09, 0x08, 0x0a - D_BYTE 0x10, 0x11, 0x10, 0x12, 0x14, 0x15, 0x14, 0x16 - D_BYTE 0x10, 0x11, 0x10, 0x12, 0x18, 0x19, 0x18, 0x1a - D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x37, 0x35, 0x37, 0x36 - D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x3b, 0x39, 0x3b, 0x3a - D_BYTE 0x2f, 0x2d, 0x2f, 0x2e, 0x27, 0x25, 0x27, 0x26 - D_BYTE 0x2f, 0x2d, 0x2f, 0x2e, 0x2b, 0x29, 0x2b, 0x2a - D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x37, 0x35, 0x37, 0x36 - D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x3b, 0x39, 0x3b, 0x3a - D_BYTE 0x1f, 0x1d, 0x1f, 0x1e, 0x17, 0x15, 0x17, 0x16 - D_BYTE 0x1f, 0x1d, 0x1f, 0x1e, 0x1b, 0x19, 0x1b, 0x1a - - - SEG_TEXT - -/* - * _mesa_x86_cliptest_points4 - * - * AL: ormask - * AH: andmask - * EBX: temp0 - * ECX: temp1 - * EDX: clipmask[] - * ESI: clip[] - * EDI: proj[] - * EBP: temp2 - */ - -#if defined(__ELF__) && defined(__PIC__) && defined(GNU_ASSEMBLER) && !defined(ELFPIC) -#define ELFPIC -#endif - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_cliptest_points4 ) -HIDDEN(_mesa_x86_cliptest_points4) -GLNAME( _mesa_x86_cliptest_points4 ): - -#ifdef ELFPIC -#define FRAME_OFFSET 20 -#else -#define FRAME_OFFSET 16 -#endif - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBP ) - PUSH_L( EBX ) - -#ifdef ELFPIC - /* store pointer to clip_table on stack */ - CALL( LLBL(ctp4_get_eip) ) - ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX ) - MOV_L( REGOFF(clip_table@GOT, EBX), EBX ) - PUSH_L( EBX ) - JMP( LLBL(ctp4_clip_table_ready) ) - -LLBL(ctp4_get_eip): - /* store eip in ebx */ - MOV_L( REGIND(ESP), EBX ) - RET - -LLBL(ctp4_clip_table_ready): -#endif - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_CLIP, EDX ) - MOV_L( ARG_OR, EBX ) - - MOV_L( ARG_AND, EBP ) - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - MOV_L( EAX, ARG_SOURCE ) /* put stride in ARG_SOURCE */ - - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDX, ECX ) - - MOV_L( ECX, ARG_CLIP ) /* put clipmask + count in ARG_CLIP */ - CMP_L( ECX, EDX ) - - MOV_B( REGIND(EBX), AL ) - MOV_B( REGIND(EBP), AH ) - - JZ( LLBL(ctp4_finish) ) - -ALIGNTEXT16 -LLBL(ctp4_top): - - FLD1 /* F3 */ - FDIV_S( SRC3 ) /* GH: don't care about div-by-zero */ - - MOV_L( SRC3, EBP ) - MOV_L( SRC2, EBX ) - - XOR_L( ECX, ECX ) - ADD_L( EBP, EBP ) /* ebp = abs(S(3))*2 ; carry = sign of S(3) */ - - ADC_L( ECX, ECX ) - ADD_L( EBX, EBX ) /* ebx = abs(S(2))*2 ; carry = sign of S(2) */ - - ADC_L( ECX, ECX ) - CMP_L( EBX, EBP ) /* carry = abs(S(2))*2 > abs(S(3))*2 */ - - ADC_L( ECX, ECX ) - MOV_L( SRC1, EBX ) - - ADD_L( EBX, EBX ) /* ebx = abs(S(1))*2 ; carry = sign of S(1) */ - - ADC_L( ECX, ECX ) - CMP_L( EBX, EBP ) /* carry = abs(S(1))*2 > abs(S(3))*2 */ - - ADC_L( ECX, ECX ) - MOV_L( SRC0, EBX ) - - ADD_L( EBX, EBX ) /* ebx = abs(S(0))*2 ; carry = sign of S(0) */ - - ADC_L( ECX, ECX ) - CMP_L( EBX, EBP ) /* carry = abs(S(0))*2 > abs(S(3))*2 */ - - ADC_L( ECX, ECX ) - -#ifdef ELFPIC - MOV_L( REGIND(ESP), EBP ) /* clip_table */ - - MOV_B( REGBI(EBP, ECX), CL ) -#else - MOV_B( REGOFF(clip_table,ECX), CL ) -#endif - - OR_B( CL, AL ) - AND_B( CL, AH ) - - TEST_B( CL, CL ) - MOV_B( CL, REGIND(EDX) ) - - JZ( LLBL(ctp4_proj) ) - -LLBL(ctp4_noproj): - - FSTP( ST(0) ) /* */ - - MOV_L( CONST(0), DST0 ) - MOV_L( CONST(0), DST1 ) - MOV_L( CONST(0), DST2 ) - MOV_L( CONST(0x3f800000), DST3 ) - - JMP( LLBL(ctp4_next) ) - -LLBL(ctp4_proj): - - FLD_S( SRC0 ) /* F0 F3 */ - FMUL2( ST(1), ST0 ) - - FLD_S( SRC1 ) /* F1 F0 F3 */ - FMUL2( ST(2), ST0 ) - - FLD_S( SRC2 ) /* F2 F1 F0 F3 */ - FMUL2( ST(3), ST0 ) - - FXCH( ST(2) ) /* F0 F1 F2 F3 */ - FSTP_S( DST0 ) /* F1 F2 F3 */ - FSTP_S( DST1 ) /* F2 F3 */ - FSTP_S( DST2 ) /* F3 */ - FSTP_S( DST3 ) /* */ - -LLBL(ctp4_next): - - INC_L( EDX ) - ADD_L( CONST(16), EDI ) - - ADD_L( ARG_SOURCE, ESI ) - CMP_L( EDX, ARG_CLIP ) - - JNZ( LLBL(ctp4_top) ) - - MOV_L( ARG_OR, ECX ) - MOV_L( ARG_AND, EDX ) - - MOV_B( AL, REGIND(ECX) ) - MOV_B( AH, REGIND(EDX) ) - -LLBL(ctp4_finish): - - MOV_L( ARG_DEST, EAX ) -#ifdef ELFPIC - POP_L( ESI ) /* discard ptr to clip_table */ -#endif - POP_L( EBX ) - POP_L( EBP ) - POP_L( EDI ) - POP_L( ESI ) - - RET - - - - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_cliptest_points4_np ) -HIDDEN(_mesa_x86_cliptest_points4_np) -GLNAME( _mesa_x86_cliptest_points4_np ): - -#ifdef ELFPIC -#define FRAME_OFFSET 20 -#else -#define FRAME_OFFSET 16 -#endif - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBP ) - PUSH_L( EBX ) - -#ifdef ELFPIC - /* store pointer to clip_table on stack */ - CALL( LLBL(ctp4_np_get_eip) ) - ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX ) - MOV_L( REGOFF(clip_table@GOT, EBX), EBX ) - PUSH_L( EBX ) - JMP( LLBL(ctp4_np_clip_table_ready) ) - -LLBL(ctp4_np_get_eip): - /* store eip in ebx */ - MOV_L( REGIND(ESP), EBX ) - RET - -LLBL(ctp4_np_clip_table_ready): -#endif - - MOV_L( ARG_SOURCE, ESI ) - /* slot */ - - MOV_L( ARG_CLIP, EDX ) - MOV_L( ARG_OR, EBX ) - - MOV_L( ARG_AND, EBP ) - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( EAX, ARG_DEST ) /* put stride in ARG_DEST */ - ADD_L( EDX, ECX ) - - MOV_L( ECX, EDI ) /* put clipmask + count in EDI */ - CMP_L( ECX, EDX ) - - MOV_B( REGIND(EBX), AL ) - MOV_B( REGIND(EBP), AH ) - - JZ( LLBL(ctp4_np_finish) ) - -ALIGNTEXT16 -LLBL(ctp4_np_top): - - MOV_L( SRC3, EBP ) - MOV_L( SRC2, EBX ) - - XOR_L( ECX, ECX ) - ADD_L( EBP, EBP ) /* ebp = abs(S(3))*2 ; carry = sign of S(3) */ - - ADC_L( ECX, ECX ) - ADD_L( EBX, EBX ) /* ebx = abs(S(2))*2 ; carry = sign of S(2) */ - - ADC_L( ECX, ECX ) - CMP_L( EBX, EBP ) /* carry = abs(S(2))*2 > abs(S(3))*2 */ - - ADC_L( ECX, ECX ) - MOV_L( SRC1, EBX ) - - ADD_L( EBX, EBX ) /* ebx = abs(S(1))*2 ; carry = sign of S(1) */ - - ADC_L( ECX, ECX ) - CMP_L( EBX, EBP ) /* carry = abs(S(1))*2 > abs(S(3))*2 */ - - ADC_L( ECX, ECX ) - MOV_L( SRC0, EBX ) - - ADD_L( EBX, EBX ) /* ebx = abs(S(0))*2 ; carry = sign of S(0) */ - - ADC_L( ECX, ECX ) - CMP_L( EBX, EBP ) /* carry = abs(S(0))*2 > abs(S(3))*2 */ - - ADC_L( ECX, ECX ) - -#ifdef ELFPIC - MOV_L( REGIND(ESP), EBP ) /* clip_table */ - - MOV_B( REGBI(EBP, ECX), CL ) -#else - MOV_B( REGOFF(clip_table,ECX), CL ) -#endif - - OR_B( CL, AL ) - AND_B( CL, AH ) - - TEST_B( CL, CL ) - MOV_B( CL, REGIND(EDX) ) - - INC_L( EDX ) - /* slot */ - - ADD_L( ARG_DEST, ESI ) - CMP_L( EDX, EDI ) - - JNZ( LLBL(ctp4_np_top) ) - - MOV_L( ARG_OR, ECX ) - MOV_L( ARG_AND, EDX ) - - MOV_B( AL, REGIND(ECX) ) - MOV_B( AH, REGIND(EDX) ) - -LLBL(ctp4_np_finish): - - MOV_L( ARG_SOURCE, EAX ) -#ifdef ELFPIC - POP_L( ESI ) /* discard ptr to clip_table */ -#endif - POP_L( EBX ) - POP_L( EBP ) - POP_L( EDI ) - POP_L( ESI ) - - RET - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * NOTE: Avoid using spaces in between '(' ')' and arguments, especially + * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces + * in there will break the build on some platforms. + */ + +#include "assyntax.h" +#include "matypes.h" +#include "clip_args.h" + +#define SRC0 REGOFF(0, ESI) +#define SRC1 REGOFF(4, ESI) +#define SRC2 REGOFF(8, ESI) +#define SRC3 REGOFF(12, ESI) +#define DST0 REGOFF(0, EDI) +#define DST1 REGOFF(4, EDI) +#define DST2 REGOFF(8, EDI) +#define DST3 REGOFF(12, EDI) +#define MAT0 REGOFF(0, EDX) +#define MAT1 REGOFF(4, EDX) +#define MAT2 REGOFF(8, EDX) +#define MAT3 REGOFF(12, EDX) + + +/* + * Table for clip test. + * + * bit6 = SRC3 < 0 + * bit5 = SRC2 < 0 + * bit4 = abs(S(2)) > abs(S(3)) + * bit3 = SRC1 < 0 + * bit2 = abs(S(1)) > abs(S(3)) + * bit1 = SRC0 < 0 + * bit0 = abs(S(0)) > abs(S(3)) + */ + + SEG_DATA + +clip_table: + D_BYTE 0x00, 0x01, 0x00, 0x02, 0x04, 0x05, 0x04, 0x06 + D_BYTE 0x00, 0x01, 0x00, 0x02, 0x08, 0x09, 0x08, 0x0a + D_BYTE 0x20, 0x21, 0x20, 0x22, 0x24, 0x25, 0x24, 0x26 + D_BYTE 0x20, 0x21, 0x20, 0x22, 0x28, 0x29, 0x28, 0x2a + D_BYTE 0x00, 0x01, 0x00, 0x02, 0x04, 0x05, 0x04, 0x06 + D_BYTE 0x00, 0x01, 0x00, 0x02, 0x08, 0x09, 0x08, 0x0a + D_BYTE 0x10, 0x11, 0x10, 0x12, 0x14, 0x15, 0x14, 0x16 + D_BYTE 0x10, 0x11, 0x10, 0x12, 0x18, 0x19, 0x18, 0x1a + D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x37, 0x35, 0x37, 0x36 + D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x3b, 0x39, 0x3b, 0x3a + D_BYTE 0x2f, 0x2d, 0x2f, 0x2e, 0x27, 0x25, 0x27, 0x26 + D_BYTE 0x2f, 0x2d, 0x2f, 0x2e, 0x2b, 0x29, 0x2b, 0x2a + D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x37, 0x35, 0x37, 0x36 + D_BYTE 0x3f, 0x3d, 0x3f, 0x3e, 0x3b, 0x39, 0x3b, 0x3a + D_BYTE 0x1f, 0x1d, 0x1f, 0x1e, 0x17, 0x15, 0x17, 0x16 + D_BYTE 0x1f, 0x1d, 0x1f, 0x1e, 0x1b, 0x19, 0x1b, 0x1a + + + SEG_TEXT + +/* + * _mesa_x86_cliptest_points4 + * + * AL: ormask + * AH: andmask + * EBX: temp0 + * ECX: temp1 + * EDX: clipmask[] + * ESI: clip[] + * EDI: proj[] + * EBP: temp2 + */ + +#if defined(__ELF__) && defined(__PIC__) && defined(GNU_ASSEMBLER) && !defined(ELFPIC) +#define ELFPIC +#endif + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_cliptest_points4 ) +HIDDEN(_mesa_x86_cliptest_points4) +GLNAME( _mesa_x86_cliptest_points4 ): + +#ifdef ELFPIC +#define FRAME_OFFSET 20 +#else +#define FRAME_OFFSET 16 +#endif + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBP ) + PUSH_L( EBX ) + +#ifdef ELFPIC + /* store pointer to clip_table on stack */ + CALL( LLBL(ctp4_get_eip) ) + ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX ) + MOV_L( REGOFF(clip_table@GOT, EBX), EBX ) + PUSH_L( EBX ) + JMP( LLBL(ctp4_clip_table_ready) ) + +LLBL(ctp4_get_eip): + /* store eip in ebx */ + MOV_L( REGIND(ESP), EBX ) + RET + +LLBL(ctp4_clip_table_ready): +#endif + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_CLIP, EDX ) + MOV_L( ARG_OR, EBX ) + + MOV_L( ARG_AND, EBP ) + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + MOV_L( EAX, ARG_SOURCE ) /* put stride in ARG_SOURCE */ + + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDX, ECX ) + + MOV_L( ECX, ARG_CLIP ) /* put clipmask + count in ARG_CLIP */ + CMP_L( ECX, EDX ) + + MOV_B( REGIND(EBX), AL ) + MOV_B( REGIND(EBP), AH ) + + JZ( LLBL(ctp4_finish) ) + +ALIGNTEXT16 +LLBL(ctp4_top): + + FLD1 /* F3 */ + FDIV_S( SRC3 ) /* GH: don't care about div-by-zero */ + + MOV_L( SRC3, EBP ) + MOV_L( SRC2, EBX ) + + XOR_L( ECX, ECX ) + ADD_L( EBP, EBP ) /* ebp = abs(S(3))*2 ; carry = sign of S(3) */ + + ADC_L( ECX, ECX ) + ADD_L( EBX, EBX ) /* ebx = abs(S(2))*2 ; carry = sign of S(2) */ + + ADC_L( ECX, ECX ) + CMP_L( EBX, EBP ) /* carry = abs(S(2))*2 > abs(S(3))*2 */ + + ADC_L( ECX, ECX ) + MOV_L( SRC1, EBX ) + + ADD_L( EBX, EBX ) /* ebx = abs(S(1))*2 ; carry = sign of S(1) */ + + ADC_L( ECX, ECX ) + CMP_L( EBX, EBP ) /* carry = abs(S(1))*2 > abs(S(3))*2 */ + + ADC_L( ECX, ECX ) + MOV_L( SRC0, EBX ) + + ADD_L( EBX, EBX ) /* ebx = abs(S(0))*2 ; carry = sign of S(0) */ + + ADC_L( ECX, ECX ) + CMP_L( EBX, EBP ) /* carry = abs(S(0))*2 > abs(S(3))*2 */ + + ADC_L( ECX, ECX ) + +#ifdef ELFPIC + MOV_L( REGIND(ESP), EBP ) /* clip_table */ + + MOV_B( REGBI(EBP, ECX), CL ) +#else + MOV_B( REGOFF(clip_table,ECX), CL ) +#endif + + OR_B( CL, AL ) + AND_B( CL, AH ) + + TEST_B( CL, CL ) + MOV_B( CL, REGIND(EDX) ) + + JZ( LLBL(ctp4_proj) ) + +LLBL(ctp4_noproj): + + FSTP( ST(0) ) /* */ + + MOV_L( CONST(0), DST0 ) + MOV_L( CONST(0), DST1 ) + MOV_L( CONST(0), DST2 ) + MOV_L( CONST(0x3f800000), DST3 ) + + JMP( LLBL(ctp4_next) ) + +LLBL(ctp4_proj): + + FLD_S( SRC0 ) /* F0 F3 */ + FMUL2( ST(1), ST0 ) + + FLD_S( SRC1 ) /* F1 F0 F3 */ + FMUL2( ST(2), ST0 ) + + FLD_S( SRC2 ) /* F2 F1 F0 F3 */ + FMUL2( ST(3), ST0 ) + + FXCH( ST(2) ) /* F0 F1 F2 F3 */ + FSTP_S( DST0 ) /* F1 F2 F3 */ + FSTP_S( DST1 ) /* F2 F3 */ + FSTP_S( DST2 ) /* F3 */ + FSTP_S( DST3 ) /* */ + +LLBL(ctp4_next): + + INC_L( EDX ) + ADD_L( CONST(16), EDI ) + + ADD_L( ARG_SOURCE, ESI ) + CMP_L( EDX, ARG_CLIP ) + + JNZ( LLBL(ctp4_top) ) + + MOV_L( ARG_OR, ECX ) + MOV_L( ARG_AND, EDX ) + + MOV_B( AL, REGIND(ECX) ) + MOV_B( AH, REGIND(EDX) ) + +LLBL(ctp4_finish): + + MOV_L( ARG_DEST, EAX ) +#ifdef ELFPIC + POP_L( ESI ) /* discard ptr to clip_table */ +#endif + POP_L( EBX ) + POP_L( EBP ) + POP_L( EDI ) + POP_L( ESI ) + + RET + + + + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_cliptest_points4_np ) +HIDDEN(_mesa_x86_cliptest_points4_np) +GLNAME( _mesa_x86_cliptest_points4_np ): + +#ifdef ELFPIC +#define FRAME_OFFSET 20 +#else +#define FRAME_OFFSET 16 +#endif + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBP ) + PUSH_L( EBX ) + +#ifdef ELFPIC + /* store pointer to clip_table on stack */ + CALL( LLBL(ctp4_np_get_eip) ) + ADD_L( CONST(_GLOBAL_OFFSET_TABLE_), EBX ) + MOV_L( REGOFF(clip_table@GOT, EBX), EBX ) + PUSH_L( EBX ) + JMP( LLBL(ctp4_np_clip_table_ready) ) + +LLBL(ctp4_np_get_eip): + /* store eip in ebx */ + MOV_L( REGIND(ESP), EBX ) + RET + +LLBL(ctp4_np_clip_table_ready): +#endif + + MOV_L( ARG_SOURCE, ESI ) + /* slot */ + + MOV_L( ARG_CLIP, EDX ) + MOV_L( ARG_OR, EBX ) + + MOV_L( ARG_AND, EBP ) + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( EAX, ARG_DEST ) /* put stride in ARG_DEST */ + ADD_L( EDX, ECX ) + + MOV_L( ECX, EDI ) /* put clipmask + count in EDI */ + CMP_L( ECX, EDX ) + + MOV_B( REGIND(EBX), AL ) + MOV_B( REGIND(EBP), AH ) + + JZ( LLBL(ctp4_np_finish) ) + +ALIGNTEXT16 +LLBL(ctp4_np_top): + + MOV_L( SRC3, EBP ) + MOV_L( SRC2, EBX ) + + XOR_L( ECX, ECX ) + ADD_L( EBP, EBP ) /* ebp = abs(S(3))*2 ; carry = sign of S(3) */ + + ADC_L( ECX, ECX ) + ADD_L( EBX, EBX ) /* ebx = abs(S(2))*2 ; carry = sign of S(2) */ + + ADC_L( ECX, ECX ) + CMP_L( EBX, EBP ) /* carry = abs(S(2))*2 > abs(S(3))*2 */ + + ADC_L( ECX, ECX ) + MOV_L( SRC1, EBX ) + + ADD_L( EBX, EBX ) /* ebx = abs(S(1))*2 ; carry = sign of S(1) */ + + ADC_L( ECX, ECX ) + CMP_L( EBX, EBP ) /* carry = abs(S(1))*2 > abs(S(3))*2 */ + + ADC_L( ECX, ECX ) + MOV_L( SRC0, EBX ) + + ADD_L( EBX, EBX ) /* ebx = abs(S(0))*2 ; carry = sign of S(0) */ + + ADC_L( ECX, ECX ) + CMP_L( EBX, EBP ) /* carry = abs(S(0))*2 > abs(S(3))*2 */ + + ADC_L( ECX, ECX ) + +#ifdef ELFPIC + MOV_L( REGIND(ESP), EBP ) /* clip_table */ + + MOV_B( REGBI(EBP, ECX), CL ) +#else + MOV_B( REGOFF(clip_table,ECX), CL ) +#endif + + OR_B( CL, AL ) + AND_B( CL, AH ) + + TEST_B( CL, CL ) + MOV_B( CL, REGIND(EDX) ) + + INC_L( EDX ) + /* slot */ + + ADD_L( ARG_DEST, ESI ) + CMP_L( EDX, EDI ) + + JNZ( LLBL(ctp4_np_top) ) + + MOV_L( ARG_OR, ECX ) + MOV_L( ARG_AND, EDX ) + + MOV_B( AL, REGIND(ECX) ) + MOV_B( AH, REGIND(EDX) ) + +LLBL(ctp4_np_finish): + + MOV_L( ARG_SOURCE, EAX ) +#ifdef ELFPIC + POP_L( ESI ) /* discard ptr to clip_table */ +#endif + POP_L( EBX ) + POP_L( EBP ) + POP_L( EDI ) + POP_L( ESI ) + + RET + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/x86_xform.c b/mesalib/src/mesa/x86/x86_xform.c index 57b10c262..3dcc55e16 100644 --- a/mesalib/src/mesa/x86/x86_xform.c +++ b/mesalib/src/mesa/x86/x86_xform.c @@ -1,126 +1,126 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * Intel x86 assembly code by Josh Vanderhoof - */ - -#include "main/glheader.h" -#include "main/context.h" -#include "math/m_xform.h" - -#include "x86_xform.h" -#include "common_x86_asm.h" - -#ifdef USE_X86_ASM -#ifdef USE_3DNOW_ASM -#include "3dnow.h" -#endif -#ifdef USE_SSE_ASM -#include "sse.h" -#endif -#endif - -#ifdef DEBUG_MATH -#include "math/m_debug.h" -#endif - - -#ifdef USE_X86_ASM -DECLARE_XFORM_GROUP( x86, 2 ) -DECLARE_XFORM_GROUP( x86, 3 ) -DECLARE_XFORM_GROUP( x86, 4 ) - - -extern GLvector4f * _ASMAPI -_mesa_x86_cliptest_points4( GLvector4f *clip_vec, - GLvector4f *proj_vec, - GLubyte clipMask[], - GLubyte *orMask, - GLubyte *andMask, - GLboolean viewport_z_clip ); - -extern GLvector4f * _ASMAPI -_mesa_x86_cliptest_points4_np( GLvector4f *clip_vec, - GLvector4f *proj_vec, - GLubyte clipMask[], - GLubyte *orMask, - GLubyte *andMask, - GLboolean viewport_z_clip ); - -extern void _ASMAPI -_mesa_v16_x86_cliptest_points4( GLfloat *first_vert, - GLfloat *last_vert, - GLubyte *or_mask, - GLubyte *and_mask, - GLubyte *clip_mask, - GLboolean viewport_z_clip ); - -extern void _ASMAPI -_mesa_v16_x86_general_xform( GLfloat *dest, - const GLfloat *m, - const GLfloat *src, - GLuint src_stride, - GLuint count ); -#endif - - -#ifdef USE_X86_ASM -static void _mesa_init_x86_transform_asm( void ) -{ - ASSIGN_XFORM_GROUP( x86, 2 ); - ASSIGN_XFORM_GROUP( x86, 3 ); - ASSIGN_XFORM_GROUP( x86, 4 ); - - _mesa_clip_tab[4] = _mesa_x86_cliptest_points4; - _mesa_clip_np_tab[4] = _mesa_x86_cliptest_points4_np; - -#ifdef DEBUG_MATH - _math_test_all_transform_functions( "x86" ); - _math_test_all_cliptest_functions( "x86" ); -#endif -} -#endif - - -void _mesa_init_all_x86_transform_asm( void ) -{ - _mesa_get_x86_features(); - -#ifdef USE_X86_ASM - if ( _mesa_x86_cpu_features ) { - _mesa_init_x86_transform_asm(); - } - - if (cpu_has_3dnow) { - _mesa_init_3dnow_transform_asm(); - } - - if ( cpu_has_xmm ) { - _mesa_init_sse_transform_asm(); - } - -#endif -} + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * Intel x86 assembly code by Josh Vanderhoof + */ + +#include "main/glheader.h" +#include "main/context.h" +#include "math/m_xform.h" + +#include "x86_xform.h" +#include "common_x86_asm.h" + +#ifdef USE_X86_ASM +#ifdef USE_3DNOW_ASM +#include "3dnow.h" +#endif +#ifdef USE_SSE_ASM +#include "sse.h" +#endif +#endif + +#ifdef DEBUG_MATH +#include "math/m_debug.h" +#endif + + +#ifdef USE_X86_ASM +DECLARE_XFORM_GROUP( x86, 2 ) +DECLARE_XFORM_GROUP( x86, 3 ) +DECLARE_XFORM_GROUP( x86, 4 ) + + +extern GLvector4f * _ASMAPI +_mesa_x86_cliptest_points4( GLvector4f *clip_vec, + GLvector4f *proj_vec, + GLubyte clipMask[], + GLubyte *orMask, + GLubyte *andMask, + GLboolean viewport_z_clip ); + +extern GLvector4f * _ASMAPI +_mesa_x86_cliptest_points4_np( GLvector4f *clip_vec, + GLvector4f *proj_vec, + GLubyte clipMask[], + GLubyte *orMask, + GLubyte *andMask, + GLboolean viewport_z_clip ); + +extern void _ASMAPI +_mesa_v16_x86_cliptest_points4( GLfloat *first_vert, + GLfloat *last_vert, + GLubyte *or_mask, + GLubyte *and_mask, + GLubyte *clip_mask, + GLboolean viewport_z_clip ); + +extern void _ASMAPI +_mesa_v16_x86_general_xform( GLfloat *dest, + const GLfloat *m, + const GLfloat *src, + GLuint src_stride, + GLuint count ); +#endif + + +#ifdef USE_X86_ASM +static void _mesa_init_x86_transform_asm( void ) +{ + ASSIGN_XFORM_GROUP( x86, 2 ); + ASSIGN_XFORM_GROUP( x86, 3 ); + ASSIGN_XFORM_GROUP( x86, 4 ); + + _mesa_clip_tab[4] = _mesa_x86_cliptest_points4; + _mesa_clip_np_tab[4] = _mesa_x86_cliptest_points4_np; + +#ifdef DEBUG_MATH + _math_test_all_transform_functions( "x86" ); + _math_test_all_cliptest_functions( "x86" ); +#endif +} +#endif + + +void _mesa_init_all_x86_transform_asm( void ) +{ + _mesa_get_x86_features(); + +#ifdef USE_X86_ASM + if ( _mesa_x86_cpu_features ) { + _mesa_init_x86_transform_asm(); + } + + if (cpu_has_3dnow) { + _mesa_init_3dnow_transform_asm(); + } + + if ( cpu_has_xmm ) { + _mesa_init_sse_transform_asm(); + } + +#endif +} diff --git a/mesalib/src/mesa/x86/x86_xform.h b/mesalib/src/mesa/x86/x86_xform.h index 055fd6a1f..e886d9add 100644 --- a/mesalib/src/mesa/x86/x86_xform.h +++ b/mesalib/src/mesa/x86/x86_xform.h @@ -1,106 +1,106 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Gareth Hughes - */ - -#ifndef X86_XFORM_H -#define X86_XFORM_H - - -/* ============================================================= - * Transformation function declarations: - */ - -#define XFORM_ARGS GLvector4f *to_vec, \ - const GLfloat m[16], \ - const GLvector4f *from_vec - -#define DECLARE_XFORM_GROUP( pfx, sz ) \ -extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_general( XFORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_identity( XFORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_3d_no_rot( XFORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_perspective( XFORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_2d( XFORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_2d_no_rot( XFORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_3d( XFORM_ARGS ); - -#define ASSIGN_XFORM_GROUP( pfx, sz ) \ - _mesa_transform_tab[sz][MATRIX_GENERAL] = \ - _mesa_##pfx##_transform_points##sz##_general; \ - _mesa_transform_tab[sz][MATRIX_IDENTITY] = \ - _mesa_##pfx##_transform_points##sz##_identity; \ - _mesa_transform_tab[sz][MATRIX_3D_NO_ROT] = \ - _mesa_##pfx##_transform_points##sz##_3d_no_rot; \ - _mesa_transform_tab[sz][MATRIX_PERSPECTIVE] = \ - _mesa_##pfx##_transform_points##sz##_perspective; \ - _mesa_transform_tab[sz][MATRIX_2D] = \ - _mesa_##pfx##_transform_points##sz##_2d; \ - _mesa_transform_tab[sz][MATRIX_2D_NO_ROT] = \ - _mesa_##pfx##_transform_points##sz##_2d_no_rot; \ - _mesa_transform_tab[sz][MATRIX_3D] = \ - _mesa_##pfx##_transform_points##sz##_3d; - - -/* ============================================================= - * Normal transformation function declarations: - */ - -#define NORM_ARGS const GLmatrix *mat, \ - GLfloat scale, \ - const GLvector4f *in, \ - const GLfloat *lengths, \ - GLvector4f *dest - -#define DECLARE_NORM_GROUP( pfx ) \ -extern void _ASMAPI _mesa_##pfx##_rescale_normals( NORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_normalize_normals( NORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_normals( NORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_normals_no_rot( NORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_rescale_normals( NORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_rescale_normals_no_rot( NORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_normalize_normals( NORM_ARGS ); \ -extern void _ASMAPI _mesa_##pfx##_transform_normalize_normals_no_rot( NORM_ARGS ); - -#define ASSIGN_NORM_GROUP( pfx ) \ - _mesa_normal_tab[NORM_RESCALE] = \ - _mesa_##pfx##_rescale_normals; \ - _mesa_normal_tab[NORM_NORMALIZE] = \ - _mesa_##pfx##_normalize_normals; \ - _mesa_normal_tab[NORM_TRANSFORM] = \ - _mesa_##pfx##_transform_normals; \ - _mesa_normal_tab[NORM_TRANSFORM_NO_ROT] = \ - _mesa_##pfx##_transform_normals_no_rot; \ - _mesa_normal_tab[NORM_TRANSFORM | NORM_RESCALE] = \ - _mesa_##pfx##_transform_rescale_normals; \ - _mesa_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE] = \ - _mesa_##pfx##_transform_rescale_normals_no_rot; \ - _mesa_normal_tab[NORM_TRANSFORM | NORM_NORMALIZE] = \ - _mesa_##pfx##_transform_normalize_normals; \ - _mesa_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE] = \ - _mesa_##pfx##_transform_normalize_normals_no_rot; - - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: + * Gareth Hughes + */ + +#ifndef X86_XFORM_H +#define X86_XFORM_H + + +/* ============================================================= + * Transformation function declarations: + */ + +#define XFORM_ARGS GLvector4f *to_vec, \ + const GLfloat m[16], \ + const GLvector4f *from_vec + +#define DECLARE_XFORM_GROUP( pfx, sz ) \ +extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_general( XFORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_identity( XFORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_3d_no_rot( XFORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_perspective( XFORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_2d( XFORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_2d_no_rot( XFORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_points##sz##_3d( XFORM_ARGS ); + +#define ASSIGN_XFORM_GROUP( pfx, sz ) \ + _mesa_transform_tab[sz][MATRIX_GENERAL] = \ + _mesa_##pfx##_transform_points##sz##_general; \ + _mesa_transform_tab[sz][MATRIX_IDENTITY] = \ + _mesa_##pfx##_transform_points##sz##_identity; \ + _mesa_transform_tab[sz][MATRIX_3D_NO_ROT] = \ + _mesa_##pfx##_transform_points##sz##_3d_no_rot; \ + _mesa_transform_tab[sz][MATRIX_PERSPECTIVE] = \ + _mesa_##pfx##_transform_points##sz##_perspective; \ + _mesa_transform_tab[sz][MATRIX_2D] = \ + _mesa_##pfx##_transform_points##sz##_2d; \ + _mesa_transform_tab[sz][MATRIX_2D_NO_ROT] = \ + _mesa_##pfx##_transform_points##sz##_2d_no_rot; \ + _mesa_transform_tab[sz][MATRIX_3D] = \ + _mesa_##pfx##_transform_points##sz##_3d; + + +/* ============================================================= + * Normal transformation function declarations: + */ + +#define NORM_ARGS const GLmatrix *mat, \ + GLfloat scale, \ + const GLvector4f *in, \ + const GLfloat *lengths, \ + GLvector4f *dest + +#define DECLARE_NORM_GROUP( pfx ) \ +extern void _ASMAPI _mesa_##pfx##_rescale_normals( NORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_normalize_normals( NORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_normals( NORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_normals_no_rot( NORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_rescale_normals( NORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_rescale_normals_no_rot( NORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_normalize_normals( NORM_ARGS ); \ +extern void _ASMAPI _mesa_##pfx##_transform_normalize_normals_no_rot( NORM_ARGS ); + +#define ASSIGN_NORM_GROUP( pfx ) \ + _mesa_normal_tab[NORM_RESCALE] = \ + _mesa_##pfx##_rescale_normals; \ + _mesa_normal_tab[NORM_NORMALIZE] = \ + _mesa_##pfx##_normalize_normals; \ + _mesa_normal_tab[NORM_TRANSFORM] = \ + _mesa_##pfx##_transform_normals; \ + _mesa_normal_tab[NORM_TRANSFORM_NO_ROT] = \ + _mesa_##pfx##_transform_normals_no_rot; \ + _mesa_normal_tab[NORM_TRANSFORM | NORM_RESCALE] = \ + _mesa_##pfx##_transform_rescale_normals; \ + _mesa_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_RESCALE] = \ + _mesa_##pfx##_transform_rescale_normals_no_rot; \ + _mesa_normal_tab[NORM_TRANSFORM | NORM_NORMALIZE] = \ + _mesa_##pfx##_transform_normalize_normals; \ + _mesa_normal_tab[NORM_TRANSFORM_NO_ROT | NORM_NORMALIZE] = \ + _mesa_##pfx##_transform_normalize_normals_no_rot; + + +#endif diff --git a/mesalib/src/mesa/x86/x86_xform2.S b/mesalib/src/mesa/x86/x86_xform2.S index 482035c1f..980725ef5 100644 --- a/mesalib/src/mesa/x86/x86_xform2.S +++ b/mesalib/src/mesa/x86/x86_xform2.S @@ -1,574 +1,574 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * NOTE: Avoid using spaces in between '(' ')' and arguments, especially - * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces - * in there will break the build on some platforms. - */ - -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FP_ONE 1065353216 -#define FP_ZERO 0 - -#define SRC0 REGOFF(0, ESI) -#define SRC1 REGOFF(4, ESI) -#define SRC2 REGOFF(8, ESI) -#define SRC3 REGOFF(12, ESI) -#define DST0 REGOFF(0, EDI) -#define DST1 REGOFF(4, EDI) -#define DST2 REGOFF(8, EDI) -#define DST3 REGOFF(12, EDI) -#define MAT0 REGOFF(0, EDX) -#define MAT1 REGOFF(4, EDX) -#define MAT2 REGOFF(8, EDX) -#define MAT3 REGOFF(12, EDX) -#define MAT4 REGOFF(16, EDX) -#define MAT5 REGOFF(20, EDX) -#define MAT6 REGOFF(24, EDX) -#define MAT7 REGOFF(28, EDX) -#define MAT8 REGOFF(32, EDX) -#define MAT9 REGOFF(36, EDX) -#define MAT10 REGOFF(40, EDX) -#define MAT11 REGOFF(44, EDX) -#define MAT12 REGOFF(48, EDX) -#define MAT13 REGOFF(52, EDX) -#define MAT14 REGOFF(56, EDX) -#define MAT15 REGOFF(60, EDX) - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points2_general ) -HIDDEN(_mesa_x86_transform_points2_general) -GLNAME( _mesa_x86_transform_points2_general ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p2_gr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p2_gr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - FLD_S( SRC0 ) /* F6 F5 F4 */ - FMUL_S( MAT2 ) - FLD_S( SRC0 ) /* F7 F6 F5 F4 */ - FMUL_S( MAT3 ) - - FLD_S( SRC1 ) /* F0 F7 F6 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT5 ) - FLD_S( SRC1 ) /* F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT6 ) - FLD_S( SRC1 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT7 ) - - FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ - FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ - - FXCH( ST(3) ) /* F4 F6 F5 F7 */ - FADD_S( MAT12 ) - FXCH( ST(2) ) /* F5 F6 F4 F7 */ - FADD_S( MAT13 ) - FXCH( ST(1) ) /* F6 F5 F4 F7 */ - FADD_S( MAT14 ) - FXCH( ST(3) ) /* F7 F5 F4 F6 */ - FADD_S( MAT15 ) - - FXCH( ST(2) ) /* F4 F5 F7 F6 */ - FSTP_S( DST0 ) /* F5 F7 F6 */ - FSTP_S( DST1 ) /* F7 F6 */ - FXCH( ST(1) ) /* F6 F7 */ - FSTP_S( DST2 ) /* F7 */ - FSTP_S( DST3 ) /* */ - -LLBL(x86_p2_gr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p2_gr_loop) ) - -LLBL(x86_p2_gr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points2_perspective ) -HIDDEN(_mesa_x86_transform_points2_perspective) -GLNAME( _mesa_x86_transform_points2_perspective ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p2_pr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - - MOV_L( MAT14, EBX ) - -ALIGNTEXT16 -LLBL(x86_p2_pr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F1 F4 */ - FMUL_S( MAT5 ) - - FXCH( ST(1) ) /* F4 F1 */ - FSTP_S( DST0 ) /* F1 */ - FSTP_S( DST1 ) /* */ - MOV_L( EBX, DST2 ) - MOV_L( CONST(FP_ZERO), DST3 ) - -LLBL(x86_p2_pr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p2_pr_loop) ) - -LLBL(x86_p2_pr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points2_3d ) -HIDDEN(_mesa_x86_transform_points2_3d) -GLNAME( _mesa_x86_transform_points2_3d ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p2_3dr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p2_3dr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - FLD_S( SRC0 ) /* F6 F5 F4 */ - FMUL_S( MAT2 ) - - FLD_S( SRC1 ) /* F0 F6 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F6 F5 F4 */ - FMUL_S( MAT5 ) - FLD_S( SRC1 ) /* F2 F1 F0 F6 F5 F4 */ - FMUL_S( MAT6 ) - - FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - FXCH( ST(2) ) /* F4 F5 F6 */ - FADD_S( MAT12 ) - FXCH( ST(1) ) /* F5 F4 F6 */ - FADD_S( MAT13 ) - FXCH( ST(2) ) /* F6 F4 F5 */ - FADD_S( MAT14 ) - - FXCH( ST(1) ) /* F4 F6 F5 */ - FSTP_S( DST0 ) /* F6 F5 */ - FXCH( ST(1) ) /* F5 F6 */ - FSTP_S( DST1 ) /* F6 */ - FSTP_S( DST2 ) /* */ - -LLBL(x86_p2_3dr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p2_3dr_loop) ) - -LLBL(x86_p2_3dr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points2_3d_no_rot ) -HIDDEN(_mesa_x86_transform_points2_3d_no_rot) -GLNAME( _mesa_x86_transform_points2_3d_no_rot ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p2_3dnrr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - - MOV_L( MAT14, EBX ) - -ALIGNTEXT16 -LLBL(x86_p2_3dnrr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F1 F4 */ - FMUL_S( MAT5 ) - - FXCH( ST(1) ) /* F4 F1 */ - FADD_S( MAT12 ) - FLD_S( MAT13 ) /* F5 F4 F1 */ - FXCH( ST(2) ) /* F1 F4 F5 */ - FADDP( ST0, ST(2) ) /* F4 F5 */ - - FSTP_S( DST0 ) /* F5 */ - FSTP_S( DST1 ) /* */ - MOV_L( EBX, DST2 ) - -LLBL(x86_p2_3dnrr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p2_3dnrr_loop) ) - -LLBL(x86_p2_3dnrr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points2_2d ) -HIDDEN(_mesa_x86_transform_points2_2d) -GLNAME( _mesa_x86_transform_points2_2d ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p2_2dr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p2_2dr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - - FLD_S( SRC1 ) /* F0 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F5 F4 */ - FMUL_S( MAT5 ) - - FXCH( ST(1) ) /* F0 F1 F5 F4 */ - FADDP( ST0, ST(3) ) /* F1 F5 F4 */ - FADDP( ST0, ST(1) ) /* F5 F4 */ - - FXCH( ST(1) ) /* F4 F5 */ - FADD_S( MAT12 ) - FXCH( ST(1) ) /* F5 F4 */ - FADD_S( MAT13 ) - - FXCH( ST(1) ) /* F4 F5 */ - FSTP_S( DST0 ) /* F5 */ - FSTP_S( DST1 ) /* */ - -LLBL(x86_p2_2dr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p2_2dr_loop) ) - -LLBL(x86_p2_2dr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT4 -GLOBL GLNAME( _mesa_x86_transform_points2_2d_no_rot ) -HIDDEN(_mesa_x86_transform_points2_2d_no_rot) -GLNAME( _mesa_x86_transform_points2_2d_no_rot ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p2_2dnrr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p2_2dnrr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F1 F4 */ - FMUL_S( MAT5 ) - - FXCH( ST(1) ) /* F4 F1 */ - FADD_S( MAT12 ) - FLD_S( MAT13 ) /* F5 F4 F1 */ - FXCH( ST(2) ) /* F1 F4 F5 */ - FADDP( ST0, ST(2) ) /* F4 F5 */ - - FSTP_S( DST0 ) /* F5 */ - FSTP_S( DST1 ) /* */ - -LLBL(x86_p2_2dnrr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p2_2dnrr_loop) ) - -LLBL(x86_p2_2dnrr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points2_identity ) -HIDDEN(_mesa_x86_transform_points2_identity) -GLNAME( _mesa_x86_transform_points2_identity ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p2_ir_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - - CMP_L( ESI, EDI ) - JE( LLBL(x86_p2_ir_done) ) - -ALIGNTEXT16 -LLBL(x86_p2_ir_loop): - - MOV_L( SRC0, EBX ) - MOV_L( SRC1, EDX ) - - MOV_L( EBX, DST0 ) - MOV_L( EDX, DST1 ) - -LLBL(x86_p2_ir_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p2_ir_loop) ) - -LLBL(x86_p2_ir_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * NOTE: Avoid using spaces in between '(' ')' and arguments, especially + * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces + * in there will break the build on some platforms. + */ + +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FP_ONE 1065353216 +#define FP_ZERO 0 + +#define SRC0 REGOFF(0, ESI) +#define SRC1 REGOFF(4, ESI) +#define SRC2 REGOFF(8, ESI) +#define SRC3 REGOFF(12, ESI) +#define DST0 REGOFF(0, EDI) +#define DST1 REGOFF(4, EDI) +#define DST2 REGOFF(8, EDI) +#define DST3 REGOFF(12, EDI) +#define MAT0 REGOFF(0, EDX) +#define MAT1 REGOFF(4, EDX) +#define MAT2 REGOFF(8, EDX) +#define MAT3 REGOFF(12, EDX) +#define MAT4 REGOFF(16, EDX) +#define MAT5 REGOFF(20, EDX) +#define MAT6 REGOFF(24, EDX) +#define MAT7 REGOFF(28, EDX) +#define MAT8 REGOFF(32, EDX) +#define MAT9 REGOFF(36, EDX) +#define MAT10 REGOFF(40, EDX) +#define MAT11 REGOFF(44, EDX) +#define MAT12 REGOFF(48, EDX) +#define MAT13 REGOFF(52, EDX) +#define MAT14 REGOFF(56, EDX) +#define MAT15 REGOFF(60, EDX) + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points2_general ) +HIDDEN(_mesa_x86_transform_points2_general) +GLNAME( _mesa_x86_transform_points2_general ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p2_gr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p2_gr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + FLD_S( SRC0 ) /* F6 F5 F4 */ + FMUL_S( MAT2 ) + FLD_S( SRC0 ) /* F7 F6 F5 F4 */ + FMUL_S( MAT3 ) + + FLD_S( SRC1 ) /* F0 F7 F6 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT5 ) + FLD_S( SRC1 ) /* F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT6 ) + FLD_S( SRC1 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT7 ) + + FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ + FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ + + FXCH( ST(3) ) /* F4 F6 F5 F7 */ + FADD_S( MAT12 ) + FXCH( ST(2) ) /* F5 F6 F4 F7 */ + FADD_S( MAT13 ) + FXCH( ST(1) ) /* F6 F5 F4 F7 */ + FADD_S( MAT14 ) + FXCH( ST(3) ) /* F7 F5 F4 F6 */ + FADD_S( MAT15 ) + + FXCH( ST(2) ) /* F4 F5 F7 F6 */ + FSTP_S( DST0 ) /* F5 F7 F6 */ + FSTP_S( DST1 ) /* F7 F6 */ + FXCH( ST(1) ) /* F6 F7 */ + FSTP_S( DST2 ) /* F7 */ + FSTP_S( DST3 ) /* */ + +LLBL(x86_p2_gr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p2_gr_loop) ) + +LLBL(x86_p2_gr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points2_perspective ) +HIDDEN(_mesa_x86_transform_points2_perspective) +GLNAME( _mesa_x86_transform_points2_perspective ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p2_pr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + + MOV_L( MAT14, EBX ) + +ALIGNTEXT16 +LLBL(x86_p2_pr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F1 F4 */ + FMUL_S( MAT5 ) + + FXCH( ST(1) ) /* F4 F1 */ + FSTP_S( DST0 ) /* F1 */ + FSTP_S( DST1 ) /* */ + MOV_L( EBX, DST2 ) + MOV_L( CONST(FP_ZERO), DST3 ) + +LLBL(x86_p2_pr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p2_pr_loop) ) + +LLBL(x86_p2_pr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points2_3d ) +HIDDEN(_mesa_x86_transform_points2_3d) +GLNAME( _mesa_x86_transform_points2_3d ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p2_3dr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p2_3dr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + FLD_S( SRC0 ) /* F6 F5 F4 */ + FMUL_S( MAT2 ) + + FLD_S( SRC1 ) /* F0 F6 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F6 F5 F4 */ + FMUL_S( MAT5 ) + FLD_S( SRC1 ) /* F2 F1 F0 F6 F5 F4 */ + FMUL_S( MAT6 ) + + FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + FXCH( ST(2) ) /* F4 F5 F6 */ + FADD_S( MAT12 ) + FXCH( ST(1) ) /* F5 F4 F6 */ + FADD_S( MAT13 ) + FXCH( ST(2) ) /* F6 F4 F5 */ + FADD_S( MAT14 ) + + FXCH( ST(1) ) /* F4 F6 F5 */ + FSTP_S( DST0 ) /* F6 F5 */ + FXCH( ST(1) ) /* F5 F6 */ + FSTP_S( DST1 ) /* F6 */ + FSTP_S( DST2 ) /* */ + +LLBL(x86_p2_3dr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p2_3dr_loop) ) + +LLBL(x86_p2_3dr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points2_3d_no_rot ) +HIDDEN(_mesa_x86_transform_points2_3d_no_rot) +GLNAME( _mesa_x86_transform_points2_3d_no_rot ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p2_3dnrr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + + MOV_L( MAT14, EBX ) + +ALIGNTEXT16 +LLBL(x86_p2_3dnrr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F1 F4 */ + FMUL_S( MAT5 ) + + FXCH( ST(1) ) /* F4 F1 */ + FADD_S( MAT12 ) + FLD_S( MAT13 ) /* F5 F4 F1 */ + FXCH( ST(2) ) /* F1 F4 F5 */ + FADDP( ST0, ST(2) ) /* F4 F5 */ + + FSTP_S( DST0 ) /* F5 */ + FSTP_S( DST1 ) /* */ + MOV_L( EBX, DST2 ) + +LLBL(x86_p2_3dnrr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p2_3dnrr_loop) ) + +LLBL(x86_p2_3dnrr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points2_2d ) +HIDDEN(_mesa_x86_transform_points2_2d) +GLNAME( _mesa_x86_transform_points2_2d ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p2_2dr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p2_2dr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + + FLD_S( SRC1 ) /* F0 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F5 F4 */ + FMUL_S( MAT5 ) + + FXCH( ST(1) ) /* F0 F1 F5 F4 */ + FADDP( ST0, ST(3) ) /* F1 F5 F4 */ + FADDP( ST0, ST(1) ) /* F5 F4 */ + + FXCH( ST(1) ) /* F4 F5 */ + FADD_S( MAT12 ) + FXCH( ST(1) ) /* F5 F4 */ + FADD_S( MAT13 ) + + FXCH( ST(1) ) /* F4 F5 */ + FSTP_S( DST0 ) /* F5 */ + FSTP_S( DST1 ) /* */ + +LLBL(x86_p2_2dr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p2_2dr_loop) ) + +LLBL(x86_p2_2dr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT4 +GLOBL GLNAME( _mesa_x86_transform_points2_2d_no_rot ) +HIDDEN(_mesa_x86_transform_points2_2d_no_rot) +GLNAME( _mesa_x86_transform_points2_2d_no_rot ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p2_2dnrr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p2_2dnrr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F1 F4 */ + FMUL_S( MAT5 ) + + FXCH( ST(1) ) /* F4 F1 */ + FADD_S( MAT12 ) + FLD_S( MAT13 ) /* F5 F4 F1 */ + FXCH( ST(2) ) /* F1 F4 F5 */ + FADDP( ST0, ST(2) ) /* F4 F5 */ + + FSTP_S( DST0 ) /* F5 */ + FSTP_S( DST1 ) /* */ + +LLBL(x86_p2_2dnrr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p2_2dnrr_loop) ) + +LLBL(x86_p2_2dnrr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points2_identity ) +HIDDEN(_mesa_x86_transform_points2_identity) +GLNAME( _mesa_x86_transform_points2_identity ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p2_ir_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_2), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(2), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + + CMP_L( ESI, EDI ) + JE( LLBL(x86_p2_ir_done) ) + +ALIGNTEXT16 +LLBL(x86_p2_ir_loop): + + MOV_L( SRC0, EBX ) + MOV_L( SRC1, EDX ) + + MOV_L( EBX, DST0 ) + MOV_L( EDX, DST1 ) + +LLBL(x86_p2_ir_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p2_ir_loop) ) + +LLBL(x86_p2_ir_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/x86_xform3.S b/mesalib/src/mesa/x86/x86_xform3.S index 6f717caf9..1c782f1c5 100644 --- a/mesalib/src/mesa/x86/x86_xform3.S +++ b/mesalib/src/mesa/x86/x86_xform3.S @@ -1,644 +1,644 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * NOTE: Avoid using spaces in between '(' ')' and arguments, especially - * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces - * in there will break the build on some platforms. - */ - -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FP_ONE 1065353216 -#define FP_ZERO 0 - -#define SRC0 REGOFF(0, ESI) -#define SRC1 REGOFF(4, ESI) -#define SRC2 REGOFF(8, ESI) -#define SRC3 REGOFF(12, ESI) -#define DST0 REGOFF(0, EDI) -#define DST1 REGOFF(4, EDI) -#define DST2 REGOFF(8, EDI) -#define DST3 REGOFF(12, EDI) -#define MAT0 REGOFF(0, EDX) -#define MAT1 REGOFF(4, EDX) -#define MAT2 REGOFF(8, EDX) -#define MAT3 REGOFF(12, EDX) -#define MAT4 REGOFF(16, EDX) -#define MAT5 REGOFF(20, EDX) -#define MAT6 REGOFF(24, EDX) -#define MAT7 REGOFF(28, EDX) -#define MAT8 REGOFF(32, EDX) -#define MAT9 REGOFF(36, EDX) -#define MAT10 REGOFF(40, EDX) -#define MAT11 REGOFF(44, EDX) -#define MAT12 REGOFF(48, EDX) -#define MAT13 REGOFF(52, EDX) -#define MAT14 REGOFF(56, EDX) -#define MAT15 REGOFF(60, EDX) - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points3_general ) -HIDDEN(_mesa_x86_transform_points3_general) -GLNAME( _mesa_x86_transform_points3_general ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p3_gr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p3_gr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - FLD_S( SRC0 ) /* F6 F5 F4 */ - FMUL_S( MAT2 ) - FLD_S( SRC0 ) /* F7 F6 F5 F4 */ - FMUL_S( MAT3 ) - - FLD_S( SRC1 ) /* F0 F7 F6 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT5 ) - FLD_S( SRC1 ) /* F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT6 ) - FLD_S( SRC1 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT7 ) - - FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ - FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ - - FLD_S( SRC2 ) /* F0 F7 F6 F5 F4 */ - FMUL_S( MAT8 ) - FLD_S( SRC2 ) /* F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT9 ) - FLD_S( SRC2 ) /* F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT10 ) - FLD_S( SRC2 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT11 ) - - FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ - FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ - - FXCH( ST(3) ) /* F4 F6 F5 F7 */ - FADD_S( MAT12 ) - FXCH( ST(2) ) /* F5 F6 F4 F7 */ - FADD_S( MAT13 ) - FXCH( ST(1) ) /* F6 F5 F4 F7 */ - FADD_S( MAT14 ) - FXCH( ST(3) ) /* F7 F5 F4 F6 */ - FADD_S( MAT15 ) - - FXCH( ST(2) ) /* F4 F5 F7 F6 */ - FSTP_S( DST0 ) /* F5 F7 F6 */ - FSTP_S( DST1 ) /* F7 F6 */ - FXCH( ST(1) ) /* F6 F7 */ - FSTP_S( DST2 ) /* F7 */ - FSTP_S( DST3 ) /* */ - -LLBL(x86_p3_gr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p3_gr_loop) ) - -LLBL(x86_p3_gr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points3_perspective ) -HIDDEN(_mesa_x86_transform_points3_perspective) -GLNAME( _mesa_x86_transform_points3_perspective ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p3_pr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p3_pr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F5 F4 */ - FMUL_S( MAT5 ) - - FLD_S( SRC2 ) /* F0 F5 F4 */ - FMUL_S( MAT8 ) - FLD_S( SRC2 ) /* F1 F0 F5 F4 */ - FMUL_S( MAT9 ) - FLD_S( SRC2 ) /* F2 F1 F0 F5 F4 */ - FMUL_S( MAT10 ) - - FXCH( ST(2) ) /* F0 F1 F2 F5 F4 */ - FADDP( ST0, ST(4) ) /* F1 F2 F5 F4 */ - FADDP( ST0, ST(2) ) /* F2 F5 F4 */ - FLD_S( MAT14 ) /* F6 F2 F5 F4 */ - FXCH( ST(1) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - MOV_L( SRC2, EBX ) - XOR_L( CONST(-2147483648), EBX )/* change sign */ - - FXCH( ST(2) ) /* F4 F5 F6 */ - FSTP_S( DST0 ) /* F5 F6 */ - FSTP_S( DST1 ) /* F6 */ - FSTP_S( DST2 ) /* */ - MOV_L( EBX, DST3 ) - -LLBL(x86_p3_pr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p3_pr_loop) ) - -LLBL(x86_p3_pr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points3_3d ) -HIDDEN(_mesa_x86_transform_points3_3d) -GLNAME( _mesa_x86_transform_points3_3d ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p3_3dr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p3_3dr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - FLD_S( SRC0 ) /* F6 F5 F4 */ - FMUL_S( MAT2 ) - - FLD_S( SRC1 ) /* F0 F6 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F6 F5 F4 */ - FMUL_S( MAT5 ) - FLD_S( SRC1 ) /* F2 F1 F0 F6 F5 F4 */ - FMUL_S( MAT6 ) - - FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - FLD_S( SRC2 ) /* F0 F6 F5 F4 */ - FMUL_S( MAT8 ) - FLD_S( SRC2 ) /* F1 F0 F6 F5 F4 */ - FMUL_S( MAT9 ) - FLD_S( SRC2 ) /* F2 F1 F0 F6 F5 F4 */ - FMUL_S( MAT10 ) - - FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - FXCH( ST(2) ) /* F4 F5 F6 */ - FADD_S( MAT12 ) - FXCH( ST(1) ) /* F5 F4 F6 */ - FADD_S( MAT13 ) - FXCH( ST(2) ) /* F6 F4 F5 */ - FADD_S( MAT14 ) - - FXCH( ST(1) ) /* F4 F6 F5 */ - FSTP_S( DST0 ) /* F6 F5 */ - FXCH( ST(1) ) /* F5 F6 */ - FSTP_S( DST1 ) /* F6 */ - FSTP_S( DST2 ) /* */ - -LLBL(x86_p3_3dr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p3_3dr_loop) ) - -LLBL(x86_p3_3dr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points3_3d_no_rot ) -HIDDEN(_mesa_x86_transform_points3_3d_no_rot) -GLNAME( _mesa_x86_transform_points3_3d_no_rot ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p3_3dnrr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p3_3dnrr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F1 F4 */ - FMUL_S( MAT5 ) - - FLD_S( SRC2 ) /* F2 F1 F4 */ - FMUL_S( MAT10 ) - - FXCH( ST(2) ) /* F4 F1 F2 */ - FADD_S( MAT12 ) - FLD_S( MAT13 ) /* F5 F4 F1 F2 */ - FXCH( ST(2) ) /* F1 F4 F5 F2 */ - FADDP( ST0, ST(2) ) /* F4 F5 F2 */ - FLD_S( MAT14 ) /* F6 F4 F5 F2 */ - FXCH( ST(3) ) /* F2 F4 F5 F6 */ - FADDP( ST0, ST(3) ) /* F4 F5 F6 */ - - FSTP_S( DST0 ) /* F5 F6 */ - FSTP_S( DST1 ) /* F6 */ - FSTP_S( DST2 ) /* */ - -LLBL(x86_p3_3dnrr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p3_3dnrr_loop) ) - -LLBL(x86_p3_3dnrr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points3_2d ) -HIDDEN(_mesa_x86_transform_points3_2d) -GLNAME( _mesa_x86_transform_points3_2d ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p3_2dr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p3_2dr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - - FLD_S( SRC1 ) /* F0 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F5 F4 */ - FMUL_S( MAT5 ) - - FXCH( ST(1) ) /* F0 F1 F5 F4 */ - FADDP( ST0, ST(3) ) /* F1 F5 F4 */ - FADDP( ST0, ST(1) ) /* F5 F4 */ - - FXCH( ST(1) ) /* F4 F5 */ - FADD_S( MAT12 ) - FXCH( ST(1) ) /* F5 F4 */ - FADD_S( MAT13 ) - - MOV_L( SRC2, EBX ) - - FXCH( ST(1) ) /* F4 F5 */ - FSTP_S( DST0 ) /* F5 */ - FSTP_S( DST1 ) /* */ - MOV_L( EBX, DST2 ) - -LLBL(x86_p3_2dr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p3_2dr_loop) ) - -LLBL(x86_p3_2dr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points3_2d_no_rot ) -HIDDEN(_mesa_x86_transform_points3_2d_no_rot) -GLNAME( _mesa_x86_transform_points3_2d_no_rot ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p3_2dnrr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p3_2dnrr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F1 F4 */ - FMUL_S( MAT5 ) - - FXCH( ST(1) ) /* F4 F1 */ - FADD_S( MAT12 ) - FLD_S( MAT13 ) /* F5 F4 F1 */ - - FXCH( ST(2) ) /* F1 F4 F5 */ - FADDP( ST0, ST(2) ) /* F4 F5 */ - - MOV_L( SRC2, EBX ) - - FSTP_S( DST0 ) /* F5 */ - FSTP_S( DST1 ) /* */ - MOV_L( EBX, DST2 ) - -LLBL(x86_p3_2dnrr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p3_2dnrr_loop) ) - -LLBL(x86_p3_2dnrr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points3_identity ) -HIDDEN(_mesa_x86_transform_points3_identity) -GLNAME(_mesa_x86_transform_points3_identity ): - -#define FRAME_OFFSET 16 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - PUSH_L( EBP ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p3_ir_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - - CMP_L( ESI, EDI ) - JE( LLBL(x86_p3_ir_done) ) - -ALIGNTEXT16 -LLBL(x86_p3_ir_loop): - -#if 1 - MOV_L( SRC0, EBX ) - MOV_L( SRC1, EBP ) - MOV_L( SRC2, EDX ) - - MOV_L( EBX, DST0 ) - MOV_L( EBP, DST1 ) - MOV_L( EDX, DST2 ) -#else - FLD_S( SRC0 ) - FLD_S( SRC1 ) - FLD_S( SRC2 ) - - FSTP_S( DST2 ) - FSTP_S( DST1 ) - FSTP_S( DST0 ) -#endif - -LLBL(x86_p3_ir_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p3_ir_loop) ) - -LLBL(x86_p3_ir_done): - - POP_L( EBP ) - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * NOTE: Avoid using spaces in between '(' ')' and arguments, especially + * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces + * in there will break the build on some platforms. + */ + +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FP_ONE 1065353216 +#define FP_ZERO 0 + +#define SRC0 REGOFF(0, ESI) +#define SRC1 REGOFF(4, ESI) +#define SRC2 REGOFF(8, ESI) +#define SRC3 REGOFF(12, ESI) +#define DST0 REGOFF(0, EDI) +#define DST1 REGOFF(4, EDI) +#define DST2 REGOFF(8, EDI) +#define DST3 REGOFF(12, EDI) +#define MAT0 REGOFF(0, EDX) +#define MAT1 REGOFF(4, EDX) +#define MAT2 REGOFF(8, EDX) +#define MAT3 REGOFF(12, EDX) +#define MAT4 REGOFF(16, EDX) +#define MAT5 REGOFF(20, EDX) +#define MAT6 REGOFF(24, EDX) +#define MAT7 REGOFF(28, EDX) +#define MAT8 REGOFF(32, EDX) +#define MAT9 REGOFF(36, EDX) +#define MAT10 REGOFF(40, EDX) +#define MAT11 REGOFF(44, EDX) +#define MAT12 REGOFF(48, EDX) +#define MAT13 REGOFF(52, EDX) +#define MAT14 REGOFF(56, EDX) +#define MAT15 REGOFF(60, EDX) + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points3_general ) +HIDDEN(_mesa_x86_transform_points3_general) +GLNAME( _mesa_x86_transform_points3_general ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p3_gr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p3_gr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + FLD_S( SRC0 ) /* F6 F5 F4 */ + FMUL_S( MAT2 ) + FLD_S( SRC0 ) /* F7 F6 F5 F4 */ + FMUL_S( MAT3 ) + + FLD_S( SRC1 ) /* F0 F7 F6 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT5 ) + FLD_S( SRC1 ) /* F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT6 ) + FLD_S( SRC1 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT7 ) + + FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ + FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ + + FLD_S( SRC2 ) /* F0 F7 F6 F5 F4 */ + FMUL_S( MAT8 ) + FLD_S( SRC2 ) /* F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT9 ) + FLD_S( SRC2 ) /* F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT10 ) + FLD_S( SRC2 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT11 ) + + FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ + FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ + + FXCH( ST(3) ) /* F4 F6 F5 F7 */ + FADD_S( MAT12 ) + FXCH( ST(2) ) /* F5 F6 F4 F7 */ + FADD_S( MAT13 ) + FXCH( ST(1) ) /* F6 F5 F4 F7 */ + FADD_S( MAT14 ) + FXCH( ST(3) ) /* F7 F5 F4 F6 */ + FADD_S( MAT15 ) + + FXCH( ST(2) ) /* F4 F5 F7 F6 */ + FSTP_S( DST0 ) /* F5 F7 F6 */ + FSTP_S( DST1 ) /* F7 F6 */ + FXCH( ST(1) ) /* F6 F7 */ + FSTP_S( DST2 ) /* F7 */ + FSTP_S( DST3 ) /* */ + +LLBL(x86_p3_gr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p3_gr_loop) ) + +LLBL(x86_p3_gr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points3_perspective ) +HIDDEN(_mesa_x86_transform_points3_perspective) +GLNAME( _mesa_x86_transform_points3_perspective ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p3_pr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p3_pr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F5 F4 */ + FMUL_S( MAT5 ) + + FLD_S( SRC2 ) /* F0 F5 F4 */ + FMUL_S( MAT8 ) + FLD_S( SRC2 ) /* F1 F0 F5 F4 */ + FMUL_S( MAT9 ) + FLD_S( SRC2 ) /* F2 F1 F0 F5 F4 */ + FMUL_S( MAT10 ) + + FXCH( ST(2) ) /* F0 F1 F2 F5 F4 */ + FADDP( ST0, ST(4) ) /* F1 F2 F5 F4 */ + FADDP( ST0, ST(2) ) /* F2 F5 F4 */ + FLD_S( MAT14 ) /* F6 F2 F5 F4 */ + FXCH( ST(1) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + MOV_L( SRC2, EBX ) + XOR_L( CONST(-2147483648), EBX )/* change sign */ + + FXCH( ST(2) ) /* F4 F5 F6 */ + FSTP_S( DST0 ) /* F5 F6 */ + FSTP_S( DST1 ) /* F6 */ + FSTP_S( DST2 ) /* */ + MOV_L( EBX, DST3 ) + +LLBL(x86_p3_pr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p3_pr_loop) ) + +LLBL(x86_p3_pr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points3_3d ) +HIDDEN(_mesa_x86_transform_points3_3d) +GLNAME( _mesa_x86_transform_points3_3d ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p3_3dr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p3_3dr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + FLD_S( SRC0 ) /* F6 F5 F4 */ + FMUL_S( MAT2 ) + + FLD_S( SRC1 ) /* F0 F6 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F6 F5 F4 */ + FMUL_S( MAT5 ) + FLD_S( SRC1 ) /* F2 F1 F0 F6 F5 F4 */ + FMUL_S( MAT6 ) + + FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + FLD_S( SRC2 ) /* F0 F6 F5 F4 */ + FMUL_S( MAT8 ) + FLD_S( SRC2 ) /* F1 F0 F6 F5 F4 */ + FMUL_S( MAT9 ) + FLD_S( SRC2 ) /* F2 F1 F0 F6 F5 F4 */ + FMUL_S( MAT10 ) + + FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + FXCH( ST(2) ) /* F4 F5 F6 */ + FADD_S( MAT12 ) + FXCH( ST(1) ) /* F5 F4 F6 */ + FADD_S( MAT13 ) + FXCH( ST(2) ) /* F6 F4 F5 */ + FADD_S( MAT14 ) + + FXCH( ST(1) ) /* F4 F6 F5 */ + FSTP_S( DST0 ) /* F6 F5 */ + FXCH( ST(1) ) /* F5 F6 */ + FSTP_S( DST1 ) /* F6 */ + FSTP_S( DST2 ) /* */ + +LLBL(x86_p3_3dr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p3_3dr_loop) ) + +LLBL(x86_p3_3dr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points3_3d_no_rot ) +HIDDEN(_mesa_x86_transform_points3_3d_no_rot) +GLNAME( _mesa_x86_transform_points3_3d_no_rot ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p3_3dnrr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p3_3dnrr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F1 F4 */ + FMUL_S( MAT5 ) + + FLD_S( SRC2 ) /* F2 F1 F4 */ + FMUL_S( MAT10 ) + + FXCH( ST(2) ) /* F4 F1 F2 */ + FADD_S( MAT12 ) + FLD_S( MAT13 ) /* F5 F4 F1 F2 */ + FXCH( ST(2) ) /* F1 F4 F5 F2 */ + FADDP( ST0, ST(2) ) /* F4 F5 F2 */ + FLD_S( MAT14 ) /* F6 F4 F5 F2 */ + FXCH( ST(3) ) /* F2 F4 F5 F6 */ + FADDP( ST0, ST(3) ) /* F4 F5 F6 */ + + FSTP_S( DST0 ) /* F5 F6 */ + FSTP_S( DST1 ) /* F6 */ + FSTP_S( DST2 ) /* */ + +LLBL(x86_p3_3dnrr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p3_3dnrr_loop) ) + +LLBL(x86_p3_3dnrr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points3_2d ) +HIDDEN(_mesa_x86_transform_points3_2d) +GLNAME( _mesa_x86_transform_points3_2d ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p3_2dr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p3_2dr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + + FLD_S( SRC1 ) /* F0 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F5 F4 */ + FMUL_S( MAT5 ) + + FXCH( ST(1) ) /* F0 F1 F5 F4 */ + FADDP( ST0, ST(3) ) /* F1 F5 F4 */ + FADDP( ST0, ST(1) ) /* F5 F4 */ + + FXCH( ST(1) ) /* F4 F5 */ + FADD_S( MAT12 ) + FXCH( ST(1) ) /* F5 F4 */ + FADD_S( MAT13 ) + + MOV_L( SRC2, EBX ) + + FXCH( ST(1) ) /* F4 F5 */ + FSTP_S( DST0 ) /* F5 */ + FSTP_S( DST1 ) /* */ + MOV_L( EBX, DST2 ) + +LLBL(x86_p3_2dr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p3_2dr_loop) ) + +LLBL(x86_p3_2dr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points3_2d_no_rot ) +HIDDEN(_mesa_x86_transform_points3_2d_no_rot) +GLNAME( _mesa_x86_transform_points3_2d_no_rot ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p3_2dnrr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p3_2dnrr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F1 F4 */ + FMUL_S( MAT5 ) + + FXCH( ST(1) ) /* F4 F1 */ + FADD_S( MAT12 ) + FLD_S( MAT13 ) /* F5 F4 F1 */ + + FXCH( ST(2) ) /* F1 F4 F5 */ + FADDP( ST0, ST(2) ) /* F4 F5 */ + + MOV_L( SRC2, EBX ) + + FSTP_S( DST0 ) /* F5 */ + FSTP_S( DST1 ) /* */ + MOV_L( EBX, DST2 ) + +LLBL(x86_p3_2dnrr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p3_2dnrr_loop) ) + +LLBL(x86_p3_2dnrr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points3_identity ) +HIDDEN(_mesa_x86_transform_points3_identity) +GLNAME(_mesa_x86_transform_points3_identity ): + +#define FRAME_OFFSET 16 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + PUSH_L( EBP ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p3_ir_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_3), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(3), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + + CMP_L( ESI, EDI ) + JE( LLBL(x86_p3_ir_done) ) + +ALIGNTEXT16 +LLBL(x86_p3_ir_loop): + +#if 1 + MOV_L( SRC0, EBX ) + MOV_L( SRC1, EBP ) + MOV_L( SRC2, EDX ) + + MOV_L( EBX, DST0 ) + MOV_L( EBP, DST1 ) + MOV_L( EDX, DST2 ) +#else + FLD_S( SRC0 ) + FLD_S( SRC1 ) + FLD_S( SRC2 ) + + FSTP_S( DST2 ) + FSTP_S( DST1 ) + FSTP_S( DST0 ) +#endif + +LLBL(x86_p3_ir_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p3_ir_loop) ) + +LLBL(x86_p3_ir_done): + + POP_L( EBP ) + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/x86_xform4.S b/mesalib/src/mesa/x86/x86_xform4.S index 687426eb5..97a841138 100644 --- a/mesalib/src/mesa/x86/x86_xform4.S +++ b/mesalib/src/mesa/x86/x86_xform4.S @@ -1,677 +1,677 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * NOTE: Avoid using spaces in between '(' ')' and arguments, especially - * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces - * in there will break the build on some platforms. - */ - -#include "assyntax.h" -#include "matypes.h" -#include "xform_args.h" - - SEG_TEXT - -#define FP_ONE 1065353216 -#define FP_ZERO 0 - -#define SRC0 REGOFF(0, ESI) -#define SRC1 REGOFF(4, ESI) -#define SRC2 REGOFF(8, ESI) -#define SRC3 REGOFF(12, ESI) -#define DST0 REGOFF(0, EDI) -#define DST1 REGOFF(4, EDI) -#define DST2 REGOFF(8, EDI) -#define DST3 REGOFF(12, EDI) -#define MAT0 REGOFF(0, EDX) -#define MAT1 REGOFF(4, EDX) -#define MAT2 REGOFF(8, EDX) -#define MAT3 REGOFF(12, EDX) -#define MAT4 REGOFF(16, EDX) -#define MAT5 REGOFF(20, EDX) -#define MAT6 REGOFF(24, EDX) -#define MAT7 REGOFF(28, EDX) -#define MAT8 REGOFF(32, EDX) -#define MAT9 REGOFF(36, EDX) -#define MAT10 REGOFF(40, EDX) -#define MAT11 REGOFF(44, EDX) -#define MAT12 REGOFF(48, EDX) -#define MAT13 REGOFF(52, EDX) -#define MAT14 REGOFF(56, EDX) -#define MAT15 REGOFF(60, EDX) - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points4_general ) -HIDDEN(_mesa_x86_transform_points4_general) -GLNAME( _mesa_x86_transform_points4_general ): - -#define FRAME_OFFSET 8 - PUSH_L( ESI ) - PUSH_L( EDI ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p4_gr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p4_gr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - FLD_S( SRC0 ) /* F6 F5 F4 */ - FMUL_S( MAT2 ) - FLD_S( SRC0 ) /* F7 F6 F5 F4 */ - FMUL_S( MAT3 ) - - FLD_S( SRC1 ) /* F0 F7 F6 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT5 ) - FLD_S( SRC1 ) /* F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT6 ) - FLD_S( SRC1 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT7 ) - - FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ - FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ - - FLD_S( SRC2 ) /* F0 F7 F6 F5 F4 */ - FMUL_S( MAT8 ) - FLD_S( SRC2 ) /* F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT9 ) - FLD_S( SRC2 ) /* F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT10 ) - FLD_S( SRC2 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT11 ) - - FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ - FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ - - FLD_S( SRC3 ) /* F0 F7 F6 F5 F4 */ - FMUL_S( MAT12 ) - FLD_S( SRC3 ) /* F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT13 ) - FLD_S( SRC3 ) /* F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT14 ) - FLD_S( SRC3 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ - FMUL_S( MAT15 ) - - FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ - FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ - - FXCH( ST(3) ) /* F4 F6 F5 F7 */ - FSTP_S( DST0 ) /* F6 F5 F7 */ - FXCH( ST(1) ) /* F5 F6 F7 */ - FSTP_S( DST1 ) /* F6 F7 */ - FSTP_S( DST2 ) /* F7 */ - FSTP_S( DST3 ) /* */ - -LLBL(x86_p4_gr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p4_gr_loop) ) - -LLBL(x86_p4_gr_done): - - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points4_perspective ) -HIDDEN(_mesa_x86_transform_points4_perspective) -GLNAME( _mesa_x86_transform_points4_perspective ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p4_pr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p4_pr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F5 F4 */ - FMUL_S( MAT5 ) - - FLD_S( SRC2 ) /* F0 F5 F4 */ - FMUL_S( MAT8 ) - FLD_S( SRC2 ) /* F1 F0 F5 F4 */ - FMUL_S( MAT9 ) - FLD_S( SRC2 ) /* F6 F1 F0 F5 F4 */ - FMUL_S( MAT10 ) - - FXCH( ST(2) ) /* F0 F1 F6 F5 F4 */ - FADDP( ST0, ST(4) ) /* F1 F6 F5 F4 */ - FADDP( ST0, ST(2) ) /* F6 F5 F4 */ - - FLD_S( SRC3 ) /* F2 F6 F5 F4 */ - FMUL_S( MAT14 ) - - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - MOV_L( SRC2, EBX ) - XOR_L( CONST(-2147483648), EBX )/* change sign */ - - FXCH( ST(2) ) /* F4 F5 F6 */ - FSTP_S( DST0 ) /* F5 F6 */ - FSTP_S( DST1 ) /* F6 */ - FSTP_S( DST2 ) /* */ - MOV_L( EBX, DST3 ) - -LLBL(x86_p4_pr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p4_pr_loop) ) - -LLBL(x86_p4_pr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points4_3d ) -HIDDEN(_mesa_x86_transform_points4_3d) -GLNAME( _mesa_x86_transform_points4_3d ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p4_3dr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p4_3dr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - FLD_S( SRC0 ) /* F6 F5 F4 */ - FMUL_S( MAT2 ) - - FLD_S( SRC1 ) /* F0 F6 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F6 F5 F4 */ - FMUL_S( MAT5 ) - FLD_S( SRC1 ) /* F2 F1 F0 F6 F5 F4 */ - FMUL_S( MAT6 ) - - FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - FLD_S( SRC2 ) /* F0 F6 F5 F4 */ - FMUL_S( MAT8 ) - FLD_S( SRC2 ) /* F1 F0 F6 F5 F4 */ - FMUL_S( MAT9 ) - FLD_S( SRC2 ) /* F2 F1 F0 F6 F5 F4 */ - FMUL_S( MAT10 ) - - FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - FLD_S( SRC3 ) /* F0 F6 F5 F4 */ - FMUL_S( MAT12 ) - FLD_S( SRC3 ) /* F1 F0 F6 F5 F4 */ - FMUL_S( MAT13 ) - FLD_S( SRC3 ) /* F2 F1 F0 F6 F5 F4 */ - FMUL_S( MAT14 ) - - FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - MOV_L( SRC3, EBX ) - - FXCH( ST(2) ) /* F4 F5 F6 */ - FSTP_S( DST0 ) /* F5 F6 */ - FSTP_S( DST1 ) /* F6 */ - FSTP_S( DST2 ) /* */ - MOV_L( EBX, DST3 ) - -LLBL(x86_p4_3dr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p4_3dr_loop) ) - -LLBL(x86_p4_3dr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME(_mesa_x86_transform_points4_3d_no_rot) -HIDDEN(_mesa_x86_transform_points4_3d_no_rot) -GLNAME(_mesa_x86_transform_points4_3d_no_rot): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p4_3dnrr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p4_3dnrr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F5 F4 */ - FMUL_S( MAT5 ) - - FLD_S( SRC2 ) /* F6 F5 F4 */ - FMUL_S( MAT10 ) - - FLD_S( SRC3 ) /* F0 F6 F5 F4 */ - FMUL_S( MAT12 ) - FLD_S( SRC3 ) /* F1 F0 F6 F5 F4 */ - FMUL_S( MAT13 ) - FLD_S( SRC3 ) /* F2 F1 F0 F6 F5 F4 */ - FMUL_S( MAT14 ) - - FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ - FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ - FADDP( ST0, ST(1) ) /* F6 F5 F4 */ - - MOV_L( SRC3, EBX ) - - FXCH( ST(2) ) /* F4 F5 F6 */ - FSTP_S( DST0 ) /* F5 F6 */ - FSTP_S( DST1 ) /* F6 */ - FSTP_S( DST2 ) /* */ - MOV_L( EBX, DST3 ) - -LLBL(x86_p4_3dnrr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p4_3dnrr_loop) ) - -LLBL(x86_p4_3dnrr_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points4_2d ) -HIDDEN(_mesa_x86_transform_points4_2d) -GLNAME( _mesa_x86_transform_points4_2d ): - -#define FRAME_OFFSET 16 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - PUSH_L( EBP ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p4_2dr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p4_2dr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - FLD_S( SRC0 ) /* F5 F4 */ - FMUL_S( MAT1 ) - - FLD_S( SRC1 ) /* F0 F5 F4 */ - FMUL_S( MAT4 ) - FLD_S( SRC1 ) /* F1 F0 F5 F4 */ - FMUL_S( MAT5 ) - - FXCH( ST(1) ) /* F0 F1 F5 F4 */ - FADDP( ST0, ST(3) ) /* F1 F5 F4 */ - FADDP( ST0, ST(1) ) /* F5 F4 */ - - FLD_S( SRC3 ) /* F0 F5 F4 */ - FMUL_S( MAT12 ) - FLD_S( SRC3 ) /* F1 F0 F5 F4 */ - FMUL_S( MAT13 ) - - FXCH( ST(1) ) /* F0 F1 F5 F4 */ - FADDP( ST0, ST(3) ) /* F1 F5 F4 */ - FADDP( ST0, ST(1) ) /* F5 F4 */ - - MOV_L( SRC2, EBX ) - MOV_L( SRC3, EBP ) - - FXCH( ST(1) ) /* F4 F5 */ - FSTP_S( DST0 ) /* F5 */ - FSTP_S( DST1 ) /* */ - MOV_L( EBX, DST2 ) - MOV_L( EBP, DST3 ) - -LLBL(x86_p4_2dr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p4_2dr_loop) ) - -LLBL(x86_p4_2dr_done): - - POP_L( EBP ) - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points4_2d_no_rot ) -HIDDEN(_mesa_x86_transform_points4_2d_no_rot) -GLNAME( _mesa_x86_transform_points4_2d_no_rot ): - -#define FRAME_OFFSET 16 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - PUSH_L( EBP ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p4_2dnrr_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - -ALIGNTEXT16 -LLBL(x86_p4_2dnrr_loop): - - FLD_S( SRC0 ) /* F4 */ - FMUL_S( MAT0 ) - - FLD_S( SRC1 ) /* F5 F4 */ - FMUL_S( MAT5 ) - - FLD_S( SRC3 ) /* F0 F5 F4 */ - FMUL_S( MAT12 ) - FLD_S( SRC3 ) /* F1 F0 F5 F4 */ - FMUL_S( MAT13 ) - - FXCH( ST(1) ) /* F0 F1 F5 F4 */ - FADDP( ST0, ST(3) ) /* F1 F5 F4 */ - FADDP( ST0, ST(1) ) /* F5 F4 */ - - MOV_L( SRC2, EBX ) - MOV_L( SRC3, EBP ) - - FXCH( ST(1) ) /* F4 F5 */ - FSTP_S( DST0 ) /* F5 */ - FSTP_S( DST1 ) /* */ - MOV_L( EBX, DST2 ) - MOV_L( EBP, DST3 ) - -LLBL(x86_p4_2dnrr_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p4_2dnrr_loop) ) - -LLBL(x86_p4_2dnrr_done): - - POP_L( EBP ) - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET -#undef FRAME_OFFSET - - - - -ALIGNTEXT16 -GLOBL GLNAME( _mesa_x86_transform_points4_identity ) -HIDDEN(_mesa_x86_transform_points4_identity) -GLNAME( _mesa_x86_transform_points4_identity ): - -#define FRAME_OFFSET 12 - PUSH_L( ESI ) - PUSH_L( EDI ) - PUSH_L( EBX ) - - MOV_L( ARG_SOURCE, ESI ) - MOV_L( ARG_DEST, EDI ) - - MOV_L( ARG_MATRIX, EDX ) - MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) - - TEST_L( ECX, ECX ) - JZ( LLBL(x86_p4_ir_done) ) - - MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) - OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) - - MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) - MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) - - SHL_L( CONST(4), ECX ) - MOV_L( REGOFF(V4F_START, ESI), ESI ) - - MOV_L( REGOFF(V4F_START, EDI), EDI ) - ADD_L( EDI, ECX ) - - CMP_L( ESI, EDI ) - JE( LLBL(x86_p4_ir_done) ) - -ALIGNTEXT16 -LLBL(x86_p4_ir_loop): - - MOV_L( SRC0, EBX ) - MOV_L( SRC1, EDX ) - - MOV_L( EBX, DST0 ) - MOV_L( EDX, DST1 ) - - MOV_L( SRC2, EBX ) - MOV_L( SRC3, EDX ) - - MOV_L( EBX, DST2 ) - MOV_L( EDX, DST3 ) - -LLBL(x86_p4_ir_skip): - - ADD_L( CONST(16), EDI ) - ADD_L( EAX, ESI ) - CMP_L( ECX, EDI ) - JNE( LLBL(x86_p4_ir_loop) ) - -LLBL(x86_p4_ir_done): - - POP_L( EBX ) - POP_L( EDI ) - POP_L( ESI ) - RET - -#if defined (__ELF__) && defined (__linux__) - .section .note.GNU-stack,"",%progbits -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * NOTE: Avoid using spaces in between '(' ')' and arguments, especially + * with macros like CONST, LLBL that expand to CONCAT(...). Putting spaces + * in there will break the build on some platforms. + */ + +#include "assyntax.h" +#include "matypes.h" +#include "xform_args.h" + + SEG_TEXT + +#define FP_ONE 1065353216 +#define FP_ZERO 0 + +#define SRC0 REGOFF(0, ESI) +#define SRC1 REGOFF(4, ESI) +#define SRC2 REGOFF(8, ESI) +#define SRC3 REGOFF(12, ESI) +#define DST0 REGOFF(0, EDI) +#define DST1 REGOFF(4, EDI) +#define DST2 REGOFF(8, EDI) +#define DST3 REGOFF(12, EDI) +#define MAT0 REGOFF(0, EDX) +#define MAT1 REGOFF(4, EDX) +#define MAT2 REGOFF(8, EDX) +#define MAT3 REGOFF(12, EDX) +#define MAT4 REGOFF(16, EDX) +#define MAT5 REGOFF(20, EDX) +#define MAT6 REGOFF(24, EDX) +#define MAT7 REGOFF(28, EDX) +#define MAT8 REGOFF(32, EDX) +#define MAT9 REGOFF(36, EDX) +#define MAT10 REGOFF(40, EDX) +#define MAT11 REGOFF(44, EDX) +#define MAT12 REGOFF(48, EDX) +#define MAT13 REGOFF(52, EDX) +#define MAT14 REGOFF(56, EDX) +#define MAT15 REGOFF(60, EDX) + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points4_general ) +HIDDEN(_mesa_x86_transform_points4_general) +GLNAME( _mesa_x86_transform_points4_general ): + +#define FRAME_OFFSET 8 + PUSH_L( ESI ) + PUSH_L( EDI ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p4_gr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p4_gr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + FLD_S( SRC0 ) /* F6 F5 F4 */ + FMUL_S( MAT2 ) + FLD_S( SRC0 ) /* F7 F6 F5 F4 */ + FMUL_S( MAT3 ) + + FLD_S( SRC1 ) /* F0 F7 F6 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT5 ) + FLD_S( SRC1 ) /* F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT6 ) + FLD_S( SRC1 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT7 ) + + FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ + FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ + + FLD_S( SRC2 ) /* F0 F7 F6 F5 F4 */ + FMUL_S( MAT8 ) + FLD_S( SRC2 ) /* F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT9 ) + FLD_S( SRC2 ) /* F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT10 ) + FLD_S( SRC2 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT11 ) + + FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ + FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ + + FLD_S( SRC3 ) /* F0 F7 F6 F5 F4 */ + FMUL_S( MAT12 ) + FLD_S( SRC3 ) /* F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT13 ) + FLD_S( SRC3 ) /* F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT14 ) + FLD_S( SRC3 ) /* F3 F2 F1 F0 F7 F6 F5 F4 */ + FMUL_S( MAT15 ) + + FXCH( ST(3) ) /* F0 F2 F1 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(7) ) /* F2 F1 F3 F7 F6 F5 F4 */ + FXCH( ST(1) ) /* F1 F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F2 F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F3 F7 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F7 F6 F5 F4 */ + + FXCH( ST(3) ) /* F4 F6 F5 F7 */ + FSTP_S( DST0 ) /* F6 F5 F7 */ + FXCH( ST(1) ) /* F5 F6 F7 */ + FSTP_S( DST1 ) /* F6 F7 */ + FSTP_S( DST2 ) /* F7 */ + FSTP_S( DST3 ) /* */ + +LLBL(x86_p4_gr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p4_gr_loop) ) + +LLBL(x86_p4_gr_done): + + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points4_perspective ) +HIDDEN(_mesa_x86_transform_points4_perspective) +GLNAME( _mesa_x86_transform_points4_perspective ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p4_pr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p4_pr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F5 F4 */ + FMUL_S( MAT5 ) + + FLD_S( SRC2 ) /* F0 F5 F4 */ + FMUL_S( MAT8 ) + FLD_S( SRC2 ) /* F1 F0 F5 F4 */ + FMUL_S( MAT9 ) + FLD_S( SRC2 ) /* F6 F1 F0 F5 F4 */ + FMUL_S( MAT10 ) + + FXCH( ST(2) ) /* F0 F1 F6 F5 F4 */ + FADDP( ST0, ST(4) ) /* F1 F6 F5 F4 */ + FADDP( ST0, ST(2) ) /* F6 F5 F4 */ + + FLD_S( SRC3 ) /* F2 F6 F5 F4 */ + FMUL_S( MAT14 ) + + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + MOV_L( SRC2, EBX ) + XOR_L( CONST(-2147483648), EBX )/* change sign */ + + FXCH( ST(2) ) /* F4 F5 F6 */ + FSTP_S( DST0 ) /* F5 F6 */ + FSTP_S( DST1 ) /* F6 */ + FSTP_S( DST2 ) /* */ + MOV_L( EBX, DST3 ) + +LLBL(x86_p4_pr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p4_pr_loop) ) + +LLBL(x86_p4_pr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points4_3d ) +HIDDEN(_mesa_x86_transform_points4_3d) +GLNAME( _mesa_x86_transform_points4_3d ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p4_3dr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p4_3dr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + FLD_S( SRC0 ) /* F6 F5 F4 */ + FMUL_S( MAT2 ) + + FLD_S( SRC1 ) /* F0 F6 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F6 F5 F4 */ + FMUL_S( MAT5 ) + FLD_S( SRC1 ) /* F2 F1 F0 F6 F5 F4 */ + FMUL_S( MAT6 ) + + FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + FLD_S( SRC2 ) /* F0 F6 F5 F4 */ + FMUL_S( MAT8 ) + FLD_S( SRC2 ) /* F1 F0 F6 F5 F4 */ + FMUL_S( MAT9 ) + FLD_S( SRC2 ) /* F2 F1 F0 F6 F5 F4 */ + FMUL_S( MAT10 ) + + FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + FLD_S( SRC3 ) /* F0 F6 F5 F4 */ + FMUL_S( MAT12 ) + FLD_S( SRC3 ) /* F1 F0 F6 F5 F4 */ + FMUL_S( MAT13 ) + FLD_S( SRC3 ) /* F2 F1 F0 F6 F5 F4 */ + FMUL_S( MAT14 ) + + FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + MOV_L( SRC3, EBX ) + + FXCH( ST(2) ) /* F4 F5 F6 */ + FSTP_S( DST0 ) /* F5 F6 */ + FSTP_S( DST1 ) /* F6 */ + FSTP_S( DST2 ) /* */ + MOV_L( EBX, DST3 ) + +LLBL(x86_p4_3dr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p4_3dr_loop) ) + +LLBL(x86_p4_3dr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME(_mesa_x86_transform_points4_3d_no_rot) +HIDDEN(_mesa_x86_transform_points4_3d_no_rot) +GLNAME(_mesa_x86_transform_points4_3d_no_rot): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p4_3dnrr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p4_3dnrr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F5 F4 */ + FMUL_S( MAT5 ) + + FLD_S( SRC2 ) /* F6 F5 F4 */ + FMUL_S( MAT10 ) + + FLD_S( SRC3 ) /* F0 F6 F5 F4 */ + FMUL_S( MAT12 ) + FLD_S( SRC3 ) /* F1 F0 F6 F5 F4 */ + FMUL_S( MAT13 ) + FLD_S( SRC3 ) /* F2 F1 F0 F6 F5 F4 */ + FMUL_S( MAT14 ) + + FXCH( ST(2) ) /* F0 F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(5) ) /* F1 F2 F6 F5 F4 */ + FADDP( ST0, ST(3) ) /* F2 F6 F5 F4 */ + FADDP( ST0, ST(1) ) /* F6 F5 F4 */ + + MOV_L( SRC3, EBX ) + + FXCH( ST(2) ) /* F4 F5 F6 */ + FSTP_S( DST0 ) /* F5 F6 */ + FSTP_S( DST1 ) /* F6 */ + FSTP_S( DST2 ) /* */ + MOV_L( EBX, DST3 ) + +LLBL(x86_p4_3dnrr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p4_3dnrr_loop) ) + +LLBL(x86_p4_3dnrr_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points4_2d ) +HIDDEN(_mesa_x86_transform_points4_2d) +GLNAME( _mesa_x86_transform_points4_2d ): + +#define FRAME_OFFSET 16 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + PUSH_L( EBP ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p4_2dr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p4_2dr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + FLD_S( SRC0 ) /* F5 F4 */ + FMUL_S( MAT1 ) + + FLD_S( SRC1 ) /* F0 F5 F4 */ + FMUL_S( MAT4 ) + FLD_S( SRC1 ) /* F1 F0 F5 F4 */ + FMUL_S( MAT5 ) + + FXCH( ST(1) ) /* F0 F1 F5 F4 */ + FADDP( ST0, ST(3) ) /* F1 F5 F4 */ + FADDP( ST0, ST(1) ) /* F5 F4 */ + + FLD_S( SRC3 ) /* F0 F5 F4 */ + FMUL_S( MAT12 ) + FLD_S( SRC3 ) /* F1 F0 F5 F4 */ + FMUL_S( MAT13 ) + + FXCH( ST(1) ) /* F0 F1 F5 F4 */ + FADDP( ST0, ST(3) ) /* F1 F5 F4 */ + FADDP( ST0, ST(1) ) /* F5 F4 */ + + MOV_L( SRC2, EBX ) + MOV_L( SRC3, EBP ) + + FXCH( ST(1) ) /* F4 F5 */ + FSTP_S( DST0 ) /* F5 */ + FSTP_S( DST1 ) /* */ + MOV_L( EBX, DST2 ) + MOV_L( EBP, DST3 ) + +LLBL(x86_p4_2dr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p4_2dr_loop) ) + +LLBL(x86_p4_2dr_done): + + POP_L( EBP ) + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points4_2d_no_rot ) +HIDDEN(_mesa_x86_transform_points4_2d_no_rot) +GLNAME( _mesa_x86_transform_points4_2d_no_rot ): + +#define FRAME_OFFSET 16 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + PUSH_L( EBP ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p4_2dnrr_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + +ALIGNTEXT16 +LLBL(x86_p4_2dnrr_loop): + + FLD_S( SRC0 ) /* F4 */ + FMUL_S( MAT0 ) + + FLD_S( SRC1 ) /* F5 F4 */ + FMUL_S( MAT5 ) + + FLD_S( SRC3 ) /* F0 F5 F4 */ + FMUL_S( MAT12 ) + FLD_S( SRC3 ) /* F1 F0 F5 F4 */ + FMUL_S( MAT13 ) + + FXCH( ST(1) ) /* F0 F1 F5 F4 */ + FADDP( ST0, ST(3) ) /* F1 F5 F4 */ + FADDP( ST0, ST(1) ) /* F5 F4 */ + + MOV_L( SRC2, EBX ) + MOV_L( SRC3, EBP ) + + FXCH( ST(1) ) /* F4 F5 */ + FSTP_S( DST0 ) /* F5 */ + FSTP_S( DST1 ) /* */ + MOV_L( EBX, DST2 ) + MOV_L( EBP, DST3 ) + +LLBL(x86_p4_2dnrr_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p4_2dnrr_loop) ) + +LLBL(x86_p4_2dnrr_done): + + POP_L( EBP ) + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET +#undef FRAME_OFFSET + + + + +ALIGNTEXT16 +GLOBL GLNAME( _mesa_x86_transform_points4_identity ) +HIDDEN(_mesa_x86_transform_points4_identity) +GLNAME( _mesa_x86_transform_points4_identity ): + +#define FRAME_OFFSET 12 + PUSH_L( ESI ) + PUSH_L( EDI ) + PUSH_L( EBX ) + + MOV_L( ARG_SOURCE, ESI ) + MOV_L( ARG_DEST, EDI ) + + MOV_L( ARG_MATRIX, EDX ) + MOV_L( REGOFF(V4F_COUNT, ESI), ECX ) + + TEST_L( ECX, ECX ) + JZ( LLBL(x86_p4_ir_done) ) + + MOV_L( REGOFF(V4F_STRIDE, ESI), EAX ) + OR_L( CONST(VEC_SIZE_4), REGOFF(V4F_FLAGS, EDI) ) + + MOV_L( ECX, REGOFF(V4F_COUNT, EDI) ) + MOV_L( CONST(4), REGOFF(V4F_SIZE, EDI) ) + + SHL_L( CONST(4), ECX ) + MOV_L( REGOFF(V4F_START, ESI), ESI ) + + MOV_L( REGOFF(V4F_START, EDI), EDI ) + ADD_L( EDI, ECX ) + + CMP_L( ESI, EDI ) + JE( LLBL(x86_p4_ir_done) ) + +ALIGNTEXT16 +LLBL(x86_p4_ir_loop): + + MOV_L( SRC0, EBX ) + MOV_L( SRC1, EDX ) + + MOV_L( EBX, DST0 ) + MOV_L( EDX, DST1 ) + + MOV_L( SRC2, EBX ) + MOV_L( SRC3, EDX ) + + MOV_L( EBX, DST2 ) + MOV_L( EDX, DST3 ) + +LLBL(x86_p4_ir_skip): + + ADD_L( CONST(16), EDI ) + ADD_L( EAX, ESI ) + CMP_L( ECX, EDI ) + JNE( LLBL(x86_p4_ir_loop) ) + +LLBL(x86_p4_ir_done): + + POP_L( EBX ) + POP_L( EDI ) + POP_L( ESI ) + RET + +#if defined (__ELF__) && defined (__linux__) + .section .note.GNU-stack,"",%progbits +#endif diff --git a/mesalib/src/mesa/x86/xform_args.h b/mesalib/src/mesa/x86/xform_args.h index 772ab9d1f..b773f5198 100644 --- a/mesalib/src/mesa/x86/xform_args.h +++ b/mesalib/src/mesa/x86/xform_args.h @@ -1,51 +1,51 @@ - -/* - * Mesa 3-D graphics library - * Version: 3.5 - * - * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included - * in all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS - * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN - * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN - * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. - */ - -/* - * Transform function interface for assembly code. Simply define - * FRAME_OFFSET to the number of bytes pushed onto the stack before - * using the ARG_* argument macros. - * - * Gareth Hughes - */ - -#ifndef __XFORM_ARGS_H__ -#define __XFORM_ARGS_H__ - -/* Offsets for transform_func arguments - * - * typedef void (*transform_func)( GLvector4f *to_vec, - * const GLfloat m[16], - * const GLvector4f *from_vec ); - */ -#define OFFSET_DEST 4 -#define OFFSET_MATRIX 8 -#define OFFSET_SOURCE 12 - -#define ARG_DEST REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP) -#define ARG_MATRIX REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP) -#define ARG_SOURCE REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP) - -#endif + +/* + * Mesa 3-D graphics library + * Version: 3.5 + * + * Copyright (C) 1999-2001 Brian Paul All Rights Reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +/* + * Transform function interface for assembly code. Simply define + * FRAME_OFFSET to the number of bytes pushed onto the stack before + * using the ARG_* argument macros. + * + * Gareth Hughes + */ + +#ifndef __XFORM_ARGS_H__ +#define __XFORM_ARGS_H__ + +/* Offsets for transform_func arguments + * + * typedef void (*transform_func)( GLvector4f *to_vec, + * const GLfloat m[16], + * const GLvector4f *from_vec ); + */ +#define OFFSET_DEST 4 +#define OFFSET_MATRIX 8 +#define OFFSET_SOURCE 12 + +#define ARG_DEST REGOFF(FRAME_OFFSET+OFFSET_DEST, ESP) +#define ARG_MATRIX REGOFF(FRAME_OFFSET+OFFSET_MATRIX, ESP) +#define ARG_SOURCE REGOFF(FRAME_OFFSET+OFFSET_SOURCE, ESP) + +#endif -- cgit v1.2.3