aboutsummaryrefslogtreecommitdiff
path: root/xorg-server/hw/dmx/glxProxy/render2swap.c
diff options
context:
space:
mode:
Diffstat (limited to 'xorg-server/hw/dmx/glxProxy/render2swap.c')
-rw-r--r--xorg-server/hw/dmx/glxProxy/render2swap.c572
1 files changed, 286 insertions, 286 deletions
diff --git a/xorg-server/hw/dmx/glxProxy/render2swap.c b/xorg-server/hw/dmx/glxProxy/render2swap.c
index 81bb501ea..70ed5e802 100644
--- a/xorg-server/hw/dmx/glxProxy/render2swap.c
+++ b/xorg-server/hw/dmx/glxProxy/render2swap.c
@@ -1,286 +1,286 @@
-/*
- * SGI FREE SOFTWARE LICENSE B (Version 2.0, Sept. 18, 2008)
- * Copyright (C) 1991-2000 Silicon 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
- * 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 including the dates of first publication and
- * either this permission notice or a reference to
- * http://oss.sgi.com/projects/FreeB/
- * shall be included in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS 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
- * SILICON GRAPHICS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
- * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF
- * OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- *
- * Except as contained in this notice, the name of Silicon Graphics, Inc.
- * shall not be used in advertising or otherwise to promote the sale, use or
- * other dealings in this Software without prior written authorization from
- * Silicon Graphics, Inc.
- */
-
-#include "glxserver.h"
-#include "unpack.h"
-#include "g_disptab.h"
-
-GLint __glEvalComputeK(GLenum target)
-{
- switch (target) {
- case GL_MAP1_VERTEX_4:
- case GL_MAP1_COLOR_4:
- case GL_MAP1_TEXTURE_COORD_4:
- case GL_MAP2_VERTEX_4:
- case GL_MAP2_COLOR_4:
- case GL_MAP2_TEXTURE_COORD_4:
- return 4;
- case GL_MAP1_VERTEX_3:
- case GL_MAP1_TEXTURE_COORD_3:
- case GL_MAP1_NORMAL:
- case GL_MAP2_VERTEX_3:
- case GL_MAP2_TEXTURE_COORD_3:
- case GL_MAP2_NORMAL:
- return 3;
- case GL_MAP1_TEXTURE_COORD_2:
- case GL_MAP2_TEXTURE_COORD_2:
- return 2;
- case GL_MAP1_TEXTURE_COORD_1:
- case GL_MAP2_TEXTURE_COORD_1:
- case GL_MAP1_INDEX:
- case GL_MAP2_INDEX:
- return 1;
- default:
- return 0;
- }
-}
-
-void __glXDispSwap_Map1f(GLbyte *pc)
-{
- GLint order, k;
- GLfloat u1, u2, *points;
- GLenum target;
- GLint compsize;
- __GLX_DECLARE_SWAP_VARIABLES;
- __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
-
- __GLX_SWAP_INT(pc + 0);
- __GLX_SWAP_INT(pc + 12);
- __GLX_SWAP_FLOAT(pc + 4);
- __GLX_SWAP_FLOAT(pc + 8);
-
- target = *(GLenum *)(pc + 0);
- order = *(GLint *)(pc + 12);
- u1 = *(GLfloat *)(pc + 4);
- u2 = *(GLfloat *)(pc + 8);
- points = (GLfloat *)(pc + 16);
- k = __glEvalComputeK(target);
-
- if (order <= 0 || k < 0) {
- /* Erroneous command. */
- compsize = 0;
- } else {
- compsize = order * k;
- }
- __GLX_SWAP_FLOAT_ARRAY(points, compsize);
-
-}
-
-void __glXDispSwap_Map2f(GLbyte *pc)
-{
- GLint uorder, vorder, ustride, vstride, k;
- GLfloat u1, u2, v1, v2, *points;
- GLenum target;
- GLint compsize;
- __GLX_DECLARE_SWAP_VARIABLES;
- __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
-
- __GLX_SWAP_INT(pc + 0);
- __GLX_SWAP_INT(pc + 12);
- __GLX_SWAP_INT(pc + 24);
- __GLX_SWAP_FLOAT(pc + 4);
- __GLX_SWAP_FLOAT(pc + 8);
- __GLX_SWAP_FLOAT(pc + 16);
- __GLX_SWAP_FLOAT(pc + 20);
-
- target = *(GLenum *)(pc + 0);
- uorder = *(GLint *)(pc + 12);
- vorder = *(GLint *)(pc + 24);
- u1 = *(GLfloat *)(pc + 4);
- u2 = *(GLfloat *)(pc + 8);
- v1 = *(GLfloat *)(pc + 16);
- v2 = *(GLfloat *)(pc + 20);
- points = (GLfloat *)(pc + 28);
-
- k = __glEvalComputeK(target);
- ustride = vorder * k;
- vstride = k;
-
- if (vorder <= 0 || uorder <= 0 || k < 0) {
- /* Erroneous command. */
- compsize = 0;
- } else {
- compsize = uorder * vorder * k;
- }
- __GLX_SWAP_FLOAT_ARRAY(points, compsize);
-
-}
-
-void __glXDispSwap_Map1d(GLbyte *pc)
-{
- GLint order, k, compsize;
- GLenum target;
- GLdouble u1, u2, *points;
- __GLX_DECLARE_SWAP_VARIABLES;
- __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
-
- __GLX_SWAP_DOUBLE(pc + 0);
- __GLX_SWAP_DOUBLE(pc + 8);
- __GLX_SWAP_INT(pc + 16);
- __GLX_SWAP_INT(pc + 20);
-
- target = *(GLenum*) (pc + 16);
- order = *(GLint*) (pc + 20);
- k = __glEvalComputeK(target);
- if (order <= 0 || k < 0) {
- /* Erroneous command. */
- compsize = 0;
- } else {
- compsize = order * k;
- }
- __GLX_GET_DOUBLE(u1,pc);
- __GLX_GET_DOUBLE(u2,pc+8);
- __GLX_SWAP_DOUBLE_ARRAY(pc+24, compsize);
- pc += 24;
-
-#ifdef __GLX_ALIGN64
- if (((unsigned long)pc) & 7) {
- /*
- ** Copy the doubles up 4 bytes, trashing the command but aligning
- ** the data in the process
- */
- __GLX_MEM_COPY(pc-4, pc, compsize*8);
- points = (GLdouble*) (pc - 4);
- } else {
- points = (GLdouble*) pc;
- }
-#else
- points = (GLdouble*) pc;
-#endif
-}
-
-void __glXDispSwap_Map2d(GLbyte *pc)
-{
- GLdouble u1, u2, v1, v2, *points;
- GLint uorder, vorder, ustride, vstride, k, compsize;
- GLenum target;
- __GLX_DECLARE_SWAP_VARIABLES;
- __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
-
- __GLX_SWAP_DOUBLE(pc + 0);
- __GLX_SWAP_DOUBLE(pc + 8);
- __GLX_SWAP_DOUBLE(pc + 16);
- __GLX_SWAP_DOUBLE(pc + 24);
- __GLX_SWAP_INT(pc + 32);
- __GLX_SWAP_INT(pc + 36);
- __GLX_SWAP_INT(pc + 40);
-
- target = *(GLenum *)(pc + 32);
- uorder = *(GLint *)(pc + 36);
- vorder = *(GLint *)(pc + 40);
- k = __glEvalComputeK(target);
- if (vorder <= 0 || uorder <= 0 || k < 0) {
- /* Erroneous command. */
- compsize = 0;
- } else {
- compsize = uorder * vorder * k;
- }
- __GLX_GET_DOUBLE(u1,pc);
- __GLX_GET_DOUBLE(u2,pc+8);
- __GLX_GET_DOUBLE(v1,pc+16);
- __GLX_GET_DOUBLE(v2,pc+24);
- __GLX_SWAP_DOUBLE_ARRAY(pc+44, compsize);
- pc += 44;
- ustride = vorder * k;
- vstride = k;
-
-#ifdef __GLX_ALIGN64
- if (((unsigned long)pc) & 7) {
- /*
- ** Copy the doubles up 4 bytes, trashing the command but aligning
- ** the data in the process
- */
- __GLX_MEM_COPY(pc-4, pc, compsize*8);
- points = (GLdouble*) (pc - 4);
- } else {
- points = (GLdouble*) pc;
- }
-#else
- points = (GLdouble*) pc;
-#endif
-}
-
-void __glXDispSwap_CallLists(GLbyte *pc)
-{
- GLenum type;
- GLsizei n;
- __GLX_DECLARE_SWAP_VARIABLES;
- __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
-
- __GLX_SWAP_INT(pc + 4);
- __GLX_SWAP_INT(pc + 0);
- type = *(GLenum *)(pc + 4);
- n = *(GLsizei *)(pc + 0);
-
- switch (type) {
- case GL_BYTE:
- case GL_UNSIGNED_BYTE:
- case GL_2_BYTES:
- case GL_3_BYTES:
- case GL_4_BYTES:
- break;
- case GL_SHORT:
- case GL_UNSIGNED_SHORT:
- __GLX_SWAP_SHORT_ARRAY(pc+8, n);
- break;
- case GL_INT:
- case GL_UNSIGNED_INT:
- __GLX_SWAP_INT_ARRAY(pc+8, n);
- break;
- case GL_FLOAT:
- __GLX_SWAP_FLOAT_ARRAY(pc+8, n);
- break;
- }
-
-}
-
-void __glXDispSwap_DrawArrays(GLbyte *pc)
-{
- __GLXdispatchDrawArraysHeader *hdr = (__GLXdispatchDrawArraysHeader *)pc;
- __GLXdispatchDrawArraysComponentHeader *compHeader;
- int i;
- __GLX_DECLARE_SWAP_VARIABLES;
-
- __GLX_SWAP_INT(&hdr->numVertexes);
- __GLX_SWAP_INT(&hdr->numComponents);
- __GLX_SWAP_INT(&hdr->primType);
-
- pc += sizeof(__GLXdispatchDrawArraysHeader);
- compHeader = (__GLXdispatchDrawArraysComponentHeader *) pc;
-
- /* compute stride (same for all component arrays) */
- for (i=0; i<hdr->numComponents; i++) {
- __GLX_SWAP_INT(&compHeader[i].datatype);
- __GLX_SWAP_INT(&compHeader[i].numVals);
- __GLX_SWAP_INT(&compHeader[i].component);
-
- }
-
-}
+/*
+ * SGI FREE SOFTWARE LICENSE B (Version 2.0, Sept. 18, 2008)
+ * Copyright (C) 1991-2000 Silicon 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
+ * 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 including the dates of first publication and
+ * either this permission notice or a reference to
+ * http://oss.sgi.com/projects/FreeB/
+ * shall be included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS 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
+ * SILICON GRAPHICS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
+ * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF
+ * OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ *
+ * Except as contained in this notice, the name of Silicon Graphics, Inc.
+ * shall not be used in advertising or otherwise to promote the sale, use or
+ * other dealings in this Software without prior written authorization from
+ * Silicon Graphics, Inc.
+ */
+
+#include "glxserver.h"
+#include "unpack.h"
+#include "g_disptab.h"
+
+GLint __glEvalComputeK(GLenum target)
+{
+ switch (target) {
+ case GL_MAP1_VERTEX_4:
+ case GL_MAP1_COLOR_4:
+ case GL_MAP1_TEXTURE_COORD_4:
+ case GL_MAP2_VERTEX_4:
+ case GL_MAP2_COLOR_4:
+ case GL_MAP2_TEXTURE_COORD_4:
+ return 4;
+ case GL_MAP1_VERTEX_3:
+ case GL_MAP1_TEXTURE_COORD_3:
+ case GL_MAP1_NORMAL:
+ case GL_MAP2_VERTEX_3:
+ case GL_MAP2_TEXTURE_COORD_3:
+ case GL_MAP2_NORMAL:
+ return 3;
+ case GL_MAP1_TEXTURE_COORD_2:
+ case GL_MAP2_TEXTURE_COORD_2:
+ return 2;
+ case GL_MAP1_TEXTURE_COORD_1:
+ case GL_MAP2_TEXTURE_COORD_1:
+ case GL_MAP1_INDEX:
+ case GL_MAP2_INDEX:
+ return 1;
+ default:
+ return 0;
+ }
+}
+
+void __glXDispSwap_Map1f(GLbyte *pc)
+{
+ GLint order, k;
+ GLfloat u1, u2, *points;
+ GLenum target;
+ GLint compsize;
+ __GLX_DECLARE_SWAP_VARIABLES;
+ __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
+
+ __GLX_SWAP_INT(pc + 0);
+ __GLX_SWAP_INT(pc + 12);
+ __GLX_SWAP_FLOAT(pc + 4);
+ __GLX_SWAP_FLOAT(pc + 8);
+
+ target = *(GLenum *)(pc + 0);
+ order = *(GLint *)(pc + 12);
+ u1 = *(GLfloat *)(pc + 4);
+ u2 = *(GLfloat *)(pc + 8);
+ points = (GLfloat *)(pc + 16);
+ k = __glEvalComputeK(target);
+
+ if (order <= 0 || k < 0) {
+ /* Erroneous command. */
+ compsize = 0;
+ } else {
+ compsize = order * k;
+ }
+ __GLX_SWAP_FLOAT_ARRAY(points, compsize);
+
+}
+
+void __glXDispSwap_Map2f(GLbyte *pc)
+{
+ GLint uorder, vorder, ustride, vstride, k;
+ GLfloat u1, u2, v1, v2, *points;
+ GLenum target;
+ GLint compsize;
+ __GLX_DECLARE_SWAP_VARIABLES;
+ __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
+
+ __GLX_SWAP_INT(pc + 0);
+ __GLX_SWAP_INT(pc + 12);
+ __GLX_SWAP_INT(pc + 24);
+ __GLX_SWAP_FLOAT(pc + 4);
+ __GLX_SWAP_FLOAT(pc + 8);
+ __GLX_SWAP_FLOAT(pc + 16);
+ __GLX_SWAP_FLOAT(pc + 20);
+
+ target = *(GLenum *)(pc + 0);
+ uorder = *(GLint *)(pc + 12);
+ vorder = *(GLint *)(pc + 24);
+ u1 = *(GLfloat *)(pc + 4);
+ u2 = *(GLfloat *)(pc + 8);
+ v1 = *(GLfloat *)(pc + 16);
+ v2 = *(GLfloat *)(pc + 20);
+ points = (GLfloat *)(pc + 28);
+
+ k = __glEvalComputeK(target);
+ ustride = vorder * k;
+ vstride = k;
+
+ if (vorder <= 0 || uorder <= 0 || k < 0) {
+ /* Erroneous command. */
+ compsize = 0;
+ } else {
+ compsize = uorder * vorder * k;
+ }
+ __GLX_SWAP_FLOAT_ARRAY(points, compsize);
+
+}
+
+void __glXDispSwap_Map1d(GLbyte *pc)
+{
+ GLint order, k, compsize;
+ GLenum target;
+ GLdouble u1, u2, *points;
+ __GLX_DECLARE_SWAP_VARIABLES;
+ __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
+
+ __GLX_SWAP_DOUBLE(pc + 0);
+ __GLX_SWAP_DOUBLE(pc + 8);
+ __GLX_SWAP_INT(pc + 16);
+ __GLX_SWAP_INT(pc + 20);
+
+ target = *(GLenum*) (pc + 16);
+ order = *(GLint*) (pc + 20);
+ k = __glEvalComputeK(target);
+ if (order <= 0 || k < 0) {
+ /* Erroneous command. */
+ compsize = 0;
+ } else {
+ compsize = order * k;
+ }
+ __GLX_GET_DOUBLE(u1,pc);
+ __GLX_GET_DOUBLE(u2,pc+8);
+ __GLX_SWAP_DOUBLE_ARRAY(pc+24, compsize);
+ pc += 24;
+
+#ifdef __GLX_ALIGN64
+ if (((unsigned long)pc) & 7) {
+ /*
+ ** Copy the doubles up 4 bytes, trashing the command but aligning
+ ** the data in the process
+ */
+ __GLX_MEM_COPY(pc-4, pc, compsize*8);
+ points = (GLdouble*) (pc - 4);
+ } else {
+ points = (GLdouble*) pc;
+ }
+#else
+ points = (GLdouble*) pc;
+#endif
+}
+
+void __glXDispSwap_Map2d(GLbyte *pc)
+{
+ GLdouble u1, u2, v1, v2, *points;
+ GLint uorder, vorder, ustride, vstride, k, compsize;
+ GLenum target;
+ __GLX_DECLARE_SWAP_VARIABLES;
+ __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
+
+ __GLX_SWAP_DOUBLE(pc + 0);
+ __GLX_SWAP_DOUBLE(pc + 8);
+ __GLX_SWAP_DOUBLE(pc + 16);
+ __GLX_SWAP_DOUBLE(pc + 24);
+ __GLX_SWAP_INT(pc + 32);
+ __GLX_SWAP_INT(pc + 36);
+ __GLX_SWAP_INT(pc + 40);
+
+ target = *(GLenum *)(pc + 32);
+ uorder = *(GLint *)(pc + 36);
+ vorder = *(GLint *)(pc + 40);
+ k = __glEvalComputeK(target);
+ if (vorder <= 0 || uorder <= 0 || k < 0) {
+ /* Erroneous command. */
+ compsize = 0;
+ } else {
+ compsize = uorder * vorder * k;
+ }
+ __GLX_GET_DOUBLE(u1,pc);
+ __GLX_GET_DOUBLE(u2,pc+8);
+ __GLX_GET_DOUBLE(v1,pc+16);
+ __GLX_GET_DOUBLE(v2,pc+24);
+ __GLX_SWAP_DOUBLE_ARRAY(pc+44, compsize);
+ pc += 44;
+ ustride = vorder * k;
+ vstride = k;
+
+#ifdef __GLX_ALIGN64
+ if (((unsigned long)pc) & 7) {
+ /*
+ ** Copy the doubles up 4 bytes, trashing the command but aligning
+ ** the data in the process
+ */
+ __GLX_MEM_COPY(pc-4, pc, compsize*8);
+ points = (GLdouble*) (pc - 4);
+ } else {
+ points = (GLdouble*) pc;
+ }
+#else
+ points = (GLdouble*) pc;
+#endif
+}
+
+void __glXDispSwap_CallLists(GLbyte *pc)
+{
+ GLenum type;
+ GLsizei n;
+ __GLX_DECLARE_SWAP_VARIABLES;
+ __GLX_DECLARE_SWAP_ARRAY_VARIABLES;
+
+ __GLX_SWAP_INT(pc + 4);
+ __GLX_SWAP_INT(pc + 0);
+ type = *(GLenum *)(pc + 4);
+ n = *(GLsizei *)(pc + 0);
+
+ switch (type) {
+ case GL_BYTE:
+ case GL_UNSIGNED_BYTE:
+ case GL_2_BYTES:
+ case GL_3_BYTES:
+ case GL_4_BYTES:
+ break;
+ case GL_SHORT:
+ case GL_UNSIGNED_SHORT:
+ __GLX_SWAP_SHORT_ARRAY(pc+8, n);
+ break;
+ case GL_INT:
+ case GL_UNSIGNED_INT:
+ __GLX_SWAP_INT_ARRAY(pc+8, n);
+ break;
+ case GL_FLOAT:
+ __GLX_SWAP_FLOAT_ARRAY(pc+8, n);
+ break;
+ }
+
+}
+
+void __glXDispSwap_DrawArrays(GLbyte *pc)
+{
+ __GLXdispatchDrawArraysHeader *hdr = (__GLXdispatchDrawArraysHeader *)pc;
+ __GLXdispatchDrawArraysComponentHeader *compHeader;
+ int i;
+ __GLX_DECLARE_SWAP_VARIABLES;
+
+ __GLX_SWAP_INT(&hdr->numVertexes);
+ __GLX_SWAP_INT(&hdr->numComponents);
+ __GLX_SWAP_INT(&hdr->primType);
+
+ pc += sizeof(__GLXdispatchDrawArraysHeader);
+ compHeader = (__GLXdispatchDrawArraysComponentHeader *) pc;
+
+ /* compute stride (same for all component arrays) */
+ for (i=0; i<hdr->numComponents; i++) {
+ __GLX_SWAP_INT(&compHeader[i].datatype);
+ __GLX_SWAP_INT(&compHeader[i].numVals);
+ __GLX_SWAP_INT(&compHeader[i].component);
+
+ }
+
+}