mesa update 11 july 2011

40 files changed, 10146 insertions, 8832 deletions

diff --git a/mesalib/.gitattributes b/mesalib/.gitattributes
new file mode 100644
index 000000000..23c6280f6
--- /dev/null
+++ b/mesalib/.gitattributes
@@ -0,0 +1,4 @@
+*.dsp -crlf
+*.dsw -crlf
+*.sln -crlf
+*.vcproj -crlf
diff --git a/mesalib/configure.ac b/mesalib/configure.ac
index f19f6478b..dde13c9b5 100644
--- a/mesalib/configure.ac
+++ b/mesalib/configure.ac
@@ -22,8 +22,8 @@ LIBDRM_REQUIRED=2.4.24
 LIBDRM_RADEON_REQUIRED=2.4.24
 LIBDRM_INTEL_REQUIRED=2.4.24
 LIBDRM_NOUVEAU_REQUIRED=0.6
-DRI2PROTO_REQUIRED=2.1
-GLPROTO_REQUIRED=1.4.11
+DRI2PROTO_REQUIRED=2.6
+GLPROTO_REQUIRED=1.4.14
 LIBDRM_XORG_REQUIRED=2.4.24
 LIBKMS_XORG_REQUIRED=1.0.0
 
diff --git a/mesalib/docs/GL3.txt b/mesalib/docs/GL3.txt
index 49b48472a..135bc4bab 100644
--- a/mesalib/docs/GL3.txt
+++ b/mesalib/docs/GL3.txt
@@ -19,7 +19,7 @@ Clamping controls (GL_ARB_color_buffer_float)         DONE
 Float textures, renderbuffers (GL_ARB_texture_float)  DONE (gallium r300)
 GL_EXT_packed_float                                   DONE (gallium r600)
 GL_EXT_texture_shared_exponent                        DONE (gallium, swrast)
-Float depth buffers (GL_ARB_depth_buffer_float)       not started
+Float depth buffers (GL_ARB_depth_buffer_float)       DONE
 Framebuffer objects (GL_EXT_framebuffer_object)       DONE
 Half-float                                            DONE
 Multisample blit                                      DONE
diff --git a/mesalib/src/gallium/auxiliary/util/u_pack_color.h b/mesalib/src/gallium/auxiliary/util/u_pack_color.h
index c51cc7bcb..9391f1b80 100644
--- a/mesalib/src/gallium/auxiliary/util/u_pack_color.h
+++ b/mesalib/src/gallium/auxiliary/util/u_pack_color.h
@@ -1,584 +1,648 @@
-/**************************************************************************

- *

- * 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.

- *

- **************************************************************************/

-

-/**

- * @file

- * Functions to produce packed colors/Z from floats.

- */

-

-

-#ifndef U_PACK_COLOR_H

-#define U_PACK_COLOR_H

-

-

-#include "pipe/p_compiler.h"

-#include "pipe/p_format.h"

-#include "util/u_debug.h"

-#include "util/u_format.h"

-#include "util/u_math.h"

-

-

-/**

- * Helper union for packing pixel values.

- * Will often contain values in formats which are too complex to be described

- * in simple terms, hence might just effectively contain a number of bytes.

- * Must be big enough to hold data for all formats (currently 256 bits).

- */

-union util_color {

-   ubyte ub;

-   ushort us;

-   uint ui;

-   float f[4];

-   double d[4];

-};

-

-/**

- * Pack ubyte R,G,B,A into dest pixel.

- */

-static INLINE void

-util_pack_color_ub(ubyte r, ubyte g, ubyte b, ubyte a,

-                   enum pipe_format format, union util_color *uc)

-{

-   switch (format) {

-   case PIPE_FORMAT_A8B8G8R8_UNORM:

-      {

-         uc->ui = (r << 24) | (g << 16) | (b << 8) | a;

-      }

-      return;

-   case PIPE_FORMAT_X8B8G8R8_UNORM:

-      {

-         uc->ui = (r << 24) | (g << 16) | (b << 8) | 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B8G8R8A8_UNORM:

-      {

-         uc->ui = (a << 24) | (r << 16) | (g << 8) | b;

-      }

-      return;

-   case PIPE_FORMAT_B8G8R8X8_UNORM:

-      {

-         uc->ui = (0xff << 24) | (r << 16) | (g << 8) | b;

-      }

-      return;

-   case PIPE_FORMAT_A8R8G8B8_UNORM:

-      {

-         uc->ui = (b << 24) | (g << 16) | (r << 8) | a;

-      }

-      return;

-   case PIPE_FORMAT_X8R8G8B8_UNORM:

-      {

-         uc->ui = (b << 24) | (g << 16) | (r << 8) | 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B5G6R5_UNORM:

-      {

-         uc->us = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);

-      }

-      return;

-   case PIPE_FORMAT_B5G5R5X1_UNORM:

-      {

-         uc->us = ((0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);

-      }

-      return;

-   case PIPE_FORMAT_B5G5R5A1_UNORM:

-      {

-         uc->us = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);

-      }

-      return;

-   case PIPE_FORMAT_B4G4R4A4_UNORM:

-      {

-         uc->us = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);

-      }

-      return;

-   case PIPE_FORMAT_A8_UNORM:

-      {

-         uc->ub = a;

-      }

-      return;

-   case PIPE_FORMAT_L8_UNORM:

-   case PIPE_FORMAT_I8_UNORM:

-      {

-         uc->ub = a;

-      }

-      return;

-   case PIPE_FORMAT_R32G32B32A32_FLOAT:

-      {

-         uc->f[0] = (float)r / 255.0f;

-         uc->f[1] = (float)g / 255.0f;

-         uc->f[2] = (float)b / 255.0f;

-         uc->f[3] = (float)a / 255.0f;

-      }

-      return;

-   case PIPE_FORMAT_R32G32B32_FLOAT:

-      {

-         uc->f[0] = (float)r / 255.0f;

-         uc->f[1] = (float)g / 255.0f;

-         uc->f[2] = (float)b / 255.0f;

-      }

-      return;

-

-   /* Handle other cases with a generic function.

-    */

-   default:

-      {

-         ubyte src[4];

-

-         src[0] = r;

-         src[1] = g;

-         src[2] = b;

-         src[3] = a;

-         util_format_write_4ub(format, src, 0, uc, 0, 0, 0, 1, 1);

-      }

-   }

-}

- 

-

-/**

- * Unpack RGBA from a packed pixel, returning values as ubytes in [0,255].

- */

-static INLINE void

-util_unpack_color_ub(enum pipe_format format, union util_color *uc,

-                     ubyte *r, ubyte *g, ubyte *b, ubyte *a)

-{

-   switch (format) {

-   case PIPE_FORMAT_A8B8G8R8_UNORM:

-      {

-         uint p = uc->ui;

-         *r = (ubyte) ((p >> 24) & 0xff);

-         *g = (ubyte) ((p >> 16) & 0xff);

-         *b = (ubyte) ((p >>  8) & 0xff);

-         *a = (ubyte) ((p >>  0) & 0xff);

-      }

-      return;

-   case PIPE_FORMAT_X8B8G8R8_UNORM:

-      {

-         uint p = uc->ui;

-         *r = (ubyte) ((p >> 24) & 0xff);

-         *g = (ubyte) ((p >> 16) & 0xff);

-         *b = (ubyte) ((p >>  8) & 0xff);

-         *a = (ubyte) 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B8G8R8A8_UNORM:

-      {

-         uint p = uc->ui;

-         *r = (ubyte) ((p >> 16) & 0xff);

-         *g = (ubyte) ((p >>  8) & 0xff);

-         *b = (ubyte) ((p >>  0) & 0xff);

-         *a = (ubyte) ((p >> 24) & 0xff);

-      }

-      return;

-   case PIPE_FORMAT_B8G8R8X8_UNORM:

-      {

-         uint p = uc->ui;

-         *r = (ubyte) ((p >> 16) & 0xff);

-         *g = (ubyte) ((p >>  8) & 0xff);

-         *b = (ubyte) ((p >>  0) & 0xff);

-         *a = (ubyte) 0xff;

-      }

-      return;

-   case PIPE_FORMAT_A8R8G8B8_UNORM:

-      {

-         uint p = uc->ui;

-         *r = (ubyte) ((p >>  8) & 0xff);

-         *g = (ubyte) ((p >> 16) & 0xff);

-         *b = (ubyte) ((p >> 24) & 0xff);

-         *a = (ubyte) ((p >>  0) & 0xff);

-      }

-      return;

-   case PIPE_FORMAT_X8R8G8B8_UNORM:

-      {

-         uint p = uc->ui;

-         *r = (ubyte) ((p >>  8) & 0xff);

-         *g = (ubyte) ((p >> 16) & 0xff);

-         *b = (ubyte) ((p >> 24) & 0xff);

-         *a = (ubyte) 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B5G6R5_UNORM:

-      {

-         ushort p = uc->us;

-         *r = (ubyte) (((p >> 8) & 0xf8) | ((p >> 13) & 0x7));

-         *g = (ubyte) (((p >> 3) & 0xfc) | ((p >>  9) & 0x3));

-         *b = (ubyte) (((p << 3) & 0xf8) | ((p >>  2) & 0x7));

-         *a = (ubyte) 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B5G5R5X1_UNORM:

-      {

-         ushort p = uc->us;

-         *r = (ubyte) (((p >>  7) & 0xf8) | ((p >> 12) & 0x7));

-         *g = (ubyte) (((p >>  2) & 0xf8) | ((p >>  7) & 0x7));

-         *b = (ubyte) (((p <<  3) & 0xf8) | ((p >>  2) & 0x7));

-         *a = (ubyte) 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B5G5R5A1_UNORM:

-      {

-         ushort p = uc->us;

-         *r = (ubyte) (((p >>  7) & 0xf8) | ((p >> 12) & 0x7));

-         *g = (ubyte) (((p >>  2) & 0xf8) | ((p >>  7) & 0x7));

-         *b = (ubyte) (((p <<  3) & 0xf8) | ((p >>  2) & 0x7));

-         *a = (ubyte) (0xff * (p >> 15));

-      }

-      return;

-   case PIPE_FORMAT_B4G4R4A4_UNORM:

-      {

-         ushort p = uc->us;

-         *r = (ubyte) (((p >> 4) & 0xf0) | ((p >>  8) & 0xf));

-         *g = (ubyte) (((p >> 0) & 0xf0) | ((p >>  4) & 0xf));

-         *b = (ubyte) (((p << 4) & 0xf0) | ((p >>  0) & 0xf));

-         *a = (ubyte) (((p >> 8) & 0xf0) | ((p >> 12) & 0xf));

-      }

-      return;

-   case PIPE_FORMAT_A8_UNORM:

-      {

-         ubyte p = uc->ub;

-         *r = *g = *b = (ubyte) 0xff;

-         *a = p;

-      }

-      return;

-   case PIPE_FORMAT_L8_UNORM:

-      {

-         ubyte p = uc->ub;

-         *r = *g = *b = p;

-         *a = (ubyte) 0xff;

-      }

-      return;

-   case PIPE_FORMAT_I8_UNORM:

-      {

-         ubyte p = uc->ub;

-         *r = *g = *b = *a = p;

-      }

-      return;

-   case PIPE_FORMAT_R32G32B32A32_FLOAT:

-      {

-         const float *p = &uc->f[0];

-         *r = float_to_ubyte(p[0]);

-         *g = float_to_ubyte(p[1]);

-         *b = float_to_ubyte(p[2]);

-         *a = float_to_ubyte(p[3]);

-      }

-      return;

-   case PIPE_FORMAT_R32G32B32_FLOAT:

-      {

-         const float *p = &uc->f[0];

-         *r = float_to_ubyte(p[0]);

-         *g = float_to_ubyte(p[1]);

-         *b = float_to_ubyte(p[2]);

-         *a = (ubyte) 0xff;

-      }

-      return;

-

-   case PIPE_FORMAT_R32G32_FLOAT:

-      {

-         const float *p = &uc->f[0];

-         *r = float_to_ubyte(p[0]);

-         *g = float_to_ubyte(p[1]);

-         *b = *a = (ubyte) 0xff;

-      }

-      return;

-

-   case PIPE_FORMAT_R32_FLOAT:

-      {

-         const float *p = &uc->f[0];

-         *r = float_to_ubyte(p[0]);

-         *g = *b = *a = (ubyte) 0xff;

-      }

-      return;

-

-   /* Handle other cases with a generic function.

-    */

-   default:

-      {

-         ubyte dst[4];

-

-         util_format_read_4ub(format, dst, 0, uc, 0, 0, 0, 1, 1);

-         *r = dst[0];

-         *g = dst[1];

-         *b = dst[2];

-         *a = dst[3];

-      }

-   }

-}

-

-

-/**

- * Note rgba outside [0,1] will be clamped for int pixel formats.

- */

-static INLINE void

-util_pack_color(const float rgba[4], enum pipe_format format, union util_color *uc)

-{

-   ubyte r = 0;

-   ubyte g = 0;

-   ubyte b = 0;

-   ubyte a = 0;

-

-   if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0) <= 8) {

-      /* format uses 8-bit components or less */

-      r = float_to_ubyte(rgba[0]);

-      g = float_to_ubyte(rgba[1]);

-      b = float_to_ubyte(rgba[2]);

-      a = float_to_ubyte(rgba[3]);

-   }

-

-   switch (format) {

-   case PIPE_FORMAT_A8B8G8R8_UNORM:

-      {

-         uc->ui = (r << 24) | (g << 16) | (b << 8) | a;

-      }

-      return;

-   case PIPE_FORMAT_X8B8G8R8_UNORM:

-      {

-         uc->ui = (r << 24) | (g << 16) | (b << 8) | 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B8G8R8A8_UNORM:

-      {

-         uc->ui = (a << 24) | (r << 16) | (g << 8) | b;

-      }

-      return;

-   case PIPE_FORMAT_B8G8R8X8_UNORM:

-      {

-         uc->ui = (0xff << 24) | (r << 16) | (g << 8) | b;

-      }

-      return;

-   case PIPE_FORMAT_A8R8G8B8_UNORM:

-      {

-         uc->ui = (b << 24) | (g << 16) | (r << 8) | a;

-      }

-      return;

-   case PIPE_FORMAT_X8R8G8B8_UNORM:

-      {

-         uc->ui = (b << 24) | (g << 16) | (r << 8) | 0xff;

-      }

-      return;

-   case PIPE_FORMAT_B5G6R5_UNORM:

-      {

-         uc->us = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);

-      }

-      return;

-   case PIPE_FORMAT_B5G5R5X1_UNORM:

-      {

-         uc->us = ((0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);

-      }

-      return;

-   case PIPE_FORMAT_B5G5R5A1_UNORM:

-      {

-         uc->us = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);

-      }

-      return;

-   case PIPE_FORMAT_B4G4R4A4_UNORM:

-      {

-         uc->us = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);

-      }

-      return;

-   case PIPE_FORMAT_A8_UNORM:

-      {

-         uc->ub = a;

-      }

-      return;

-   case PIPE_FORMAT_L8_UNORM:

-   case PIPE_FORMAT_I8_UNORM:

-      {

-         uc->ub = r;

-      }

-      return;

-   case PIPE_FORMAT_R32G32B32A32_FLOAT:

-      {

-         uc->f[0] = rgba[0];

-         uc->f[1] = rgba[1];

-         uc->f[2] = rgba[2];

-         uc->f[3] = rgba[3];

-      }

-      return;

-   case PIPE_FORMAT_R32G32B32_FLOAT:

-      {

-         uc->f[0] = rgba[0];

-         uc->f[1] = rgba[1];

-         uc->f[2] = rgba[2];

-      }

-      return;

-

-   /* Handle other cases with a generic function.

-    */

-   default:

-      util_format_write_4f(format, rgba, 0, uc, 0, 0, 0, 1, 1);

-   }

-}

- 

-/* Integer versions of util_pack_z and util_pack_z_stencil - useful for

- * constructing clear masks.

- */

-static INLINE uint32_t

-util_pack_mask_z(enum pipe_format format, uint32_t z)

-{

-   switch (format) {

-   case PIPE_FORMAT_Z16_UNORM:

-      return z & 0xffff;

-   case PIPE_FORMAT_Z32_UNORM:

-   case PIPE_FORMAT_Z32_FLOAT:

-      return z;

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-   case PIPE_FORMAT_Z24X8_UNORM:

-      return z & 0xffffff;

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-   case PIPE_FORMAT_X8Z24_UNORM:

-      return (z & 0xffffff) << 8;

-   case PIPE_FORMAT_S8_USCALED:

-      return 0;

-   default:

-      debug_print_format("gallium: unhandled format in util_pack_mask_z()", format);

-      assert(0);

-      return 0;

-   }

-}

-

-static INLINE uint32_t

-util_pack_mask_z_stencil(enum pipe_format format, uint32_t z, uint8_t s)

-{

-   uint32_t packed = util_pack_mask_z(format, z);

-

-   switch (format) {

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-      packed |= (uint32_t)s << 24;

-      break;

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-      packed |= s;

-      break;

-   case PIPE_FORMAT_S8_USCALED:

-      packed |= s;

-      break;

-   default:

-      break;

-   }

-

-   return packed;

-}

-

-

-

-/**

- * Note: it's assumed that z is in [0,1]

- */

-static INLINE uint32_t

-util_pack_z(enum pipe_format format, double z)

-{

-   union fi fui;

-

-   if (z == 0.0)

-      return 0;

-

-   switch (format) {

-   case PIPE_FORMAT_Z16_UNORM:

-      if (z == 1.0)

-         return 0xffff;

-      return (uint32_t) (z * 0xffff);

-   case PIPE_FORMAT_Z32_UNORM:

-      /* special-case to avoid overflow */

-      if (z == 1.0)

-         return 0xffffffff;

-      return (uint32_t) (z * 0xffffffff);

-   case PIPE_FORMAT_Z32_FLOAT:

-      fui.f = (float)z;

-      return fui.ui;

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-   case PIPE_FORMAT_Z24X8_UNORM:

-      if (z == 1.0)

-         return 0xffffff;

-      return (uint32_t) (z * 0xffffff);

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-   case PIPE_FORMAT_X8Z24_UNORM:

-      if (z == 1.0)

-         return 0xffffff00;

-      return ((uint32_t) (z * 0xffffff)) << 8;

-   case PIPE_FORMAT_S8_USCALED:

-      /* this case can get it via util_pack_z_stencil() */

-      return 0;

-   default:

-      debug_print_format("gallium: unhandled format in util_pack_z()", format);

-      assert(0);

-      return 0;

-   }

-}

- 

-

-/**

- * Pack Z and/or stencil values into a 32-bit value described by format.

- * Note: it's assumed that z is in [0,1] and s in [0,255]

- */

-static INLINE uint32_t

-util_pack_z_stencil(enum pipe_format format, double z, uint8_t s)

-{

-   uint32_t packed = util_pack_z(format, z);

-

-   switch (format) {

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-      packed |= (uint32_t)s << 24;

-      break;

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-      packed |= s;

-      break;

-   case PIPE_FORMAT_S8_USCALED:

-      packed |= s;

-      break;

-   default:

-      break;

-   }

-

-   return packed;

-}

-

-

-/**

- * Pack 4 ubytes into a 4-byte word

- */

-static INLINE unsigned

-pack_ub4(ubyte b0, ubyte b1, ubyte b2, ubyte b3)

-{

-   return ((((unsigned int)b0) << 0) |

-	   (((unsigned int)b1) << 8) |

-	   (((unsigned int)b2) << 16) |

-	   (((unsigned int)b3) << 24));

-}

-

-

-/**

- * Pack/convert 4 floats into one 4-byte word.

- */

-static INLINE unsigned

-pack_ui32_float4(float a, float b, float c, float d)

-{

-   return pack_ub4( float_to_ubyte(a),

-		    float_to_ubyte(b),

-		    float_to_ubyte(c),

-		    float_to_ubyte(d) );

-}

-

-

-

-#endif /* U_PACK_COLOR_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.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * Functions to produce packed colors/Z from floats.
+ */
+
+
+#ifndef U_PACK_COLOR_H
+#define U_PACK_COLOR_H
+
+
+#include "pipe/p_compiler.h"
+#include "pipe/p_format.h"
+#include "util/u_debug.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
+
+
+/**
+ * Helper union for packing pixel values.
+ * Will often contain values in formats which are too complex to be described
+ * in simple terms, hence might just effectively contain a number of bytes.
+ * Must be big enough to hold data for all formats (currently 256 bits).
+ */
+union util_color {
+   ubyte ub;
+   ushort us;
+   uint ui;
+   float f[4];
+   double d[4];
+};
+
+/**
+ * Pack ubyte R,G,B,A into dest pixel.
+ */
+static INLINE void
+util_pack_color_ub(ubyte r, ubyte g, ubyte b, ubyte a,
+                   enum pipe_format format, union util_color *uc)
+{
+   switch (format) {
+   case PIPE_FORMAT_A8B8G8R8_UNORM:
+      {
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | a;
+      }
+      return;
+   case PIPE_FORMAT_X8B8G8R8_UNORM:
+      {
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B8G8R8A8_UNORM:
+      {
+         uc->ui = (a << 24) | (r << 16) | (g << 8) | b;
+      }
+      return;
+   case PIPE_FORMAT_B8G8R8X8_UNORM:
+      {
+         uc->ui = (0xff << 24) | (r << 16) | (g << 8) | b;
+      }
+      return;
+   case PIPE_FORMAT_A8R8G8B8_UNORM:
+      {
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | a;
+      }
+      return;
+   case PIPE_FORMAT_X8R8G8B8_UNORM:
+      {
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B5G6R5_UNORM:
+      {
+         uc->us = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);
+      }
+      return;
+   case PIPE_FORMAT_B5G5R5X1_UNORM:
+      {
+         uc->us = ((0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
+      }
+      return;
+   case PIPE_FORMAT_B5G5R5A1_UNORM:
+      {
+         uc->us = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
+      }
+      return;
+   case PIPE_FORMAT_B4G4R4A4_UNORM:
+      {
+         uc->us = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);
+      }
+      return;
+   case PIPE_FORMAT_A8_UNORM:
+      {
+         uc->ub = a;
+      }
+      return;
+   case PIPE_FORMAT_L8_UNORM:
+   case PIPE_FORMAT_I8_UNORM:
+      {
+         uc->ub = a;
+      }
+      return;
+   case PIPE_FORMAT_R32G32B32A32_FLOAT:
+      {
+         uc->f[0] = (float)r / 255.0f;
+         uc->f[1] = (float)g / 255.0f;
+         uc->f[2] = (float)b / 255.0f;
+         uc->f[3] = (float)a / 255.0f;
+      }
+      return;
+   case PIPE_FORMAT_R32G32B32_FLOAT:
+      {
+         uc->f[0] = (float)r / 255.0f;
+         uc->f[1] = (float)g / 255.0f;
+         uc->f[2] = (float)b / 255.0f;
+      }
+      return;
+
+   /* Handle other cases with a generic function.
+    */
+   default:
+      {
+         ubyte src[4];
+
+         src[0] = r;
+         src[1] = g;
+         src[2] = b;
+         src[3] = a;
+         util_format_write_4ub(format, src, 0, uc, 0, 0, 0, 1, 1);
+      }
+   }
+}
+ 
+
+/**
+ * Unpack RGBA from a packed pixel, returning values as ubytes in [0,255].
+ */
+static INLINE void
+util_unpack_color_ub(enum pipe_format format, union util_color *uc,
+                     ubyte *r, ubyte *g, ubyte *b, ubyte *a)
+{
+   switch (format) {
+   case PIPE_FORMAT_A8B8G8R8_UNORM:
+      {
+         uint p = uc->ui;
+         *r = (ubyte) ((p >> 24) & 0xff);
+         *g = (ubyte) ((p >> 16) & 0xff);
+         *b = (ubyte) ((p >>  8) & 0xff);
+         *a = (ubyte) ((p >>  0) & 0xff);
+      }
+      return;
+   case PIPE_FORMAT_X8B8G8R8_UNORM:
+      {
+         uint p = uc->ui;
+         *r = (ubyte) ((p >> 24) & 0xff);
+         *g = (ubyte) ((p >> 16) & 0xff);
+         *b = (ubyte) ((p >>  8) & 0xff);
+         *a = (ubyte) 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B8G8R8A8_UNORM:
+      {
+         uint p = uc->ui;
+         *r = (ubyte) ((p >> 16) & 0xff);
+         *g = (ubyte) ((p >>  8) & 0xff);
+         *b = (ubyte) ((p >>  0) & 0xff);
+         *a = (ubyte) ((p >> 24) & 0xff);
+      }
+      return;
+   case PIPE_FORMAT_B8G8R8X8_UNORM:
+      {
+         uint p = uc->ui;
+         *r = (ubyte) ((p >> 16) & 0xff);
+         *g = (ubyte) ((p >>  8) & 0xff);
+         *b = (ubyte) ((p >>  0) & 0xff);
+         *a = (ubyte) 0xff;
+      }
+      return;
+   case PIPE_FORMAT_A8R8G8B8_UNORM:
+      {
+         uint p = uc->ui;
+         *r = (ubyte) ((p >>  8) & 0xff);
+         *g = (ubyte) ((p >> 16) & 0xff);
+         *b = (ubyte) ((p >> 24) & 0xff);
+         *a = (ubyte) ((p >>  0) & 0xff);
+      }
+      return;
+   case PIPE_FORMAT_X8R8G8B8_UNORM:
+      {
+         uint p = uc->ui;
+         *r = (ubyte) ((p >>  8) & 0xff);
+         *g = (ubyte) ((p >> 16) & 0xff);
+         *b = (ubyte) ((p >> 24) & 0xff);
+         *a = (ubyte) 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B5G6R5_UNORM:
+      {
+         ushort p = uc->us;
+         *r = (ubyte) (((p >> 8) & 0xf8) | ((p >> 13) & 0x7));
+         *g = (ubyte) (((p >> 3) & 0xfc) | ((p >>  9) & 0x3));
+         *b = (ubyte) (((p << 3) & 0xf8) | ((p >>  2) & 0x7));
+         *a = (ubyte) 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B5G5R5X1_UNORM:
+      {
+         ushort p = uc->us;
+         *r = (ubyte) (((p >>  7) & 0xf8) | ((p >> 12) & 0x7));
+         *g = (ubyte) (((p >>  2) & 0xf8) | ((p >>  7) & 0x7));
+         *b = (ubyte) (((p <<  3) & 0xf8) | ((p >>  2) & 0x7));
+         *a = (ubyte) 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B5G5R5A1_UNORM:
+      {
+         ushort p = uc->us;
+         *r = (ubyte) (((p >>  7) & 0xf8) | ((p >> 12) & 0x7));
+         *g = (ubyte) (((p >>  2) & 0xf8) | ((p >>  7) & 0x7));
+         *b = (ubyte) (((p <<  3) & 0xf8) | ((p >>  2) & 0x7));
+         *a = (ubyte) (0xff * (p >> 15));
+      }
+      return;
+   case PIPE_FORMAT_B4G4R4A4_UNORM:
+      {
+         ushort p = uc->us;
+         *r = (ubyte) (((p >> 4) & 0xf0) | ((p >>  8) & 0xf));
+         *g = (ubyte) (((p >> 0) & 0xf0) | ((p >>  4) & 0xf));
+         *b = (ubyte) (((p << 4) & 0xf0) | ((p >>  0) & 0xf));
+         *a = (ubyte) (((p >> 8) & 0xf0) | ((p >> 12) & 0xf));
+      }
+      return;
+   case PIPE_FORMAT_A8_UNORM:
+      {
+         ubyte p = uc->ub;
+         *r = *g = *b = (ubyte) 0xff;
+         *a = p;
+      }
+      return;
+   case PIPE_FORMAT_L8_UNORM:
+      {
+         ubyte p = uc->ub;
+         *r = *g = *b = p;
+         *a = (ubyte) 0xff;
+      }
+      return;
+   case PIPE_FORMAT_I8_UNORM:
+      {
+         ubyte p = uc->ub;
+         *r = *g = *b = *a = p;
+      }
+      return;
+   case PIPE_FORMAT_R32G32B32A32_FLOAT:
+      {
+         const float *p = &uc->f[0];
+         *r = float_to_ubyte(p[0]);
+         *g = float_to_ubyte(p[1]);
+         *b = float_to_ubyte(p[2]);
+         *a = float_to_ubyte(p[3]);
+      }
+      return;
+   case PIPE_FORMAT_R32G32B32_FLOAT:
+      {
+         const float *p = &uc->f[0];
+         *r = float_to_ubyte(p[0]);
+         *g = float_to_ubyte(p[1]);
+         *b = float_to_ubyte(p[2]);
+         *a = (ubyte) 0xff;
+      }
+      return;
+
+   case PIPE_FORMAT_R32G32_FLOAT:
+      {
+         const float *p = &uc->f[0];
+         *r = float_to_ubyte(p[0]);
+         *g = float_to_ubyte(p[1]);
+         *b = *a = (ubyte) 0xff;
+      }
+      return;
+
+   case PIPE_FORMAT_R32_FLOAT:
+      {
+         const float *p = &uc->f[0];
+         *r = float_to_ubyte(p[0]);
+         *g = *b = *a = (ubyte) 0xff;
+      }
+      return;
+
+   /* Handle other cases with a generic function.
+    */
+   default:
+      {
+         ubyte dst[4];
+
+         util_format_read_4ub(format, dst, 0, uc, 0, 0, 0, 1, 1);
+         *r = dst[0];
+         *g = dst[1];
+         *b = dst[2];
+         *a = dst[3];
+      }
+   }
+}
+
+
+/**
+ * Note rgba outside [0,1] will be clamped for int pixel formats.
+ */
+static INLINE void
+util_pack_color(const float rgba[4], enum pipe_format format, union util_color *uc)
+{
+   ubyte r = 0;
+   ubyte g = 0;
+   ubyte b = 0;
+   ubyte a = 0;
+
+   if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0) <= 8) {
+      /* format uses 8-bit components or less */
+      r = float_to_ubyte(rgba[0]);
+      g = float_to_ubyte(rgba[1]);
+      b = float_to_ubyte(rgba[2]);
+      a = float_to_ubyte(rgba[3]);
+   }
+
+   switch (format) {
+   case PIPE_FORMAT_A8B8G8R8_UNORM:
+      {
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | a;
+      }
+      return;
+   case PIPE_FORMAT_X8B8G8R8_UNORM:
+      {
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B8G8R8A8_UNORM:
+      {
+         uc->ui = (a << 24) | (r << 16) | (g << 8) | b;
+      }
+      return;
+   case PIPE_FORMAT_B8G8R8X8_UNORM:
+      {
+         uc->ui = (0xff << 24) | (r << 16) | (g << 8) | b;
+      }
+      return;
+   case PIPE_FORMAT_A8R8G8B8_UNORM:
+      {
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | a;
+      }
+      return;
+   case PIPE_FORMAT_X8R8G8B8_UNORM:
+      {
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | 0xff;
+      }
+      return;
+   case PIPE_FORMAT_B5G6R5_UNORM:
+      {
+         uc->us = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);
+      }
+      return;
+   case PIPE_FORMAT_B5G5R5X1_UNORM:
+      {
+         uc->us = ((0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
+      }
+      return;
+   case PIPE_FORMAT_B5G5R5A1_UNORM:
+      {
+         uc->us = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
+      }
+      return;
+   case PIPE_FORMAT_B4G4R4A4_UNORM:
+      {
+         uc->us = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);
+      }
+      return;
+   case PIPE_FORMAT_A8_UNORM:
+      {
+         uc->ub = a;
+      }
+      return;
+   case PIPE_FORMAT_L8_UNORM:
+   case PIPE_FORMAT_I8_UNORM:
+      {
+         uc->ub = r;
+      }
+      return;
+   case PIPE_FORMAT_R32G32B32A32_FLOAT:
+      {
+         uc->f[0] = rgba[0];
+         uc->f[1] = rgba[1];
+         uc->f[2] = rgba[2];
+         uc->f[3] = rgba[3];
+      }
+      return;
+   case PIPE_FORMAT_R32G32B32_FLOAT:
+      {
+         uc->f[0] = rgba[0];
+         uc->f[1] = rgba[1];
+         uc->f[2] = rgba[2];
+      }
+      return;
+
+   /* Handle other cases with a generic function.
+    */
+   default:
+      util_format_write_4f(format, rgba, 0, uc, 0, 0, 0, 1, 1);
+   }
+}
+ 
+/* Integer versions of util_pack_z and util_pack_z_stencil - useful for
+ * constructing clear masks.
+ */
+static INLINE uint32_t
+util_pack_mask_z(enum pipe_format format, uint32_t z)
+{
+   switch (format) {
+   case PIPE_FORMAT_Z16_UNORM:
+      return z & 0xffff;
+   case PIPE_FORMAT_Z32_UNORM:
+   case PIPE_FORMAT_Z32_FLOAT:
+      return z;
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+   case PIPE_FORMAT_Z24X8_UNORM:
+      return z & 0xffffff;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+   case PIPE_FORMAT_X8Z24_UNORM:
+      return (z & 0xffffff) << 8;
+   case PIPE_FORMAT_S8_USCALED:
+      return 0;
+   default:
+      debug_print_format("gallium: unhandled format in util_pack_mask_z()", format);
+      assert(0);
+      return 0;
+   }
+}
+
+
+static INLINE uint64_t
+util_pack64_mask_z(enum pipe_format format, uint32_t z)
+{
+   switch (format) {
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      return z;
+   default:
+      return util_pack_mask_z(format, z);
+   }
+}
+
+
+static INLINE uint32_t
+util_pack_mask_z_stencil(enum pipe_format format, uint32_t z, uint8_t s)
+{
+   uint32_t packed = util_pack_mask_z(format, z);
+
+   switch (format) {
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+      packed |= (uint32_t)s << 24;
+      break;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+      packed |= s;
+      break;
+   case PIPE_FORMAT_S8_USCALED:
+      packed |= s;
+      break;
+   default:
+      break;
+   }
+
+   return packed;
+}
+
+
+static INLINE uint64_t
+util_pack64_mask_z_stencil(enum pipe_format format, uint32_t z, uint8_t s)
+{
+   uint64_t packed;
+
+   switch (format) {
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      packed = util_pack64_mask_z(format, z);
+      packed |= (uint64_t)s << 32ull;
+      return packed;
+   default:
+      return util_pack_mask_z_stencil(format, z, s);
+   }
+}
+
+
+/**
+ * Note: it's assumed that z is in [0,1]
+ */
+static INLINE uint32_t
+util_pack_z(enum pipe_format format, double z)
+{
+   union fi fui;
+
+   if (z == 0.0)
+      return 0;
+
+   switch (format) {
+   case PIPE_FORMAT_Z16_UNORM:
+      if (z == 1.0)
+         return 0xffff;
+      return (uint32_t) (z * 0xffff);
+   case PIPE_FORMAT_Z32_UNORM:
+      /* special-case to avoid overflow */
+      if (z == 1.0)
+         return 0xffffffff;
+      return (uint32_t) (z * 0xffffffff);
+   case PIPE_FORMAT_Z32_FLOAT:
+      fui.f = (float)z;
+      return fui.ui;
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+   case PIPE_FORMAT_Z24X8_UNORM:
+      if (z == 1.0)
+         return 0xffffff;
+      return (uint32_t) (z * 0xffffff);
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+   case PIPE_FORMAT_X8Z24_UNORM:
+      if (z == 1.0)
+         return 0xffffff00;
+      return ((uint32_t) (z * 0xffffff)) << 8;
+   case PIPE_FORMAT_S8_USCALED:
+      /* this case can get it via util_pack_z_stencil() */
+      return 0;
+   default:
+      debug_print_format("gallium: unhandled format in util_pack_z()", format);
+      assert(0);
+      return 0;
+   }
+}
+
+
+static INLINE uint64_t
+util_pack64_z(enum pipe_format format, double z)
+{
+   union fi fui;
+
+   if (z == 0)
+      return 0;
+
+   switch (format) {
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      fui.f = (float)z;
+      return fui.ui;
+   default:
+      return util_pack_z(format, z);
+   }
+}
+ 
+
+/**
+ * Pack Z and/or stencil values into a 32-bit value described by format.
+ * Note: it's assumed that z is in [0,1] and s in [0,255]
+ */
+static INLINE uint32_t
+util_pack_z_stencil(enum pipe_format format, double z, uint8_t s)
+{
+   uint32_t packed = util_pack_z(format, z);
+
+   switch (format) {
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+      packed |= (uint32_t)s << 24;
+      break;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+      packed |= s;
+      break;
+   case PIPE_FORMAT_S8_USCALED:
+      packed |= s;
+      break;
+   default:
+      break;
+   }
+
+   return packed;
+}
+
+
+static INLINE uint64_t
+util_pack64_z_stencil(enum pipe_format format, double z, uint8_t s)
+{
+   uint64_t packed;
+
+   switch (format) {
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      packed = util_pack64_z(format, z);
+      packed |= (uint64_t)s << 32ull;
+      break;
+   default:
+      return util_pack_z_stencil(format, z, s);
+   }
+
+   return packed;
+}
+
+
+/**
+ * Pack 4 ubytes into a 4-byte word
+ */
+static INLINE unsigned
+pack_ub4(ubyte b0, ubyte b1, ubyte b2, ubyte b3)
+{
+   return ((((unsigned int)b0) << 0) |
+	   (((unsigned int)b1) << 8) |
+	   (((unsigned int)b2) << 16) |
+	   (((unsigned int)b3) << 24));
+}
+
+
+/**
+ * Pack/convert 4 floats into one 4-byte word.
+ */
+static INLINE unsigned
+pack_ui32_float4(float a, float b, float c, float d)
+{
+   return pack_ub4( float_to_ubyte(a),
+		    float_to_ubyte(b),
+		    float_to_ubyte(c),
+		    float_to_ubyte(d) );
+}
+
+
+
+#endif /* U_PACK_COLOR_H */
diff --git a/mesalib/src/gallium/auxiliary/util/u_surface.c b/mesalib/src/gallium/auxiliary/util/u_surface.c
index 26f2aee99..8e123867d 100644
--- a/mesalib/src/gallium/auxiliary/util/u_surface.c
+++ b/mesalib/src/gallium/auxiliary/util/u_surface.c
@@ -1,371 +1,404 @@
-/**************************************************************************

- *

- * 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 VMWARE 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.

- *

- **************************************************************************/

-

-/**

- * @file

- * Surface utility functions.

- *  

- * @author Brian Paul

- */

-

-

-#include "pipe/p_defines.h"

-#include "pipe/p_screen.h"

-#include "pipe/p_state.h"

-

-#include "util/u_format.h"

-#include "util/u_inlines.h"

-#include "util/u_rect.h"

-#include "util/u_surface.h"

-#include "util/u_pack_color.h"

-

-void

-u_surface_default_template(struct pipe_surface *view,

-                           const struct pipe_resource *texture,

-                           unsigned bind)

-{

-   view->format = texture->format;

-   view->u.tex.level = 0;

-   view->u.tex.first_layer = 0;

-   view->u.tex.last_layer = 0;

-   /* XXX should filter out all non-rt/ds bind flags ? */

-   view->usage = bind;

-}

-

-/**

- * Helper to quickly create an RGBA rendering surface of a certain size.

- * \param textureOut  returns the new texture

- * \param surfaceOut  returns the new surface

- * \return TRUE for success, FALSE if failure

- */

-boolean

-util_create_rgba_surface(struct pipe_context *pipe,

-                         uint width, uint height,

-                         uint bind,

-                         struct pipe_resource **textureOut,

-                         struct pipe_surface **surfaceOut)

-{

-   static const enum pipe_format rgbaFormats[] = {

-      PIPE_FORMAT_B8G8R8A8_UNORM,

-      PIPE_FORMAT_A8R8G8B8_UNORM,

-      PIPE_FORMAT_A8B8G8R8_UNORM,

-      PIPE_FORMAT_NONE

-   };

-   const uint target = PIPE_TEXTURE_2D;

-   enum pipe_format format = PIPE_FORMAT_NONE;

-   struct pipe_resource templ;

-   struct pipe_surface surf_templ;

-   struct pipe_screen *screen = pipe->screen;

-   uint i;

-

-   /* Choose surface format */

-   for (i = 0; rgbaFormats[i]; i++) {

-      if (screen->is_format_supported(screen, rgbaFormats[i],

-                                      target, 0, bind)) {

-         format = rgbaFormats[i];

-         break;

-      }

-   }

-   if (format == PIPE_FORMAT_NONE)

-      return FALSE;  /* unable to get an rgba format!?! */

-

-   /* create texture */

-   memset(&templ, 0, sizeof(templ));

-   templ.target = target;

-   templ.format = format;

-   templ.last_level = 0;

-   templ.width0 = width;

-   templ.height0 = height;

-   templ.depth0 = 1;

-   templ.array_size = 1;

-   templ.bind = bind;

-

-   *textureOut = screen->resource_create(screen, &templ);

-   if (!*textureOut)

-      return FALSE;

-

-   /* create surface */

-   memset(&surf_templ, 0, sizeof(surf_templ));

-   u_surface_default_template(&surf_templ, *textureOut, bind);

-   /* create surface / view into texture */

-   *surfaceOut = pipe->create_surface(pipe,

-                                      *textureOut,

-                                      &surf_templ);

-   if (!*surfaceOut) {

-      pipe_resource_reference(textureOut, NULL);

-      return FALSE;

-   }

-

-   return TRUE;

-}

-

-

-/**

- * Release the surface and texture from util_create_rgba_surface().

- */

-void

-util_destroy_rgba_surface(struct pipe_resource *texture,

-                          struct pipe_surface *surface)

-{

-   pipe_surface_reference(&surface, NULL);

-   pipe_resource_reference(&texture, NULL);

-}

-

-

-

-/**

- * Fallback function for pipe->resource_copy_region().

- * Note: (X,Y)=(0,0) is always the upper-left corner.

- */

-void

-util_resource_copy_region(struct pipe_context *pipe,

-                          struct pipe_resource *dst,

-                          unsigned dst_level,

-                          unsigned dst_x, unsigned dst_y, unsigned dst_z,

-                          struct pipe_resource *src,

-                          unsigned src_level,

-                          const struct pipe_box *src_box)

-{

-   struct pipe_transfer *src_trans, *dst_trans;

-   void *dst_map;

-   const void *src_map;

-   enum pipe_format src_format, dst_format;

-   unsigned w = src_box->width;

-   unsigned h = src_box->height;

-

-   assert(src && dst);

-   assert((src->target == PIPE_BUFFER && dst->target == PIPE_BUFFER) ||

-          (src->target != PIPE_BUFFER && dst->target != PIPE_BUFFER));

-

-   if (!src || !dst)

-      return;

-

-   src_format = src->format;

-   dst_format = dst->format;

-

-   src_trans = pipe_get_transfer(pipe,

-                                 src,

-                                 src_level,

-                                 src_box->z,

-                                 PIPE_TRANSFER_READ,

-                                 src_box->x, src_box->y, w, h);

-

-   dst_trans = pipe_get_transfer(pipe,

-                                 dst,

-                                 dst_level,

-                                 dst_z,

-                                 PIPE_TRANSFER_WRITE,

-                                 dst_x, dst_y, w, h);

-

-   assert(util_format_get_blocksize(dst_format) == util_format_get_blocksize(src_format));

-   assert(util_format_get_blockwidth(dst_format) == util_format_get_blockwidth(src_format));

-   assert(util_format_get_blockheight(dst_format) == util_format_get_blockheight(src_format));

-

-   src_map = pipe->transfer_map(pipe, src_trans);

-   dst_map = pipe->transfer_map(pipe, dst_trans);

-

-   assert(src_map);

-   assert(dst_map);

-

-   if (src_map && dst_map) {

-      if (dst->target == PIPE_BUFFER && src->target == PIPE_BUFFER) {

-         memcpy(dst_map, src_map, w);

-      } else {

-         util_copy_rect(dst_map,

-                        dst_format,

-                        dst_trans->stride,

-                        0, 0,

-                        w, h,

-                        src_map,

-                        src_trans->stride,

-                        0,

-                        0);

-      }

-   }

-

-   pipe->transfer_unmap(pipe, src_trans);

-   pipe->transfer_unmap(pipe, dst_trans);

-

-   pipe->transfer_destroy(pipe, src_trans);

-   pipe->transfer_destroy(pipe, dst_trans);

-}

-

-

-

-#define UBYTE_TO_USHORT(B) ((B) | ((B) << 8))

-

-

-/**

- * Fallback for pipe->clear_render_target() function.

- * XXX this looks too hackish to be really useful.

- * cpp > 4 looks like a gross hack at best...

- * Plus can't use these transfer fallbacks when clearing

- * multisampled surfaces for instance.

- */

-void

-util_clear_render_target(struct pipe_context *pipe,

-                         struct pipe_surface *dst,

-                         const float *rgba,

-                         unsigned dstx, unsigned dsty,

-                         unsigned width, unsigned height)

-{

-   struct pipe_transfer *dst_trans;

-   void *dst_map;

-   union util_color uc;

-

-   assert(dst->texture);

-   if (!dst->texture)

-      return;

-   /* XXX: should handle multiple layers */

-   dst_trans = pipe_get_transfer(pipe,

-                                 dst->texture,

-                                 dst->u.tex.level,

-                                 dst->u.tex.first_layer,

-                                 PIPE_TRANSFER_WRITE,

-                                 dstx, dsty, width, height);

-

-   dst_map = pipe->transfer_map(pipe, dst_trans);

-

-   assert(dst_map);

-

-   if (dst_map) {

-      assert(dst_trans->stride > 0);

-

-      util_pack_color(rgba, dst->texture->format, &uc);

-      util_fill_rect(dst_map, dst->texture->format,

-                     dst_trans->stride,

-                     0, 0, width, height, &uc);

-   }

-

-   pipe->transfer_unmap(pipe, dst_trans);

-   pipe->transfer_destroy(pipe, dst_trans);

-}

-

-/**

- * Fallback for pipe->clear_stencil() function.

- * sw fallback doesn't look terribly useful here.

- * Plus can't use these transfer fallbacks when clearing

- * multisampled surfaces for instance.

- */

-void

-util_clear_depth_stencil(struct pipe_context *pipe,

-                         struct pipe_surface *dst,

-                         unsigned clear_flags,

-                         double depth,

-                         unsigned stencil,

-                         unsigned dstx, unsigned dsty,

-                         unsigned width, unsigned height)

-{

-   struct pipe_transfer *dst_trans;

-   ubyte *dst_map;

-   boolean need_rmw = FALSE;

-

-   if ((clear_flags & PIPE_CLEAR_DEPTHSTENCIL) &&

-       ((clear_flags & PIPE_CLEAR_DEPTHSTENCIL) != PIPE_CLEAR_DEPTHSTENCIL) &&

-       util_format_is_depth_and_stencil(dst->format))

-      need_rmw = TRUE;

-

-   assert(dst->texture);

-   if (!dst->texture)

-      return;

-   dst_trans = pipe_get_transfer(pipe,

-                                 dst->texture,

-                                 dst->u.tex.level,

-                                 dst->u.tex.first_layer,

-                                 (need_rmw ? PIPE_TRANSFER_READ_WRITE :

-                                     PIPE_TRANSFER_WRITE),

-                                 dstx, dsty, width, height);

-

-   dst_map = pipe->transfer_map(pipe, dst_trans);

-

-   assert(dst_map);

-

-   if (dst_map) {

-      unsigned dst_stride = dst_trans->stride;

-      unsigned zstencil = util_pack_z_stencil(dst->texture->format, depth, stencil);

-      unsigned i, j;

-      assert(dst_trans->stride > 0);

-

-      switch (util_format_get_blocksize(dst->format)) {

-      case 1:

-         assert(dst->format == PIPE_FORMAT_S8_USCALED);

-         if(dst_stride == width)

-            memset(dst_map, (ubyte) zstencil, height * width);

-         else {

-            for (i = 0; i < height; i++) {

-               memset(dst_map, (ubyte) zstencil, width);

-               dst_map += dst_stride;

-            }

-         }

-         break;

-      case 2:

-         assert(dst->format == PIPE_FORMAT_Z16_UNORM);

-         for (i = 0; i < height; i++) {

-            uint16_t *row = (uint16_t *)dst_map;

-            for (j = 0; j < width; j++)

-               *row++ = (uint16_t) zstencil;

-            dst_map += dst_stride;

-            }

-         break;

-      case 4:

-         if (!need_rmw) {

-            for (i = 0; i < height; i++) {

-               uint32_t *row = (uint32_t *)dst_map;

-               for (j = 0; j < width; j++)

-                  *row++ = zstencil;

-               dst_map += dst_stride;

-            }

-         }

-         else {

-            uint32_t dst_mask;

-            if (dst->format == PIPE_FORMAT_Z24_UNORM_S8_USCALED)

-               dst_mask = 0xffffff00;

-            else {

-               assert(dst->format == PIPE_FORMAT_S8_USCALED_Z24_UNORM);

-               dst_mask = 0xffffff;

-            }

-            if (clear_flags & PIPE_CLEAR_DEPTH)

-               dst_mask = ~dst_mask;

-            for (i = 0; i < height; i++) {

-               uint32_t *row = (uint32_t *)dst_map;

-               for (j = 0; j < width; j++) {

-                  uint32_t tmp = *row & dst_mask;

-                  *row++ = tmp | (zstencil & ~dst_mask);

-               }

-               dst_map += dst_stride;

-            }

-         }

-        break;

-      case 8:

-      default:

-         assert(0);

-         break;

-      }

-   }

-

-   pipe->transfer_unmap(pipe, dst_trans);

-   pipe->transfer_destroy(pipe, dst_trans);

-}

+/**************************************************************************
+ *
+ * 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 VMWARE 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.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * Surface utility functions.
+ *  
+ * @author Brian Paul
+ */
+
+
+#include "pipe/p_defines.h"
+#include "pipe/p_screen.h"
+#include "pipe/p_state.h"
+
+#include "util/u_format.h"
+#include "util/u_inlines.h"
+#include "util/u_rect.h"
+#include "util/u_surface.h"
+#include "util/u_pack_color.h"
+
+void
+u_surface_default_template(struct pipe_surface *view,
+                           const struct pipe_resource *texture,
+                           unsigned bind)
+{
+   view->format = texture->format;
+   view->u.tex.level = 0;
+   view->u.tex.first_layer = 0;
+   view->u.tex.last_layer = 0;
+   /* XXX should filter out all non-rt/ds bind flags ? */
+   view->usage = bind;
+}
+
+/**
+ * Helper to quickly create an RGBA rendering surface of a certain size.
+ * \param textureOut  returns the new texture
+ * \param surfaceOut  returns the new surface
+ * \return TRUE for success, FALSE if failure
+ */
+boolean
+util_create_rgba_surface(struct pipe_context *pipe,
+                         uint width, uint height,
+                         uint bind,
+                         struct pipe_resource **textureOut,
+                         struct pipe_surface **surfaceOut)
+{
+   static const enum pipe_format rgbaFormats[] = {
+      PIPE_FORMAT_B8G8R8A8_UNORM,
+      PIPE_FORMAT_A8R8G8B8_UNORM,
+      PIPE_FORMAT_A8B8G8R8_UNORM,
+      PIPE_FORMAT_NONE
+   };
+   const uint target = PIPE_TEXTURE_2D;
+   enum pipe_format format = PIPE_FORMAT_NONE;
+   struct pipe_resource templ;
+   struct pipe_surface surf_templ;
+   struct pipe_screen *screen = pipe->screen;
+   uint i;
+
+   /* Choose surface format */
+   for (i = 0; rgbaFormats[i]; i++) {
+      if (screen->is_format_supported(screen, rgbaFormats[i],
+                                      target, 0, bind)) {
+         format = rgbaFormats[i];
+         break;
+      }
+   }
+   if (format == PIPE_FORMAT_NONE)
+      return FALSE;  /* unable to get an rgba format!?! */
+
+   /* create texture */
+   memset(&templ, 0, sizeof(templ));
+   templ.target = target;
+   templ.format = format;
+   templ.last_level = 0;
+   templ.width0 = width;
+   templ.height0 = height;
+   templ.depth0 = 1;
+   templ.array_size = 1;
+   templ.bind = bind;
+
+   *textureOut = screen->resource_create(screen, &templ);
+   if (!*textureOut)
+      return FALSE;
+
+   /* create surface */
+   memset(&surf_templ, 0, sizeof(surf_templ));
+   u_surface_default_template(&surf_templ, *textureOut, bind);
+   /* create surface / view into texture */
+   *surfaceOut = pipe->create_surface(pipe,
+                                      *textureOut,
+                                      &surf_templ);
+   if (!*surfaceOut) {
+      pipe_resource_reference(textureOut, NULL);
+      return FALSE;
+   }
+
+   return TRUE;
+}
+
+
+/**
+ * Release the surface and texture from util_create_rgba_surface().
+ */
+void
+util_destroy_rgba_surface(struct pipe_resource *texture,
+                          struct pipe_surface *surface)
+{
+   pipe_surface_reference(&surface, NULL);
+   pipe_resource_reference(&texture, NULL);
+}
+
+
+
+/**
+ * Fallback function for pipe->resource_copy_region().
+ * Note: (X,Y)=(0,0) is always the upper-left corner.
+ */
+void
+util_resource_copy_region(struct pipe_context *pipe,
+                          struct pipe_resource *dst,
+                          unsigned dst_level,
+                          unsigned dst_x, unsigned dst_y, unsigned dst_z,
+                          struct pipe_resource *src,
+                          unsigned src_level,
+                          const struct pipe_box *src_box)
+{
+   struct pipe_transfer *src_trans, *dst_trans;
+   void *dst_map;
+   const void *src_map;
+   enum pipe_format src_format, dst_format;
+   unsigned w = src_box->width;
+   unsigned h = src_box->height;
+
+   assert(src && dst);
+   assert((src->target == PIPE_BUFFER && dst->target == PIPE_BUFFER) ||
+          (src->target != PIPE_BUFFER && dst->target != PIPE_BUFFER));
+
+   if (!src || !dst)
+      return;
+
+   src_format = src->format;
+   dst_format = dst->format;
+
+   src_trans = pipe_get_transfer(pipe,
+                                 src,
+                                 src_level,
+                                 src_box->z,
+                                 PIPE_TRANSFER_READ,
+                                 src_box->x, src_box->y, w, h);
+
+   dst_trans = pipe_get_transfer(pipe,
+                                 dst,
+                                 dst_level,
+                                 dst_z,
+                                 PIPE_TRANSFER_WRITE,
+                                 dst_x, dst_y, w, h);
+
+   assert(util_format_get_blocksize(dst_format) == util_format_get_blocksize(src_format));
+   assert(util_format_get_blockwidth(dst_format) == util_format_get_blockwidth(src_format));
+   assert(util_format_get_blockheight(dst_format) == util_format_get_blockheight(src_format));
+
+   src_map = pipe->transfer_map(pipe, src_trans);
+   dst_map = pipe->transfer_map(pipe, dst_trans);
+
+   assert(src_map);
+   assert(dst_map);
+
+   if (src_map && dst_map) {
+      if (dst->target == PIPE_BUFFER && src->target == PIPE_BUFFER) {
+         memcpy(dst_map, src_map, w);
+      } else {
+         util_copy_rect(dst_map,
+                        dst_format,
+                        dst_trans->stride,
+                        0, 0,
+                        w, h,
+                        src_map,
+                        src_trans->stride,
+                        0,
+                        0);
+      }
+   }
+
+   pipe->transfer_unmap(pipe, src_trans);
+   pipe->transfer_unmap(pipe, dst_trans);
+
+   pipe->transfer_destroy(pipe, src_trans);
+   pipe->transfer_destroy(pipe, dst_trans);
+}
+
+
+
+#define UBYTE_TO_USHORT(B) ((B) | ((B) << 8))
+
+
+/**
+ * Fallback for pipe->clear_render_target() function.
+ * XXX this looks too hackish to be really useful.
+ * cpp > 4 looks like a gross hack at best...
+ * Plus can't use these transfer fallbacks when clearing
+ * multisampled surfaces for instance.
+ */
+void
+util_clear_render_target(struct pipe_context *pipe,
+                         struct pipe_surface *dst,
+                         const float *rgba,
+                         unsigned dstx, unsigned dsty,
+                         unsigned width, unsigned height)
+{
+   struct pipe_transfer *dst_trans;
+   void *dst_map;
+   union util_color uc;
+
+   assert(dst->texture);
+   if (!dst->texture)
+      return;
+   /* XXX: should handle multiple layers */
+   dst_trans = pipe_get_transfer(pipe,
+                                 dst->texture,
+                                 dst->u.tex.level,
+                                 dst->u.tex.first_layer,
+                                 PIPE_TRANSFER_WRITE,
+                                 dstx, dsty, width, height);
+
+   dst_map = pipe->transfer_map(pipe, dst_trans);
+
+   assert(dst_map);
+
+   if (dst_map) {
+      assert(dst_trans->stride > 0);
+
+      util_pack_color(rgba, dst->texture->format, &uc);
+      util_fill_rect(dst_map, dst->texture->format,
+                     dst_trans->stride,
+                     0, 0, width, height, &uc);
+   }
+
+   pipe->transfer_unmap(pipe, dst_trans);
+   pipe->transfer_destroy(pipe, dst_trans);
+}
+
+/**
+ * Fallback for pipe->clear_stencil() function.
+ * sw fallback doesn't look terribly useful here.
+ * Plus can't use these transfer fallbacks when clearing
+ * multisampled surfaces for instance.
+ */
+void
+util_clear_depth_stencil(struct pipe_context *pipe,
+                         struct pipe_surface *dst,
+                         unsigned clear_flags,
+                         double depth,
+                         unsigned stencil,
+                         unsigned dstx, unsigned dsty,
+                         unsigned width, unsigned height)
+{
+   struct pipe_transfer *dst_trans;
+   ubyte *dst_map;
+   boolean need_rmw = FALSE;
+
+   if ((clear_flags & PIPE_CLEAR_DEPTHSTENCIL) &&
+       ((clear_flags & PIPE_CLEAR_DEPTHSTENCIL) != PIPE_CLEAR_DEPTHSTENCIL) &&
+       util_format_is_depth_and_stencil(dst->format))
+      need_rmw = TRUE;
+
+   assert(dst->texture);
+   if (!dst->texture)
+      return;
+   dst_trans = pipe_get_transfer(pipe,
+                                 dst->texture,
+                                 dst->u.tex.level,
+                                 dst->u.tex.first_layer,
+                                 (need_rmw ? PIPE_TRANSFER_READ_WRITE :
+                                     PIPE_TRANSFER_WRITE),
+                                 dstx, dsty, width, height);
+
+   dst_map = pipe->transfer_map(pipe, dst_trans);
+
+   assert(dst_map);
+
+   if (dst_map) {
+      unsigned dst_stride = dst_trans->stride;
+      unsigned zstencil = util_pack_z_stencil(dst->texture->format, depth, stencil);
+      unsigned i, j;
+      assert(dst_trans->stride > 0);
+
+      switch (util_format_get_blocksize(dst->format)) {
+      case 1:
+         assert(dst->format == PIPE_FORMAT_S8_USCALED);
+         if(dst_stride == width)
+            memset(dst_map, (ubyte) zstencil, height * width);
+         else {
+            for (i = 0; i < height; i++) {
+               memset(dst_map, (ubyte) zstencil, width);
+               dst_map += dst_stride;
+            }
+         }
+         break;
+      case 2:
+         assert(dst->format == PIPE_FORMAT_Z16_UNORM);
+         for (i = 0; i < height; i++) {
+            uint16_t *row = (uint16_t *)dst_map;
+            for (j = 0; j < width; j++)
+               *row++ = (uint16_t) zstencil;
+            dst_map += dst_stride;
+            }
+         break;
+      case 4:
+         if (!need_rmw) {
+            for (i = 0; i < height; i++) {
+               uint32_t *row = (uint32_t *)dst_map;
+               for (j = 0; j < width; j++)
+                  *row++ = zstencil;
+               dst_map += dst_stride;
+            }
+         }
+         else {
+            uint32_t dst_mask;
+            if (dst->format == PIPE_FORMAT_Z24_UNORM_S8_USCALED)
+               dst_mask = 0xffffff00;
+            else {
+               assert(dst->format == PIPE_FORMAT_S8_USCALED_Z24_UNORM);
+               dst_mask = 0xffffff;
+            }
+            if (clear_flags & PIPE_CLEAR_DEPTH)
+               dst_mask = ~dst_mask;
+            for (i = 0; i < height; i++) {
+               uint32_t *row = (uint32_t *)dst_map;
+               for (j = 0; j < width; j++) {
+                  uint32_t tmp = *row & dst_mask;
+                  *row++ = tmp | (zstencil & ~dst_mask);
+               }
+               dst_map += dst_stride;
+            }
+         }
+         break;
+      case 8:
+      {
+         uint64_t zstencil = util_pack64_z_stencil(dst->texture->format,
+                                                   depth, stencil);
+
+         assert(dst->format == PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED);
+
+         if (!need_rmw) {
+            for (i = 0; i < height; i++) {
+               uint64_t *row = (uint64_t *)dst_map;
+               for (j = 0; j < width; j++)
+                  *row++ = zstencil;
+               dst_map += dst_stride;
+            }
+         }
+         else {
+            uint64_t src_mask;
+
+            if (clear_flags & PIPE_CLEAR_DEPTH)
+               src_mask = 0x00000000ffffffffull;
+            else
+               src_mask = 0x000000ff00000000ull;
+
+            for (i = 0; i < height; i++) {
+               uint64_t *row = (uint64_t *)dst_map;
+               for (j = 0; j < width; j++) {
+                  uint64_t tmp = *row & ~src_mask;
+                  *row++ = tmp | (zstencil & src_mask);
+               }
+               dst_map += dst_stride;
+            }
+         }
+         break;
+      }
+      default:
+         assert(0);
+         break;
+      }
+   }
+
+   pipe->transfer_unmap(pipe, dst_trans);
+   pipe->transfer_destroy(pipe, dst_trans);
+}
diff --git a/mesalib/src/gallium/auxiliary/util/u_tile.c b/mesalib/src/gallium/auxiliary/util/u_tile.c
index 54f634948..23f12e5f4 100644
--- a/mesalib/src/gallium/auxiliary/util/u_tile.c
+++ b/mesalib/src/gallium/auxiliary/util/u_tile.c
@@ -1,655 +1,690 @@
-/**************************************************************************

- * 

- * 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.

- * 

- **************************************************************************/

-

-/**

- * RGBA/float tile get/put functions.

- * Usable both by drivers and state trackers.

- */

-

-

-#include "pipe/p_defines.h"

-#include "util/u_inlines.h"

-

-#include "util/u_format.h"

-#include "util/u_math.h"

-#include "util/u_memory.h"

-#include "util/u_rect.h"

-#include "util/u_tile.h"

-

-

-/**

- * Move raw block of pixels from transfer object to user memory.

- */

-void

-pipe_get_tile_raw(struct pipe_context *pipe,

-                  struct pipe_transfer *pt,

-                  uint x, uint y, uint w, uint h,

-                  void *dst, int dst_stride)

-{

-   const void *src;

-

-   if (dst_stride == 0)

-      dst_stride = util_format_get_stride(pt->resource->format, w);

-

-   if (u_clip_tile(x, y, &w, &h, &pt->box))

-      return;

-

-   src = pipe->transfer_map(pipe, pt);

-   assert(src);

-   if(!src)

-      return;

-

-   util_copy_rect(dst, pt->resource->format, dst_stride, 0, 0, w, h, src, pt->stride, x, y);

-

-   pipe->transfer_unmap(pipe, pt);

-}

-

-

-/**

- * Move raw block of pixels from user memory to transfer object.

- */

-void

-pipe_put_tile_raw(struct pipe_context *pipe,

-                  struct pipe_transfer *pt,

-                  uint x, uint y, uint w, uint h,

-                  const void *src, int src_stride)

-{

-   void *dst;

-   enum pipe_format format = pt->resource->format;

-

-   if (src_stride == 0)

-      src_stride = util_format_get_stride(format, w);

-

-   if (u_clip_tile(x, y, &w, &h, &pt->box))

-      return;

-

-   dst = pipe->transfer_map(pipe, pt);

-   assert(dst);

-   if(!dst)

-      return;

-

-   util_copy_rect(dst, format, pt->stride, x, y, w, h, src, src_stride, 0, 0);

-

-   pipe->transfer_unmap(pipe, pt);

-}

-

-

-

-

-/** Convert short in [-32768,32767] to GLfloat in [-1.0,1.0] */

-#define SHORT_TO_FLOAT(S)   ((2.0F * (S) + 1.0F) * (1.0F/65535.0F))

-

-#define UNCLAMPED_FLOAT_TO_SHORT(us, f)  \

-   us = ( (short) ( CLAMP((f), -1.0, 1.0) * 32767.0F) )

-

-

-

-/*** PIPE_FORMAT_Z16_UNORM ***/

-

-/**

- * Return each Z value as four floats in [0,1].

- */

-static void

-z16_get_tile_rgba(const ushort *src,

-                  unsigned w, unsigned h,

-                  float *p,

-                  unsigned dst_stride)

-{

-   const float scale = 1.0f / 65535.0f;

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = *src++ * scale;

-      }

-      p += dst_stride;

-   }

-}

-

-

-

-

-/*** PIPE_FORMAT_Z32_UNORM ***/

-

-/**

- * Return each Z value as four floats in [0,1].

- */

-static void

-z32_get_tile_rgba(const unsigned *src,

-                  unsigned w, unsigned h,

-                  float *p,

-                  unsigned dst_stride)

-{

-   const double scale = 1.0 / (double) 0xffffffff;

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = (float) (*src++ * scale);

-      }

-      p += dst_stride;

-   }

-}

-

-

-/*** PIPE_FORMAT_Z24_UNORM_S8_USCALED ***/

-

-/**

- * Return Z component as four float in [0,1].  Stencil part ignored.

- */

-static void

-s8z24_get_tile_rgba(const unsigned *src,

-                    unsigned w, unsigned h,

-                    float *p,

-                    unsigned dst_stride)

-{

-   const double scale = 1.0 / ((1 << 24) - 1);

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = (float) (scale * (*src++ & 0xffffff));

-      }

-      p += dst_stride;

-   }

-}

-

-

-/*** PIPE_FORMAT_S8_USCALED_Z24_UNORM ***/

-

-/**

- * Return Z component as four float in [0,1].  Stencil part ignored.

- */

-static void

-z24s8_get_tile_rgba(const unsigned *src,

-                    unsigned w, unsigned h,

-                    float *p,

-                    unsigned dst_stride)

-{

-   const double scale = 1.0 / ((1 << 24) - 1);

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = (float) (scale * (*src++ >> 8));

-      }

-      p += dst_stride;

-   }

-}

-

-/*** PIPE_FORMAT_S8X24_USCALED ***/

-

-/**

- * Return S component as four uint32_t in [0..255].  Z part ignored.

- */

-static void

-s8x24_get_tile_rgba(const unsigned *src,

-                    unsigned w, unsigned h,

-                    float *p,

-                    unsigned dst_stride)

-{

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = (float)((*src++ >> 24) & 0xff);

-      }

-

-      p += dst_stride;

-   }

-}

-

-/*** PIPE_FORMAT_X24S8_USCALED ***/

-

-/**

- * Return S component as four uint32_t in [0..255].  Z part ignored.

- */

-static void

-x24s8_get_tile_rgba(const unsigned *src,

-                    unsigned w, unsigned h,

-                    float *p,

-                    unsigned dst_stride)

-{

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = (float)(*src++ & 0xff);

-      }

-      p += dst_stride;

-   }

-}

-

-

-/**

- * Return S component as four uint32_t in [0..255].  Z part ignored.

- */

-static void

-s8_get_tile_rgba(const unsigned char *src,

-		 unsigned w, unsigned h,

-		 float *p,

-		 unsigned dst_stride)

-{

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = (float)(*src++ & 0xff);

-      }

-      p += dst_stride;

-   }

-}

-

-/*** PIPE_FORMAT_Z32_FLOAT ***/

-

-/**

- * Return each Z value as four floats in [0,1].

- */

-static void

-z32f_get_tile_rgba(const float *src,

-                   unsigned w, unsigned h,

-                   float *p,

-                   unsigned dst_stride)

-{

-   unsigned i, j;

-

-   for (i = 0; i < h; i++) {

-      float *pRow = p;

-      for (j = 0; j < w; j++, pRow += 4) {

-         pRow[0] =

-         pRow[1] =

-         pRow[2] =

-         pRow[3] = *src++;

-      }

-      p += dst_stride;

-   }

-}

-

-

-void

-pipe_tile_raw_to_rgba(enum pipe_format format,

-                      void *src,

-                      uint w, uint h,

-                      float *dst, unsigned dst_stride)

-{

-   switch (format) {

-   case PIPE_FORMAT_Z16_UNORM:

-      z16_get_tile_rgba((ushort *) src, w, h, dst, dst_stride);

-      break;

-   case PIPE_FORMAT_Z32_UNORM:

-      z32_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);

-      break;

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-   case PIPE_FORMAT_Z24X8_UNORM:

-      s8z24_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);

-      break;

-   case PIPE_FORMAT_S8_USCALED:

-      s8_get_tile_rgba((unsigned char *) src, w, h, dst, dst_stride);

-      break;

-   case PIPE_FORMAT_X24S8_USCALED:

-      s8x24_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);

-      break;

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-   case PIPE_FORMAT_X8Z24_UNORM:

-      z24s8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);

-      break;

-   case PIPE_FORMAT_S8X24_USCALED:

-      x24s8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);

-      break;

-   case PIPE_FORMAT_Z32_FLOAT:

-      z32f_get_tile_rgba((float *) src, w, h, dst, dst_stride);

-      break;

-   default:

-      util_format_read_4f(format,

-                          dst, dst_stride * sizeof(float),

-                          src, util_format_get_stride(format, w),

-                          0, 0, w, h);

-   }

-}

-

-

-void

-pipe_get_tile_rgba(struct pipe_context *pipe,

-                   struct pipe_transfer *pt,

-                   uint x, uint y, uint w, uint h,

-                   float *p)

-{

-   pipe_get_tile_rgba_format(pipe, pt, x, y, w, h, pt->resource->format, p);

-}

-

-

-void

-pipe_get_tile_rgba_format(struct pipe_context *pipe,

-                          struct pipe_transfer *pt,

-                          uint x, uint y, uint w, uint h,

-                          enum pipe_format format,

-                          float *p)

-{

-   unsigned dst_stride = w * 4;

-   void *packed;

-

-   if (u_clip_tile(x, y, &w, &h, &pt->box)) {

-      return;

-   }

-

-   packed = MALLOC(util_format_get_nblocks(format, w, h) * util_format_get_blocksize(format));

-   if (!packed) {

-      return;

-   }

-

-   if (format == PIPE_FORMAT_UYVY || format == PIPE_FORMAT_YUYV) {

-      assert((x & 1) == 0);

-   }

-

-   pipe_get_tile_raw(pipe, pt, x, y, w, h, packed, 0);

-

-   pipe_tile_raw_to_rgba(format, packed, w, h, p, dst_stride);

-

-   FREE(packed);

-}

-

-

-void

-pipe_put_tile_rgba(struct pipe_context *pipe,

-                   struct pipe_transfer *pt,

-                   uint x, uint y, uint w, uint h,

-                   const float *p)

-{

-   pipe_put_tile_rgba_format(pipe, pt, x, y, w, h, pt->resource->format, p);

-}

-

-

-void

-pipe_put_tile_rgba_format(struct pipe_context *pipe,

-                          struct pipe_transfer *pt,

-                          uint x, uint y, uint w, uint h,

-                          enum pipe_format format,

-                          const float *p)

-{

-   unsigned src_stride = w * 4;

-   void *packed;

-

-   if (u_clip_tile(x, y, &w, &h, &pt->box))

-      return;

-

-   packed = MALLOC(util_format_get_nblocks(format, w, h) * util_format_get_blocksize(format));

-

-   if (!packed)

-      return;

-

-   switch (format) {

-   case PIPE_FORMAT_Z16_UNORM:

-      /*z16_put_tile_rgba((ushort *) packed, w, h, p, src_stride);*/

-      break;

-   case PIPE_FORMAT_Z32_UNORM:

-      /*z32_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/

-      break;

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-   case PIPE_FORMAT_Z24X8_UNORM:

-      /*s8z24_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/

-      break;

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-   case PIPE_FORMAT_X8Z24_UNORM:

-      /*z24s8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/

-      break;

-   default:

-      util_format_write_4f(format,

-                           p, src_stride * sizeof(float),

-                           packed, util_format_get_stride(format, w),

-                           0, 0, w, h);

-   }

-

-   pipe_put_tile_raw(pipe, pt, x, y, w, h, packed, 0);

-

-   FREE(packed);

-}

-

-

-/**

- * Get a block of Z values, converted to 32-bit range.

- */

-void

-pipe_get_tile_z(struct pipe_context *pipe,

-                struct pipe_transfer *pt,

-                uint x, uint y, uint w, uint h,

-                uint *z)

-{

-   const uint dstStride = w;

-   ubyte *map;

-   uint *pDest = z;

-   uint i, j;

-   enum pipe_format format = pt->resource->format;

-

-   if (u_clip_tile(x, y, &w, &h, &pt->box))

-      return;

-

-   map = (ubyte *)pipe->transfer_map(pipe, pt);

-   if (!map) {

-      assert(0);

-      return;

-   }

-

-   switch (format) {

-   case PIPE_FORMAT_Z32_UNORM:

-      {

-         const uint *ptrc

-            = (const uint *)(map  + y * pt->stride + x*4);

-         for (i = 0; i < h; i++) {

-            memcpy(pDest, ptrc, 4 * w);

-            pDest += dstStride;

-            ptrc += pt->stride/4;

-         }

-      }

-      break;

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-   case PIPE_FORMAT_Z24X8_UNORM:

-      {

-         const uint *ptrc

-            = (const uint *)(map + y * pt->stride + x*4);

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 24-bit Z to 32-bit Z */

-               pDest[j] = (ptrc[j] << 8) | ((ptrc[j] >> 16) & 0xff);

-            }

-            pDest += dstStride;

-            ptrc += pt->stride/4;

-         }

-      }

-      break;

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-   case PIPE_FORMAT_X8Z24_UNORM:

-      {

-         const uint *ptrc

-            = (const uint *)(map + y * pt->stride + x*4);

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 24-bit Z to 32-bit Z */

-               pDest[j] = (ptrc[j] & 0xffffff00) | ((ptrc[j] >> 24) & 0xff);

-            }

-            pDest += dstStride;

-            ptrc += pt->stride/4;

-         }

-      }

-      break;

-   case PIPE_FORMAT_Z16_UNORM:

-      {

-         const ushort *ptrc

-            = (const ushort *)(map + y * pt->stride + x*2);

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 16-bit Z to 32-bit Z */

-               pDest[j] = (ptrc[j] << 16) | ptrc[j];

-            }

-            pDest += dstStride;

-            ptrc += pt->stride/2;

-         }

-      }

-      break;

-   default:

-      assert(0);

-   }

-

-   pipe->transfer_unmap(pipe, pt);

-}

-

-

-void

-pipe_put_tile_z(struct pipe_context *pipe,

-                struct pipe_transfer *pt,

-                uint x, uint y, uint w, uint h,

-                const uint *zSrc)

-{

-   const uint srcStride = w;

-   const uint *ptrc = zSrc;

-   ubyte *map;

-   uint i, j;

-   enum pipe_format format = pt->resource->format;

-

-   if (u_clip_tile(x, y, &w, &h, &pt->box))

-      return;

-

-   map = (ubyte *)pipe->transfer_map(pipe, pt);

-   if (!map) {

-      assert(0);

-      return;

-   }

-

-   switch (format) {

-   case PIPE_FORMAT_Z32_UNORM:

-      {

-         uint *pDest = (uint *) (map + y * pt->stride + x*4);

-         for (i = 0; i < h; i++) {

-            memcpy(pDest, ptrc, 4 * w);

-            pDest += pt->stride/4;

-            ptrc += srcStride;

-         }

-      }

-      break;

-   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:

-      {

-         uint *pDest = (uint *) (map + y * pt->stride + x*4);

-         /*assert((pt->usage & PIPE_TRANSFER_READ_WRITE) == PIPE_TRANSFER_READ_WRITE);*/

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 32-bit Z to 24-bit Z, preserve stencil */

-               pDest[j] = (pDest[j] & 0xff000000) | ptrc[j] >> 8;

-            }

-            pDest += pt->stride/4;

-            ptrc += srcStride;

-         }

-      }

-      break;

-   case PIPE_FORMAT_Z24X8_UNORM:

-      {

-         uint *pDest = (uint *) (map + y * pt->stride + x*4);

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 32-bit Z to 24-bit Z (0 stencil) */

-               pDest[j] = ptrc[j] >> 8;

-            }

-            pDest += pt->stride/4;

-            ptrc += srcStride;

-         }

-      }

-      break;

-   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:

-      {

-         uint *pDest = (uint *) (map + y * pt->stride + x*4);

-         /*assert((pt->usage & PIPE_TRANSFER_READ_WRITE) == PIPE_TRANSFER_READ_WRITE);*/

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 32-bit Z to 24-bit Z, preserve stencil */

-               pDest[j] = (pDest[j] & 0xff) | (ptrc[j] & 0xffffff00);

-            }

-            pDest += pt->stride/4;

-            ptrc += srcStride;

-         }

-      }

-      break;

-   case PIPE_FORMAT_X8Z24_UNORM:

-      {

-         uint *pDest = (uint *) (map + y * pt->stride + x*4);

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 32-bit Z to 24-bit Z (0 stencil) */

-               pDest[j] = ptrc[j] & 0xffffff00;

-            }

-            pDest += pt->stride/4;

-            ptrc += srcStride;

-         }

-      }

-      break;

-   case PIPE_FORMAT_Z16_UNORM:

-      {

-         ushort *pDest = (ushort *) (map + y * pt->stride + x*2);

-         for (i = 0; i < h; i++) {

-            for (j = 0; j < w; j++) {

-               /* convert 32-bit Z to 16-bit Z */

-               pDest[j] = ptrc[j] >> 16;

-            }

-            pDest += pt->stride/2;

-            ptrc += srcStride;

-         }

-      }

-      break;

-   default:

-      assert(0);

-   }

-

-   pipe->transfer_unmap(pipe, pt);

-}

-

-

+/**************************************************************************
+ * 
+ * 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.
+ * 
+ **************************************************************************/
+
+/**
+ * RGBA/float tile get/put functions.
+ * Usable both by drivers and state trackers.
+ */
+
+
+#include "pipe/p_defines.h"
+#include "util/u_inlines.h"
+
+#include "util/u_format.h"
+#include "util/u_math.h"
+#include "util/u_memory.h"
+#include "util/u_rect.h"
+#include "util/u_tile.h"
+
+
+/**
+ * Move raw block of pixels from transfer object to user memory.
+ */
+void
+pipe_get_tile_raw(struct pipe_context *pipe,
+                  struct pipe_transfer *pt,
+                  uint x, uint y, uint w, uint h,
+                  void *dst, int dst_stride)
+{
+   const void *src;
+
+   if (dst_stride == 0)
+      dst_stride = util_format_get_stride(pt->resource->format, w);
+
+   if (u_clip_tile(x, y, &w, &h, &pt->box))
+      return;
+
+   src = pipe->transfer_map(pipe, pt);
+   assert(src);
+   if(!src)
+      return;
+
+   util_copy_rect(dst, pt->resource->format, dst_stride, 0, 0, w, h, src, pt->stride, x, y);
+
+   pipe->transfer_unmap(pipe, pt);
+}
+
+
+/**
+ * Move raw block of pixels from user memory to transfer object.
+ */
+void
+pipe_put_tile_raw(struct pipe_context *pipe,
+                  struct pipe_transfer *pt,
+                  uint x, uint y, uint w, uint h,
+                  const void *src, int src_stride)
+{
+   void *dst;
+   enum pipe_format format = pt->resource->format;
+
+   if (src_stride == 0)
+      src_stride = util_format_get_stride(format, w);
+
+   if (u_clip_tile(x, y, &w, &h, &pt->box))
+      return;
+
+   dst = pipe->transfer_map(pipe, pt);
+   assert(dst);
+   if(!dst)
+      return;
+
+   util_copy_rect(dst, format, pt->stride, x, y, w, h, src, src_stride, 0, 0);
+
+   pipe->transfer_unmap(pipe, pt);
+}
+
+
+
+
+/** Convert short in [-32768,32767] to GLfloat in [-1.0,1.0] */
+#define SHORT_TO_FLOAT(S)   ((2.0F * (S) + 1.0F) * (1.0F/65535.0F))
+
+#define UNCLAMPED_FLOAT_TO_SHORT(us, f)  \
+   us = ( (short) ( CLAMP((f), -1.0, 1.0) * 32767.0F) )
+
+
+
+/*** PIPE_FORMAT_Z16_UNORM ***/
+
+/**
+ * Return each Z value as four floats in [0,1].
+ */
+static void
+z16_get_tile_rgba(const ushort *src,
+                  unsigned w, unsigned h,
+                  float *p,
+                  unsigned dst_stride)
+{
+   const float scale = 1.0f / 65535.0f;
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = *src++ * scale;
+      }
+      p += dst_stride;
+   }
+}
+
+
+
+
+/*** PIPE_FORMAT_Z32_UNORM ***/
+
+/**
+ * Return each Z value as four floats in [0,1].
+ */
+static void
+z32_get_tile_rgba(const unsigned *src,
+                  unsigned w, unsigned h,
+                  float *p,
+                  unsigned dst_stride)
+{
+   const double scale = 1.0 / (double) 0xffffffff;
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float) (*src++ * scale);
+      }
+      p += dst_stride;
+   }
+}
+
+
+/*** PIPE_FORMAT_Z24_UNORM_S8_USCALED ***/
+
+/**
+ * Return Z component as four float in [0,1].  Stencil part ignored.
+ */
+static void
+s8z24_get_tile_rgba(const unsigned *src,
+                    unsigned w, unsigned h,
+                    float *p,
+                    unsigned dst_stride)
+{
+   const double scale = 1.0 / ((1 << 24) - 1);
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float) (scale * (*src++ & 0xffffff));
+      }
+      p += dst_stride;
+   }
+}
+
+
+/*** PIPE_FORMAT_S8_USCALED_Z24_UNORM ***/
+
+/**
+ * Return Z component as four float in [0,1].  Stencil part ignored.
+ */
+static void
+z24s8_get_tile_rgba(const unsigned *src,
+                    unsigned w, unsigned h,
+                    float *p,
+                    unsigned dst_stride)
+{
+   const double scale = 1.0 / ((1 << 24) - 1);
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float) (scale * (*src++ >> 8));
+      }
+      p += dst_stride;
+   }
+}
+
+/*** PIPE_FORMAT_S8X24_USCALED ***/
+
+/**
+ * Return S component as four uint32_t in [0..255].  Z part ignored.
+ */
+static void
+s8x24_get_tile_rgba(const unsigned *src,
+                    unsigned w, unsigned h,
+                    float *p,
+                    unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float)((*src++ >> 24) & 0xff);
+      }
+
+      p += dst_stride;
+   }
+}
+
+/*** PIPE_FORMAT_X24S8_USCALED ***/
+
+/**
+ * Return S component as four uint32_t in [0..255].  Z part ignored.
+ */
+static void
+x24s8_get_tile_rgba(const unsigned *src,
+                    unsigned w, unsigned h,
+                    float *p,
+                    unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float)(*src++ & 0xff);
+      }
+      p += dst_stride;
+   }
+}
+
+
+/**
+ * Return S component as four uint32_t in [0..255].  Z part ignored.
+ */
+static void
+s8_get_tile_rgba(const unsigned char *src,
+		 unsigned w, unsigned h,
+		 float *p,
+		 unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = (float)(*src++ & 0xff);
+      }
+      p += dst_stride;
+   }
+}
+
+/*** PIPE_FORMAT_Z32_FLOAT ***/
+
+/**
+ * Return each Z value as four floats in [0,1].
+ */
+static void
+z32f_get_tile_rgba(const float *src,
+                   unsigned w, unsigned h,
+                   float *p,
+                   unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = *src++;
+      }
+      p += dst_stride;
+   }
+}
+
+/*** PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED ***/
+
+/**
+ * Return each Z value as four floats in [0,1].
+ */
+static void
+z32f_x24s8_get_tile_rgba(const float *src,
+                         unsigned w, unsigned h,
+                         float *p,
+                         unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         pRow[0] =
+         pRow[1] =
+         pRow[2] =
+         pRow[3] = *src;
+         src += 2;
+      }
+      p += dst_stride;
+   }
+}
+
+
+void
+pipe_tile_raw_to_rgba(enum pipe_format format,
+                      void *src,
+                      uint w, uint h,
+                      float *dst, unsigned dst_stride)
+{
+   switch (format) {
+   case PIPE_FORMAT_Z16_UNORM:
+      z16_get_tile_rgba((ushort *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_Z32_UNORM:
+      z32_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+   case PIPE_FORMAT_Z24X8_UNORM:
+      s8z24_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_S8_USCALED:
+      s8_get_tile_rgba((unsigned char *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_X24S8_USCALED:
+      s8x24_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+   case PIPE_FORMAT_X8Z24_UNORM:
+      z24s8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_S8X24_USCALED:
+      x24s8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_Z32_FLOAT:
+      z32f_get_tile_rgba((float *) src, w, h, dst, dst_stride);
+      break;
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      z32f_x24s8_get_tile_rgba((float *) src, w, h, dst, dst_stride);
+      break;
+   default:
+      util_format_read_4f(format,
+                          dst, dst_stride * sizeof(float),
+                          src, util_format_get_stride(format, w),
+                          0, 0, w, h);
+   }
+}
+
+
+void
+pipe_get_tile_rgba(struct pipe_context *pipe,
+                   struct pipe_transfer *pt,
+                   uint x, uint y, uint w, uint h,
+                   float *p)
+{
+   pipe_get_tile_rgba_format(pipe, pt, x, y, w, h, pt->resource->format, p);
+}
+
+
+void
+pipe_get_tile_rgba_format(struct pipe_context *pipe,
+                          struct pipe_transfer *pt,
+                          uint x, uint y, uint w, uint h,
+                          enum pipe_format format,
+                          float *p)
+{
+   unsigned dst_stride = w * 4;
+   void *packed;
+
+   if (u_clip_tile(x, y, &w, &h, &pt->box)) {
+      return;
+   }
+
+   packed = MALLOC(util_format_get_nblocks(format, w, h) * util_format_get_blocksize(format));
+   if (!packed) {
+      return;
+   }
+
+   if (format == PIPE_FORMAT_UYVY || format == PIPE_FORMAT_YUYV) {
+      assert((x & 1) == 0);
+   }
+
+   pipe_get_tile_raw(pipe, pt, x, y, w, h, packed, 0);
+
+   pipe_tile_raw_to_rgba(format, packed, w, h, p, dst_stride);
+
+   FREE(packed);
+}
+
+
+void
+pipe_put_tile_rgba(struct pipe_context *pipe,
+                   struct pipe_transfer *pt,
+                   uint x, uint y, uint w, uint h,
+                   const float *p)
+{
+   pipe_put_tile_rgba_format(pipe, pt, x, y, w, h, pt->resource->format, p);
+}
+
+
+void
+pipe_put_tile_rgba_format(struct pipe_context *pipe,
+                          struct pipe_transfer *pt,
+                          uint x, uint y, uint w, uint h,
+                          enum pipe_format format,
+                          const float *p)
+{
+   unsigned src_stride = w * 4;
+   void *packed;
+
+   if (u_clip_tile(x, y, &w, &h, &pt->box))
+      return;
+
+   packed = MALLOC(util_format_get_nblocks(format, w, h) * util_format_get_blocksize(format));
+
+   if (!packed)
+      return;
+
+   switch (format) {
+   case PIPE_FORMAT_Z16_UNORM:
+      /*z16_put_tile_rgba((ushort *) packed, w, h, p, src_stride);*/
+      break;
+   case PIPE_FORMAT_Z32_UNORM:
+      /*z32_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/
+      break;
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+   case PIPE_FORMAT_Z24X8_UNORM:
+      /*s8z24_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/
+      break;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+   case PIPE_FORMAT_X8Z24_UNORM:
+      /*z24s8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/
+      break;
+   case PIPE_FORMAT_Z32_FLOAT:
+      /*z32f_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/
+      break;
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      /*z32f_s8x24_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/
+      break;
+   default:
+      util_format_write_4f(format,
+                           p, src_stride * sizeof(float),
+                           packed, util_format_get_stride(format, w),
+                           0, 0, w, h);
+   }
+
+   pipe_put_tile_raw(pipe, pt, x, y, w, h, packed, 0);
+
+   FREE(packed);
+}
+
+
+/**
+ * Get a block of Z values, converted to 32-bit range.
+ */
+void
+pipe_get_tile_z(struct pipe_context *pipe,
+                struct pipe_transfer *pt,
+                uint x, uint y, uint w, uint h,
+                uint *z)
+{
+   const uint dstStride = w;
+   ubyte *map;
+   uint *pDest = z;
+   uint i, j;
+   enum pipe_format format = pt->resource->format;
+
+   if (u_clip_tile(x, y, &w, &h, &pt->box))
+      return;
+
+   map = (ubyte *)pipe->transfer_map(pipe, pt);
+   if (!map) {
+      assert(0);
+      return;
+   }
+
+   switch (format) {
+   case PIPE_FORMAT_Z32_UNORM:
+      {
+         const uint *ptrc
+            = (const uint *)(map  + y * pt->stride + x*4);
+         for (i = 0; i < h; i++) {
+            memcpy(pDest, ptrc, 4 * w);
+            pDest += dstStride;
+            ptrc += pt->stride/4;
+         }
+      }
+      break;
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+   case PIPE_FORMAT_Z24X8_UNORM:
+      {
+         const uint *ptrc
+            = (const uint *)(map + y * pt->stride + x*4);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 24-bit Z to 32-bit Z */
+               pDest[j] = (ptrc[j] << 8) | ((ptrc[j] >> 16) & 0xff);
+            }
+            pDest += dstStride;
+            ptrc += pt->stride/4;
+         }
+      }
+      break;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+   case PIPE_FORMAT_X8Z24_UNORM:
+      {
+         const uint *ptrc
+            = (const uint *)(map + y * pt->stride + x*4);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 24-bit Z to 32-bit Z */
+               pDest[j] = (ptrc[j] & 0xffffff00) | ((ptrc[j] >> 24) & 0xff);
+            }
+            pDest += dstStride;
+            ptrc += pt->stride/4;
+         }
+      }
+      break;
+   case PIPE_FORMAT_Z16_UNORM:
+      {
+         const ushort *ptrc
+            = (const ushort *)(map + y * pt->stride + x*2);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 16-bit Z to 32-bit Z */
+               pDest[j] = (ptrc[j] << 16) | ptrc[j];
+            }
+            pDest += dstStride;
+            ptrc += pt->stride/2;
+         }
+      }
+      break;
+   default:
+      assert(0);
+   }
+
+   pipe->transfer_unmap(pipe, pt);
+}
+
+
+void
+pipe_put_tile_z(struct pipe_context *pipe,
+                struct pipe_transfer *pt,
+                uint x, uint y, uint w, uint h,
+                const uint *zSrc)
+{
+   const uint srcStride = w;
+   const uint *ptrc = zSrc;
+   ubyte *map;
+   uint i, j;
+   enum pipe_format format = pt->resource->format;
+
+   if (u_clip_tile(x, y, &w, &h, &pt->box))
+      return;
+
+   map = (ubyte *)pipe->transfer_map(pipe, pt);
+   if (!map) {
+      assert(0);
+      return;
+   }
+
+   switch (format) {
+   case PIPE_FORMAT_Z32_UNORM:
+      {
+         uint *pDest = (uint *) (map + y * pt->stride + x*4);
+         for (i = 0; i < h; i++) {
+            memcpy(pDest, ptrc, 4 * w);
+            pDest += pt->stride/4;
+            ptrc += srcStride;
+         }
+      }
+      break;
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+      {
+         uint *pDest = (uint *) (map + y * pt->stride + x*4);
+         /*assert((pt->usage & PIPE_TRANSFER_READ_WRITE) == PIPE_TRANSFER_READ_WRITE);*/
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 32-bit Z to 24-bit Z, preserve stencil */
+               pDest[j] = (pDest[j] & 0xff000000) | ptrc[j] >> 8;
+            }
+            pDest += pt->stride/4;
+            ptrc += srcStride;
+         }
+      }
+      break;
+   case PIPE_FORMAT_Z24X8_UNORM:
+      {
+         uint *pDest = (uint *) (map + y * pt->stride + x*4);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 32-bit Z to 24-bit Z (0 stencil) */
+               pDest[j] = ptrc[j] >> 8;
+            }
+            pDest += pt->stride/4;
+            ptrc += srcStride;
+         }
+      }
+      break;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+      {
+         uint *pDest = (uint *) (map + y * pt->stride + x*4);
+         /*assert((pt->usage & PIPE_TRANSFER_READ_WRITE) == PIPE_TRANSFER_READ_WRITE);*/
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 32-bit Z to 24-bit Z, preserve stencil */
+               pDest[j] = (pDest[j] & 0xff) | (ptrc[j] & 0xffffff00);
+            }
+            pDest += pt->stride/4;
+            ptrc += srcStride;
+         }
+      }
+      break;
+   case PIPE_FORMAT_X8Z24_UNORM:
+      {
+         uint *pDest = (uint *) (map + y * pt->stride + x*4);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 32-bit Z to 24-bit Z (0 stencil) */
+               pDest[j] = ptrc[j] & 0xffffff00;
+            }
+            pDest += pt->stride/4;
+            ptrc += srcStride;
+         }
+      }
+      break;
+   case PIPE_FORMAT_Z16_UNORM:
+      {
+         ushort *pDest = (ushort *) (map + y * pt->stride + x*2);
+         for (i = 0; i < h; i++) {
+            for (j = 0; j < w; j++) {
+               /* convert 32-bit Z to 16-bit Z */
+               pDest[j] = ptrc[j] >> 16;
+            }
+            pDest += pt->stride/2;
+            ptrc += srcStride;
+         }
+      }
+      break;
+   default:
+      assert(0);
+   }
+
+   pipe->transfer_unmap(pipe, pt);
+}
+
+
diff --git a/mesalib/src/glsl/ir_optimization.h b/mesalib/src/glsl/ir_optimization.h
index dd265673c..59a040751 100644
--- a/mesalib/src/glsl/ir_optimization.h
+++ b/mesalib/src/glsl/ir_optimization.h
@@ -56,10 +56,8 @@ bool do_if_simplification(exec_list *instructions);
 bool do_discard_simplification(exec_list *instructions);
 bool lower_if_to_cond_assign(exec_list *instructions, unsigned max_depth = 0);
 bool do_mat_op_to_vec(exec_list *instructions);
-bool do_mod_to_fract(exec_list *instructions);
 bool do_noop_swizzle(exec_list *instructions);
 bool do_structure_splitting(exec_list *instructions);
-bool do_sub_to_add_neg(exec_list *instructions);
 bool do_swizzle_swizzle(exec_list *instructions);
 bool do_tree_grafting(exec_list *instructions);
 bool do_vec_index_to_cond_assign(exec_list *instructions);
diff --git a/mesalib/src/glsl/ir_reader.cpp b/mesalib/src/glsl/ir_reader.cpp
index 201e436be..f3a621734 100644
--- a/mesalib/src/glsl/ir_reader.cpp
+++ b/mesalib/src/glsl/ir_reader.cpp
@@ -1,1005 +1,1007 @@
-/*

- * 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.

- */

-

-#include "ir_reader.h"

-#include "glsl_parser_extras.h"

-#include "glsl_types.h"

-#include "s_expression.h"

-

-const static bool debug = false;

-

-class ir_reader {

-public:

-   ir_reader(_mesa_glsl_parse_state *);

-

-   void read(exec_list *instructions, const char *src, bool scan_for_protos);

-

-private:

-   void *mem_ctx;

-   _mesa_glsl_parse_state *state;

-

-   void ir_read_error(s_expression *, const char *fmt, ...);

-

-   const glsl_type *read_type(s_expression *);

-

-   void scan_for_prototypes(exec_list *, s_expression *);

-   ir_function *read_function(s_expression *, bool skip_body);

-   void read_function_sig(ir_function *, s_expression *, bool skip_body);

-

-   void read_instructions(exec_list *, s_expression *, ir_loop *);

-   ir_instruction *read_instruction(s_expression *, ir_loop *);

-   ir_variable *read_declaration(s_expression *);

-   ir_if *read_if(s_expression *, ir_loop *);

-   ir_loop *read_loop(s_expression *);

-   ir_return *read_return(s_expression *);

-   ir_rvalue *read_rvalue(s_expression *);

-   ir_assignment *read_assignment(s_expression *);

-   ir_expression *read_expression(s_expression *);

-   ir_call *read_call(s_expression *);

-   ir_swizzle *read_swizzle(s_expression *);

-   ir_constant *read_constant(s_expression *);

-   ir_texture *read_texture(s_expression *);

-

-   ir_dereference *read_dereference(s_expression *);

-};

-

-ir_reader::ir_reader(_mesa_glsl_parse_state *state) : state(state)

-{

-   this->mem_ctx = state;

-}

-

-void

-_mesa_glsl_read_ir(_mesa_glsl_parse_state *state, exec_list *instructions,

-		   const char *src, bool scan_for_protos)

-{

-   ir_reader r(state);

-   r.read(instructions, src, scan_for_protos);

-}

-

-void

-ir_reader::read(exec_list *instructions, const char *src, bool scan_for_protos)

-{

-   s_expression *expr = s_expression::read_expression(mem_ctx, src);

-   if (expr == NULL) {

-      ir_read_error(NULL, "couldn't parse S-Expression.");

-      return;

-   }

-   

-   if (scan_for_protos) {

-      scan_for_prototypes(instructions, expr);

-      if (state->error)

-	 return;

-   }

-

-   read_instructions(instructions, expr, NULL);

-   ralloc_free(expr);

-

-   if (debug)

-      validate_ir_tree(instructions);

-}

-

-void

-ir_reader::ir_read_error(s_expression *expr, const char *fmt, ...)

-{

-   va_list ap;

-

-   state->error = true;

-

-   if (state->current_function != NULL)

-      ralloc_asprintf_append(&state->info_log, "In function %s:\n",

-			     state->current_function->function_name());

-   ralloc_strcat(&state->info_log, "error: ");

-

-   va_start(ap, fmt);

-   ralloc_vasprintf_append(&state->info_log, fmt, ap);

-   va_end(ap);

-   ralloc_strcat(&state->info_log, "\n");

-

-   if (expr != NULL) {

-      ralloc_strcat(&state->info_log, "...in this context:\n   ");

-      expr->print();

-      ralloc_strcat(&state->info_log, "\n\n");

-   }

-}

-

-const glsl_type *

-ir_reader::read_type(s_expression *expr)

-{

-   s_expression *s_base_type;

-   s_int *s_size;

-

-   s_pattern pat[] = { "array", s_base_type, s_size };

-   if (MATCH(expr, pat)) {

-      const glsl_type *base_type = read_type(s_base_type);

-      if (base_type == NULL) {

-	 ir_read_error(NULL, "when reading base type of array type");

-	 return NULL;

-      }

-

-      return glsl_type::get_array_instance(base_type, s_size->value());

-   }

-   

-   s_symbol *type_sym = SX_AS_SYMBOL(expr);

-   if (type_sym == NULL) {

-      ir_read_error(expr, "expected <type>");

-      return NULL;

-   }

-

-   const glsl_type *type = state->symbols->get_type(type_sym->value());

-   if (type == NULL)

-      ir_read_error(expr, "invalid type: %s", type_sym->value());

-

-   return type;

-}

-

-

-void

-ir_reader::scan_for_prototypes(exec_list *instructions, s_expression *expr)

-{

-   s_list *list = SX_AS_LIST(expr);

-   if (list == NULL) {

-      ir_read_error(expr, "Expected (<instruction> ...); found an atom.");

-      return;

-   }

-

-   foreach_iter(exec_list_iterator, it, list->subexpressions) {

-      s_list *sub = SX_AS_LIST(it.get());

-      if (sub == NULL)

-	 continue; // not a (function ...); ignore it.

-

-      s_symbol *tag = SX_AS_SYMBOL(sub->subexpressions.get_head());

-      if (tag == NULL || strcmp(tag->value(), "function") != 0)

-	 continue; // not a (function ...); ignore it.

-

-      ir_function *f = read_function(sub, true);

-      if (f == NULL)

-	 return;

-      instructions->push_tail(f);

-   }

-}

-

-ir_function *

-ir_reader::read_function(s_expression *expr, bool skip_body)

-{

-   bool added = false;

-   s_symbol *name;

-

-   s_pattern pat[] = { "function", name };

-   if (!PARTIAL_MATCH(expr, pat)) {

-      ir_read_error(expr, "Expected (function <name> (signature ...) ...)");

-      return NULL;

-   }

-

-   ir_function *f = state->symbols->get_function(name->value());

-   if (f == NULL) {

-      f = new(mem_ctx) ir_function(name->value());

-      added = state->symbols->add_function(f);

-      assert(added);

-   }

-

-   exec_list_iterator it = ((s_list *) expr)->subexpressions.iterator();

-   it.next(); // skip "function" tag

-   it.next(); // skip function name

-   for (/* nothing */; it.has_next(); it.next()) {

-      s_expression *s_sig = (s_expression *) it.get();

-      read_function_sig(f, s_sig, skip_body);

-   }

-   return added ? f : NULL;

-}

-

-void

-ir_reader::read_function_sig(ir_function *f, s_expression *expr, bool skip_body)

-{

-   s_expression *type_expr;

-   s_list *paramlist;

-   s_list *body_list;

-

-   s_pattern pat[] = { "signature", type_expr, paramlist, body_list };

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "Expected (signature <type> (parameters ...) "

-			  "(<instruction> ...))");

-      return;

-   }

-

-   const glsl_type *return_type = read_type(type_expr);

-   if (return_type == NULL)

-      return;

-

-   s_symbol *paramtag = SX_AS_SYMBOL(paramlist->subexpressions.get_head());

-   if (paramtag == NULL || strcmp(paramtag->value(), "parameters") != 0) {

-      ir_read_error(paramlist, "Expected (parameters ...)");

-      return;

-   }

-

-   // Read the parameters list into a temporary place.

-   exec_list hir_parameters;

-   state->symbols->push_scope();

-

-   exec_list_iterator it = paramlist->subexpressions.iterator();

-   for (it.next() /* skip "parameters" */; it.has_next(); it.next()) {

-      ir_variable *var = read_declaration((s_expression *) it.get());

-      if (var == NULL)

-	 return;

-

-      hir_parameters.push_tail(var);

-   }

-

-   ir_function_signature *sig = f->exact_matching_signature(&hir_parameters);

-   if (sig == NULL && skip_body) {

-      /* If scanning for prototypes, generate a new signature. */

-      sig = new(mem_ctx) ir_function_signature(return_type);

-      sig->is_builtin = true;

-      f->add_signature(sig);

-   } else if (sig != NULL) {

-      const char *badvar = sig->qualifiers_match(&hir_parameters);

-      if (badvar != NULL) {

-	 ir_read_error(expr, "function `%s' parameter `%s' qualifiers "

-		       "don't match prototype", f->name, badvar);

-	 return;

-      }

-

-      if (sig->return_type != return_type) {

-	 ir_read_error(expr, "function `%s' return type doesn't "

-		       "match prototype", f->name);

-	 return;

-      }

-   } else {

-      /* No prototype for this body exists - skip it. */

-      state->symbols->pop_scope();

-      return;

-   }

-   assert(sig != NULL);

-

-   sig->replace_parameters(&hir_parameters);

-

-   if (!skip_body && !body_list->subexpressions.is_empty()) {

-      if (sig->is_defined) {

-	 ir_read_error(expr, "function %s redefined", f->name);

-	 return;

-      }

-      state->current_function = sig;

-      read_instructions(&sig->body, body_list, NULL);

-      state->current_function = NULL;

-      sig->is_defined = true;

-   }

-

-   state->symbols->pop_scope();

-}

-

-void

-ir_reader::read_instructions(exec_list *instructions, s_expression *expr,

-			     ir_loop *loop_ctx)

-{

-   // Read in a list of instructions

-   s_list *list = SX_AS_LIST(expr);

-   if (list == NULL) {

-      ir_read_error(expr, "Expected (<instruction> ...); found an atom.");

-      return;

-   }

-

-   foreach_iter(exec_list_iterator, it, list->subexpressions) {

-      s_expression *sub = (s_expression*) it.get();

-      ir_instruction *ir = read_instruction(sub, loop_ctx);

-      if (ir != NULL) {

-	 /* Global variable declarations should be moved to the top, before

-	  * any functions that might use them.  Functions are added to the

-	  * instruction stream when scanning for prototypes, so without this

-	  * hack, they always appear before variable declarations.

-	  */

-	 if (state->current_function == NULL && ir->as_variable() != NULL)

-	    instructions->push_head(ir);

-	 else

-	    instructions->push_tail(ir);

-      }

-   }

-}

-

-

-ir_instruction *

-ir_reader::read_instruction(s_expression *expr, ir_loop *loop_ctx)

-{

-   s_symbol *symbol = SX_AS_SYMBOL(expr);

-   if (symbol != NULL) {

-      if (strcmp(symbol->value(), "break") == 0 && loop_ctx != NULL)

-	 return new(mem_ctx) ir_loop_jump(ir_loop_jump::jump_break);

-      if (strcmp(symbol->value(), "continue") == 0 && loop_ctx != NULL)

-	 return new(mem_ctx) ir_loop_jump(ir_loop_jump::jump_continue);

-   }

-

-   s_list *list = SX_AS_LIST(expr);

-   if (list == NULL || list->subexpressions.is_empty()) {

-      ir_read_error(expr, "Invalid instruction.\n");

-      return NULL;

-   }

-

-   s_symbol *tag = SX_AS_SYMBOL(list->subexpressions.get_head());

-   if (tag == NULL) {

-      ir_read_error(expr, "expected instruction tag");

-      return NULL;

-   }

-

-   ir_instruction *inst = NULL;

-   if (strcmp(tag->value(), "declare") == 0) {

-      inst = read_declaration(list);

-   } else if (strcmp(tag->value(), "assign") == 0) {

-      inst = read_assignment(list);

-   } else if (strcmp(tag->value(), "if") == 0) {

-      inst = read_if(list, loop_ctx);

-   } else if (strcmp(tag->value(), "loop") == 0) {

-      inst = read_loop(list);

-   } else if (strcmp(tag->value(), "return") == 0) {

-      inst = read_return(list);

-   } else if (strcmp(tag->value(), "function") == 0) {

-      inst = read_function(list, false);

-   } else {

-      inst = read_rvalue(list);

-      if (inst == NULL)

-	 ir_read_error(NULL, "when reading instruction");

-   }

-   return inst;

-}

-

-ir_variable *

-ir_reader::read_declaration(s_expression *expr)

-{

-   s_list *s_quals;

-   s_expression *s_type;

-   s_symbol *s_name;

-

-   s_pattern pat[] = { "declare", s_quals, s_type, s_name };

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (declare (<qualifiers>) <type> <name>)");

-      return NULL;

-   }

-

-   const glsl_type *type = read_type(s_type);

-   if (type == NULL)

-      return NULL;

-

-   ir_variable *var = new(mem_ctx) ir_variable(type, s_name->value(),

-					       ir_var_auto);

-

-   foreach_iter(exec_list_iterator, it, s_quals->subexpressions) {

-      s_symbol *qualifier = SX_AS_SYMBOL(it.get());

-      if (qualifier == NULL) {

-	 ir_read_error(expr, "qualifier list must contain only symbols");

-	 return NULL;

-      }

-

-      // FINISHME: Check for duplicate/conflicting qualifiers.

-      if (strcmp(qualifier->value(), "centroid") == 0) {

-	 var->centroid = 1;

-      } else if (strcmp(qualifier->value(), "invariant") == 0) {

-	 var->invariant = 1;

-      } else if (strcmp(qualifier->value(), "uniform") == 0) {

-	 var->mode = ir_var_uniform;

-      } else if (strcmp(qualifier->value(), "auto") == 0) {

-	 var->mode = ir_var_auto;

-      } else if (strcmp(qualifier->value(), "in") == 0) {

-	 var->mode = ir_var_in;

-      } else if (strcmp(qualifier->value(), "const_in") == 0) {

-	 var->mode = ir_var_const_in;

-      } else if (strcmp(qualifier->value(), "out") == 0) {

-	 var->mode = ir_var_out;

-      } else if (strcmp(qualifier->value(), "inout") == 0) {

-	 var->mode = ir_var_inout;

-      } else if (strcmp(qualifier->value(), "smooth") == 0) {

-	 var->interpolation = ir_var_smooth;

-      } else if (strcmp(qualifier->value(), "flat") == 0) {

-	 var->interpolation = ir_var_flat;

-      } else if (strcmp(qualifier->value(), "noperspective") == 0) {

-	 var->interpolation = ir_var_noperspective;

-      } else {

-	 ir_read_error(expr, "unknown qualifier: %s", qualifier->value());

-	 return NULL;

-      }

-   }

-

-   // Add the variable to the symbol table

-   state->symbols->add_variable(var);

-

-   return var;

-}

-

-

-ir_if *

-ir_reader::read_if(s_expression *expr, ir_loop *loop_ctx)

-{

-   s_expression *s_cond;

-   s_expression *s_then;

-   s_expression *s_else;

-

-   s_pattern pat[] = { "if", s_cond, s_then, s_else };

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (if <condition> (<then>...) (<else>...))");

-      return NULL;

-   }

-

-   ir_rvalue *condition = read_rvalue(s_cond);

-   if (condition == NULL) {

-      ir_read_error(NULL, "when reading condition of (if ...)");

-      return NULL;

-   }

-

-   ir_if *iff = new(mem_ctx) ir_if(condition);

-

-   read_instructions(&iff->then_instructions, s_then, loop_ctx);

-   read_instructions(&iff->else_instructions, s_else, loop_ctx);

-   if (state->error) {

-      delete iff;

-      iff = NULL;

-   }

-   return iff;

-}

-

-

-ir_loop *

-ir_reader::read_loop(s_expression *expr)

-{

-   s_expression *s_counter, *s_from, *s_to, *s_inc, *s_body;

-

-   s_pattern pat[] = { "loop", s_counter, s_from, s_to, s_inc, s_body };

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (loop <counter> <from> <to> "

-			  "<increment> <body>)");

-      return NULL;

-   }

-

-   // FINISHME: actually read the count/from/to fields.

-

-   ir_loop *loop = new(mem_ctx) ir_loop;

-   read_instructions(&loop->body_instructions, s_body, loop);

-   if (state->error) {

-      delete loop;

-      loop = NULL;

-   }

-   return loop;

-}

-

-

-ir_return *

-ir_reader::read_return(s_expression *expr)

-{

-   s_expression *s_retval;

-

-   s_pattern pat[] = { "return", s_retval};

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (return <rvalue>)");

-      return NULL;

-   }

-

-   ir_rvalue *retval = read_rvalue(s_retval);

-   if (retval == NULL) {

-      ir_read_error(NULL, "when reading return value");

-      return NULL;

-   }

-

-   return new(mem_ctx) ir_return(retval);

-}

-

-

-ir_rvalue *

-ir_reader::read_rvalue(s_expression *expr)

-{

-   s_list *list = SX_AS_LIST(expr);

-   if (list == NULL || list->subexpressions.is_empty())

-      return NULL;

-

-   s_symbol *tag = SX_AS_SYMBOL(list->subexpressions.get_head());

-   if (tag == NULL) {

-      ir_read_error(expr, "expected rvalue tag");

-      return NULL;

-   }

-

-   ir_rvalue *rvalue = read_dereference(list);

-   if (rvalue != NULL || state->error)

-      return rvalue;

-   else if (strcmp(tag->value(), "swiz") == 0) {

-      rvalue = read_swizzle(list);

-   } else if (strcmp(tag->value(), "expression") == 0) {

-      rvalue = read_expression(list);

-   } else if (strcmp(tag->value(), "call") == 0) {

-      rvalue = read_call(list);

-   } else if (strcmp(tag->value(), "constant") == 0) {

-      rvalue = read_constant(list);

-   } else {

-      rvalue = read_texture(list);

-      if (rvalue == NULL && !state->error)

-	 ir_read_error(expr, "unrecognized rvalue tag: %s", tag->value());

-   }

-

-   return rvalue;

-}

-

-ir_assignment *

-ir_reader::read_assignment(s_expression *expr)

-{

-   s_expression *cond_expr = NULL;

-   s_expression *lhs_expr, *rhs_expr;

-   s_list       *mask_list;

-

-   s_pattern pat4[] = { "assign",            mask_list, lhs_expr, rhs_expr };

-   s_pattern pat5[] = { "assign", cond_expr, mask_list, lhs_expr, rhs_expr };

-   if (!MATCH(expr, pat4) && !MATCH(expr, pat5)) {

-      ir_read_error(expr, "expected (assign [<condition>] (<write mask>) "

-			  "<lhs> <rhs>)");

-      return NULL;

-   }

-

-   ir_rvalue *condition = NULL;

-   if (cond_expr != NULL) {

-      condition = read_rvalue(cond_expr);

-      if (condition == NULL) {

-	 ir_read_error(NULL, "when reading condition of assignment");

-	 return NULL;

-      }

-   }

-

-   unsigned mask = 0;

-

-   s_symbol *mask_symbol;

-   s_pattern mask_pat[] = { mask_symbol };

-   if (MATCH(mask_list, mask_pat)) {

-      const char *mask_str = mask_symbol->value();

-      unsigned mask_length = strlen(mask_str);

-      if (mask_length > 4) {

-	 ir_read_error(expr, "invalid write mask: %s", mask_str);

-	 return NULL;

-      }

-

-      const unsigned idx_map[] = { 3, 0, 1, 2 }; /* w=bit 3, x=0, y=1, z=2 */

-

-      for (unsigned i = 0; i < mask_length; i++) {

-	 if (mask_str[i] < 'w' || mask_str[i] > 'z') {

-	    ir_read_error(expr, "write mask contains invalid character: %c",

-			  mask_str[i]);

-	    return NULL;

-	 }

-	 mask |= 1 << idx_map[mask_str[i] - 'w'];

-      }

-   } else if (!mask_list->subexpressions.is_empty()) {

-      ir_read_error(mask_list, "expected () or (<write mask>)");

-      return NULL;

-   }

-

-   ir_dereference *lhs = read_dereference(lhs_expr);

-   if (lhs == NULL) {

-      ir_read_error(NULL, "when reading left-hand side of assignment");

-      return NULL;

-   }

-

-   ir_rvalue *rhs = read_rvalue(rhs_expr);

-   if (rhs == NULL) {

-      ir_read_error(NULL, "when reading right-hand side of assignment");

-      return NULL;

-   }

-

-   if (mask == 0 && (lhs->type->is_vector() || lhs->type->is_scalar())) {

-      ir_read_error(expr, "non-zero write mask required.");

-      return NULL;

-   }

-

-   return new(mem_ctx) ir_assignment(lhs, rhs, condition, mask);

-}

-

-ir_call *

-ir_reader::read_call(s_expression *expr)

-{

-   s_symbol *name;

-   s_list *params;

-

-   s_pattern pat[] = { "call", name, params };

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (call <name> (<param> ...))");

-      return NULL;

-   }

-

-   exec_list parameters;

-

-   foreach_iter(exec_list_iterator, it, params->subexpressions) {

-      s_expression *expr = (s_expression*) it.get();

-      ir_rvalue *param = read_rvalue(expr);

-      if (param == NULL) {

-	 ir_read_error(expr, "when reading parameter to function call");

-	 return NULL;

-      }

-      parameters.push_tail(param);

-   }

-

-   ir_function *f = state->symbols->get_function(name->value());

-   if (f == NULL) {

-      ir_read_error(expr, "found call to undefined function %s",

-		    name->value());

-      return NULL;

-   }

-

-   ir_function_signature *callee = f->matching_signature(&parameters);

-   if (callee == NULL) {

-      ir_read_error(expr, "couldn't find matching signature for function "

-                    "%s", name->value());

-      return NULL;

-   }

-

-   return new(mem_ctx) ir_call(callee, &parameters);

-}

-

-ir_expression *

-ir_reader::read_expression(s_expression *expr)

-{

-   s_expression *s_type;

-   s_symbol *s_op;

-   s_expression *s_arg1;

-

-   s_pattern pat[] = { "expression", s_type, s_op, s_arg1 };

-   if (!PARTIAL_MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (expression <type> <operator> "

-			  "<operand> [<operand>])");

-      return NULL;

-   }

-   s_expression *s_arg2 = (s_expression *) s_arg1->next; // may be tail sentinel

-

-   const glsl_type *type = read_type(s_type);

-   if (type == NULL)

-      return NULL;

-

-   /* Read the operator */

-   ir_expression_operation op = ir_expression::get_operator(s_op->value());

-   if (op == (ir_expression_operation) -1) {

-      ir_read_error(expr, "invalid operator: %s", s_op->value());

-      return NULL;

-   }

-    

-   unsigned num_operands = ir_expression::get_num_operands(op);

-   if (num_operands == 1 && !s_arg1->next->is_tail_sentinel()) {

-      ir_read_error(expr, "expected (expression <type> %s <operand>)",

-		    s_op->value());

-      return NULL;

-   }

-

-   ir_rvalue *arg1 = read_rvalue(s_arg1);

-   ir_rvalue *arg2 = NULL;

-   if (arg1 == NULL) {

-      ir_read_error(NULL, "when reading first operand of %s", s_op->value());

-      return NULL;

-   }

-

-   if (num_operands == 2) {

-      if (s_arg2->is_tail_sentinel() || !s_arg2->next->is_tail_sentinel()) {

-	 ir_read_error(expr, "expected (expression <type> %s <operand> "

-			     "<operand>)", s_op->value());

-	 return NULL;

-      }

-      arg2 = read_rvalue(s_arg2);

-      if (arg2 == NULL) {

-	 ir_read_error(NULL, "when reading second operand of %s",

-		       s_op->value());

-	 return NULL;

-      }

-   }

-

-   return new(mem_ctx) ir_expression(op, type, arg1, arg2);

-}

-

-ir_swizzle *

-ir_reader::read_swizzle(s_expression *expr)

-{

-   s_symbol *swiz;

-   s_expression *sub;

-

-   s_pattern pat[] = { "swiz", swiz, sub };

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (swiz <swizzle> <rvalue>)");

-      return NULL;

-   }

-

-   if (strlen(swiz->value()) > 4) {

-      ir_read_error(expr, "expected a valid swizzle; found %s", swiz->value());

-      return NULL;

-   }

-

-   ir_rvalue *rvalue = read_rvalue(sub);

-   if (rvalue == NULL)

-      return NULL;

-

-   ir_swizzle *ir = ir_swizzle::create(rvalue, swiz->value(),

-				       rvalue->type->vector_elements);

-   if (ir == NULL)

-      ir_read_error(expr, "invalid swizzle");

-

-   return ir;

-}

-

-ir_constant *

-ir_reader::read_constant(s_expression *expr)

-{

-   s_expression *type_expr;

-   s_list *values;

-

-   s_pattern pat[] = { "constant", type_expr, values };

-   if (!MATCH(expr, pat)) {

-      ir_read_error(expr, "expected (constant <type> (...))");

-      return NULL;

-   }

-

-   const glsl_type *type = read_type(type_expr);

-   if (type == NULL)

-      return NULL;

-

-   if (values == NULL) {

-      ir_read_error(expr, "expected (constant <type> (...))");

-      return NULL;

-   }

-

-   if (type->is_array()) {

-      unsigned elements_supplied = 0;

-      exec_list elements;

-      foreach_iter(exec_list_iterator, it, values->subexpressions) {

-	 s_expression *elt = (s_expression *) it.get();

-	 ir_constant *ir_elt = read_constant(elt);

-	 if (ir_elt == NULL)

-	    return NULL;

-	 elements.push_tail(ir_elt);

-	 elements_supplied++;

-      }

-

-      if (elements_supplied != type->length) {

-	 ir_read_error(values, "expected exactly %u array elements, "

-		       "given %u", type->length, elements_supplied);

-	 return NULL;

-      }

-      return new(mem_ctx) ir_constant(type, &elements);

-   }

-

-   const glsl_type *const base_type = type->get_base_type();

-

-   ir_constant_data data = { { 0 } };

-

-   // Read in list of values (at most 16).

-   int k = 0;

-   foreach_iter(exec_list_iterator, it, values->subexpressions) {

-      if (k >= 16) {

-	 ir_read_error(values, "expected at most 16 numbers");

-	 return NULL;

-      }

-

-      s_expression *expr = (s_expression*) it.get();

-

-      if (base_type->base_type == GLSL_TYPE_FLOAT) {

-	 s_number *value = SX_AS_NUMBER(expr);

-	 if (value == NULL) {

-	    ir_read_error(values, "expected numbers");

-	    return NULL;

-	 }

-	 data.f[k] = value->fvalue();

-      } else {

-	 s_int *value = SX_AS_INT(expr);

-	 if (value == NULL) {

-	    ir_read_error(values, "expected integers");

-	    return NULL;

-	 }

-

-	 switch (base_type->base_type) {

-	 case GLSL_TYPE_UINT: {

-	    data.u[k] = value->value();

-	    break;

-	 }

-	 case GLSL_TYPE_INT: {

-	    data.i[k] = value->value();

-	    break;

-	 }

-	 case GLSL_TYPE_BOOL: {

-	    data.b[k] = value->value();

-	    break;

-	 }

-	 default:

-	    ir_read_error(values, "unsupported constant type");

-	    return NULL;

-	 }

-      }

-      ++k;

-   }

-

-   return new(mem_ctx) ir_constant(type, &data);

-}

-

-ir_dereference *

-ir_reader::read_dereference(s_expression *expr)

-{

-   s_symbol *s_var;

-   s_expression *s_subject;

-   s_expression *s_index;

-   s_symbol *s_field;

-

-   s_pattern var_pat[] = { "var_ref", s_var };

-   s_pattern array_pat[] = { "array_ref", s_subject, s_index };

-   s_pattern record_pat[] = { "record_ref", s_subject, s_field };

-

-   if (MATCH(expr, var_pat)) {

-      ir_variable *var = state->symbols->get_variable(s_var->value());

-      if (var == NULL) {

-	 ir_read_error(expr, "undeclared variable: %s", s_var->value());

-	 return NULL;

-      }

-      return new(mem_ctx) ir_dereference_variable(var);

-   } else if (MATCH(expr, array_pat)) {

-      ir_rvalue *subject = read_rvalue(s_subject);

-      if (subject == NULL) {

-	 ir_read_error(NULL, "when reading the subject of an array_ref");

-	 return NULL;

-      }

-

-      ir_rvalue *idx = read_rvalue(s_index);

-      if (subject == NULL) {

-	 ir_read_error(NULL, "when reading the index of an array_ref");

-	 return NULL;

-      }

-      return new(mem_ctx) ir_dereference_array(subject, idx);

-   } else if (MATCH(expr, record_pat)) {

-      ir_rvalue *subject = read_rvalue(s_subject);

-      if (subject == NULL) {

-	 ir_read_error(NULL, "when reading the subject of a record_ref");

-	 return NULL;

-      }

-      return new(mem_ctx) ir_dereference_record(subject, s_field->value());

-   }

-   return NULL;

-}

-

-ir_texture *

-ir_reader::read_texture(s_expression *expr)

-{

-   s_symbol *tag = NULL;

-   s_expression *s_type = NULL;

-   s_expression *s_sampler = NULL;

-   s_expression *s_coord = NULL;

-   s_expression *s_offset = NULL;

-   s_expression *s_proj = NULL;

-   s_list *s_shadow = NULL;

-   s_expression *s_lod = NULL;

-

-   ir_texture_opcode op = ir_tex; /* silence warning */

-

-   s_pattern tex_pattern[] =

-      { "tex", s_type, s_sampler, s_coord, s_offset, s_proj, s_shadow };

-   s_pattern txf_pattern[] =

-      { "txf", s_type, s_sampler, s_coord, s_offset, s_lod };

-   s_pattern other_pattern[] =

-      { tag, s_type, s_sampler, s_coord, s_offset, s_proj, s_shadow, s_lod };

-

-   if (MATCH(expr, tex_pattern)) {

-      op = ir_tex;

-   } else if (MATCH(expr, txf_pattern)) {

-      op = ir_txf;

-   } else if (MATCH(expr, other_pattern)) {

-      op = ir_texture::get_opcode(tag->value());

-      if (op == -1)

-	 return NULL;

-   } else {

-      ir_read_error(NULL, "unexpected texture pattern");

-      return NULL;

-   }

-

-   ir_texture *tex = new(mem_ctx) ir_texture(op);

-

-   // Read return type

-   const glsl_type *type = read_type(s_type);

-   if (type == NULL) {

-      ir_read_error(NULL, "when reading type in (%s ...)",

-		    tex->opcode_string());

-      return NULL;

-   }

-

-   // Read sampler (must be a deref)

-   ir_dereference *sampler = read_dereference(s_sampler);

-   if (sampler == NULL) {

-      ir_read_error(NULL, "when reading sampler in (%s ...)",

-		    tex->opcode_string());

-      return NULL;

-   }

-   tex->set_sampler(sampler, type);

-

-   // Read coordinate (any rvalue)

-   tex->coordinate = read_rvalue(s_coord);

-   if (tex->coordinate == NULL) {

-      ir_read_error(NULL, "when reading coordinate in (%s ...)",

-		    tex->opcode_string());

-      return NULL;

-   }

-

-   // Read texel offset - either 0 or an rvalue.

-   s_int *si_offset = SX_AS_INT(s_offset);

-   if (si_offset == NULL || si_offset->value() != 0) {

-      tex->offset = read_rvalue(s_offset);

-      if (tex->offset == NULL) {

-	 ir_read_error(s_offset, "expected 0 or an expression");

-	 return NULL;

-      }

-   }

-

-   if (op != ir_txf) {

-      s_int *proj_as_int = SX_AS_INT(s_proj);

-      if (proj_as_int && proj_as_int->value() == 1) {

-	 tex->projector = NULL;

-      } else {

-	 tex->projector = read_rvalue(s_proj);

-	 if (tex->projector == NULL) {

-	    ir_read_error(NULL, "when reading projective divide in (%s ..)",

-	                  tex->opcode_string());

-	    return NULL;

-	 }

-      }

-

-      if (s_shadow->subexpressions.is_empty()) {

-	 tex->shadow_comparitor = NULL;

-      } else {

-	 tex->shadow_comparitor = read_rvalue(s_shadow);

-	 if (tex->shadow_comparitor == NULL) {

-	    ir_read_error(NULL, "when reading shadow comparitor in (%s ..)",

-			  tex->opcode_string());

-	    return NULL;

-	 }

-      }

-   }

-

-   switch (op) {

-   case ir_txb:

-      tex->lod_info.bias = read_rvalue(s_lod);

-      if (tex->lod_info.bias == NULL) {

-	 ir_read_error(NULL, "when reading LOD bias in (txb ...)");

-	 return NULL;

-      }

-      break;

-   case ir_txl:

-   case ir_txf:

-      tex->lod_info.lod = read_rvalue(s_lod);

-      if (tex->lod_info.lod == NULL) {

-	 ir_read_error(NULL, "when reading LOD in (%s ...)",

-		       tex->opcode_string());

-	 return NULL;

-      }

-      break;

-   case ir_txd: {

-      s_expression *s_dx, *s_dy;

-      s_pattern dxdy_pat[] = { s_dx, s_dy };

-      if (!MATCH(s_lod, dxdy_pat)) {

-	 ir_read_error(s_lod, "expected (dPdx dPdy) in (txd ...)");

-	 return NULL;

-      }

-      tex->lod_info.grad.dPdx = read_rvalue(s_dx);

-      if (tex->lod_info.grad.dPdx == NULL) {

-	 ir_read_error(NULL, "when reading dPdx in (txd ...)");

-	 return NULL;

-      }

-      tex->lod_info.grad.dPdy = read_rvalue(s_dy);

-      if (tex->lod_info.grad.dPdy == NULL) {

-	 ir_read_error(NULL, "when reading dPdy in (txd ...)");

-	 return NULL;

-      }

-      break;

-   }

-   default:

-      // tex doesn't have any extra parameters.

-      break;

-   };

-   return tex;

-}

+/*
+ * 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.
+ */
+
+#include "ir_reader.h"
+#include "glsl_parser_extras.h"
+#include "glsl_types.h"
+#include "s_expression.h"
+
+const static bool debug = false;
+
+class ir_reader {
+public:
+   ir_reader(_mesa_glsl_parse_state *);
+
+   void read(exec_list *instructions, const char *src, bool scan_for_protos);
+
+private:
+   void *mem_ctx;
+   _mesa_glsl_parse_state *state;
+
+   void ir_read_error(s_expression *, const char *fmt, ...);
+
+   const glsl_type *read_type(s_expression *);
+
+   void scan_for_prototypes(exec_list *, s_expression *);
+   ir_function *read_function(s_expression *, bool skip_body);
+   void read_function_sig(ir_function *, s_expression *, bool skip_body);
+
+   void read_instructions(exec_list *, s_expression *, ir_loop *);
+   ir_instruction *read_instruction(s_expression *, ir_loop *);
+   ir_variable *read_declaration(s_expression *);
+   ir_if *read_if(s_expression *, ir_loop *);
+   ir_loop *read_loop(s_expression *);
+   ir_return *read_return(s_expression *);
+   ir_rvalue *read_rvalue(s_expression *);
+   ir_assignment *read_assignment(s_expression *);
+   ir_expression *read_expression(s_expression *);
+   ir_call *read_call(s_expression *);
+   ir_swizzle *read_swizzle(s_expression *);
+   ir_constant *read_constant(s_expression *);
+   ir_texture *read_texture(s_expression *);
+
+   ir_dereference *read_dereference(s_expression *);
+};
+
+ir_reader::ir_reader(_mesa_glsl_parse_state *state) : state(state)
+{
+   this->mem_ctx = state;
+}
+
+void
+_mesa_glsl_read_ir(_mesa_glsl_parse_state *state, exec_list *instructions,
+		   const char *src, bool scan_for_protos)
+{
+   ir_reader r(state);
+   r.read(instructions, src, scan_for_protos);
+}
+
+void
+ir_reader::read(exec_list *instructions, const char *src, bool scan_for_protos)
+{
+   s_expression *expr = s_expression::read_expression(mem_ctx, src);
+   if (expr == NULL) {
+      ir_read_error(NULL, "couldn't parse S-Expression.");
+      return;
+   }
+   
+   if (scan_for_protos) {
+      scan_for_prototypes(instructions, expr);
+      if (state->error)
+	 return;
+   }
+
+   read_instructions(instructions, expr, NULL);
+   ralloc_free(expr);
+
+   if (debug)
+      validate_ir_tree(instructions);
+}
+
+void
+ir_reader::ir_read_error(s_expression *expr, const char *fmt, ...)
+{
+   va_list ap;
+
+   state->error = true;
+
+   if (state->current_function != NULL)
+      ralloc_asprintf_append(&state->info_log, "In function %s:\n",
+			     state->current_function->function_name());
+   ralloc_strcat(&state->info_log, "error: ");
+
+   va_start(ap, fmt);
+   ralloc_vasprintf_append(&state->info_log, fmt, ap);
+   va_end(ap);
+   ralloc_strcat(&state->info_log, "\n");
+
+   if (expr != NULL) {
+      ralloc_strcat(&state->info_log, "...in this context:\n   ");
+      expr->print();
+      ralloc_strcat(&state->info_log, "\n\n");
+   }
+}
+
+const glsl_type *
+ir_reader::read_type(s_expression *expr)
+{
+   s_expression *s_base_type;
+   s_int *s_size;
+
+   s_pattern pat[] = { "array", s_base_type, s_size };
+   if (MATCH(expr, pat)) {
+      const glsl_type *base_type = read_type(s_base_type);
+      if (base_type == NULL) {
+	 ir_read_error(NULL, "when reading base type of array type");
+	 return NULL;
+      }
+
+      return glsl_type::get_array_instance(base_type, s_size->value());
+   }
+   
+   s_symbol *type_sym = SX_AS_SYMBOL(expr);
+   if (type_sym == NULL) {
+      ir_read_error(expr, "expected <type>");
+      return NULL;
+   }
+
+   const glsl_type *type = state->symbols->get_type(type_sym->value());
+   if (type == NULL)
+      ir_read_error(expr, "invalid type: %s", type_sym->value());
+
+   return type;
+}
+
+
+void
+ir_reader::scan_for_prototypes(exec_list *instructions, s_expression *expr)
+{
+   s_list *list = SX_AS_LIST(expr);
+   if (list == NULL) {
+      ir_read_error(expr, "Expected (<instruction> ...); found an atom.");
+      return;
+   }
+
+   foreach_iter(exec_list_iterator, it, list->subexpressions) {
+      s_list *sub = SX_AS_LIST(it.get());
+      if (sub == NULL)
+	 continue; // not a (function ...); ignore it.
+
+      s_symbol *tag = SX_AS_SYMBOL(sub->subexpressions.get_head());
+      if (tag == NULL || strcmp(tag->value(), "function") != 0)
+	 continue; // not a (function ...); ignore it.
+
+      ir_function *f = read_function(sub, true);
+      if (f == NULL)
+	 return;
+      instructions->push_tail(f);
+   }
+}
+
+ir_function *
+ir_reader::read_function(s_expression *expr, bool skip_body)
+{
+   bool added = false;
+   s_symbol *name;
+
+   s_pattern pat[] = { "function", name };
+   if (!PARTIAL_MATCH(expr, pat)) {
+      ir_read_error(expr, "Expected (function <name> (signature ...) ...)");
+      return NULL;
+   }
+
+   ir_function *f = state->symbols->get_function(name->value());
+   if (f == NULL) {
+      f = new(mem_ctx) ir_function(name->value());
+      added = state->symbols->add_function(f);
+      assert(added);
+   }
+
+   exec_list_iterator it = ((s_list *) expr)->subexpressions.iterator();
+   it.next(); // skip "function" tag
+   it.next(); // skip function name
+   for (/* nothing */; it.has_next(); it.next()) {
+      s_expression *s_sig = (s_expression *) it.get();
+      read_function_sig(f, s_sig, skip_body);
+   }
+   return added ? f : NULL;
+}
+
+void
+ir_reader::read_function_sig(ir_function *f, s_expression *expr, bool skip_body)
+{
+   s_expression *type_expr;
+   s_list *paramlist;
+   s_list *body_list;
+
+   s_pattern pat[] = { "signature", type_expr, paramlist, body_list };
+   if (!MATCH(expr, pat)) {
+      ir_read_error(expr, "Expected (signature <type> (parameters ...) "
+			  "(<instruction> ...))");
+      return;
+   }
+
+   const glsl_type *return_type = read_type(type_expr);
+   if (return_type == NULL)
+      return;
+
+   s_symbol *paramtag = SX_AS_SYMBOL(paramlist->subexpressions.get_head());
+   if (paramtag == NULL || strcmp(paramtag->value(), "parameters") != 0) {
+      ir_read_error(paramlist, "Expected (parameters ...)");
+      return;
+   }
+
+   // Read the parameters list into a temporary place.
+   exec_list hir_parameters;
+   state->symbols->push_scope();
+
+   exec_list_iterator it = paramlist->subexpressions.iterator();
+   for (it.next() /* skip "parameters" */; it.has_next(); it.next()) {
+      ir_variable *var = read_declaration((s_expression *) it.get());
+      if (var == NULL)
+	 return;
+
+      hir_parameters.push_tail(var);
+   }
+
+   ir_function_signature *sig = f->exact_matching_signature(&hir_parameters);
+   if (sig == NULL && skip_body) {
+      /* If scanning for prototypes, generate a new signature. */
+      sig = new(mem_ctx) ir_function_signature(return_type);
+      sig->is_builtin = true;
+      f->add_signature(sig);
+   } else if (sig != NULL) {
+      const char *badvar = sig->qualifiers_match(&hir_parameters);
+      if (badvar != NULL) {
+	 ir_read_error(expr, "function `%s' parameter `%s' qualifiers "
+		       "don't match prototype", f->name, badvar);
+	 return;
+      }
+
+      if (sig->return_type != return_type) {
+	 ir_read_error(expr, "function `%s' return type doesn't "
+		       "match prototype", f->name);
+	 return;
+      }
+   } else {
+      /* No prototype for this body exists - skip it. */
+      state->symbols->pop_scope();
+      return;
+   }
+   assert(sig != NULL);
+
+   sig->replace_parameters(&hir_parameters);
+
+   if (!skip_body && !body_list->subexpressions.is_empty()) {
+      if (sig->is_defined) {
+	 ir_read_error(expr, "function %s redefined", f->name);
+	 return;
+      }
+      state->current_function = sig;
+      read_instructions(&sig->body, body_list, NULL);
+      state->current_function = NULL;
+      sig->is_defined = true;
+   }
+
+   state->symbols->pop_scope();
+}
+
+void
+ir_reader::read_instructions(exec_list *instructions, s_expression *expr,
+			     ir_loop *loop_ctx)
+{
+   // Read in a list of instructions
+   s_list *list = SX_AS_LIST(expr);
+   if (list == NULL) {
+      ir_read_error(expr, "Expected (<instruction> ...); found an atom.");
+      return;
+   }
+
+   foreach_iter(exec_list_iterator, it, list->subexpressions) {
+      s_expression *sub = (s_expression*) it.get();
+      ir_instruction *ir = read_instruction(sub, loop_ctx);
+      if (ir != NULL) {
+	 /* Global variable declarations should be moved to the top, before
+	  * any functions that might use them.  Functions are added to the
+	  * instruction stream when scanning for prototypes, so without this
+	  * hack, they always appear before variable declarations.
+	  */
+	 if (state->current_function == NULL && ir->as_variable() != NULL)
+	    instructions->push_head(ir);
+	 else
+	    instructions->push_tail(ir);
+      }
+   }
+}
+
+
+ir_instruction *
+ir_reader::read_instruction(s_expression *expr, ir_loop *loop_ctx)
+{
+   s_symbol *symbol = SX_AS_SYMBOL(expr);
+   if (symbol != NULL) {
+      if (strcmp(symbol->value(), "break") == 0 && loop_ctx != NULL)
+	 return new(mem_ctx) ir_loop_jump(ir_loop_jump::jump_break);
+      if (strcmp(symbol->value(), "continue") == 0 && loop_ctx != NULL)
+	 return new(mem_ctx) ir_loop_jump(ir_loop_jump::jump_continue);
+   }
+
+   s_list *list = SX_AS_LIST(expr);
+   if (list == NULL || list->subexpressions.is_empty()) {
+      ir_read_error(expr, "Invalid instruction.\n");
+      return NULL;
+   }
+
+   s_symbol *tag = SX_AS_SYMBOL(list->subexpressions.get_head());
+   if (tag == NULL) {
+      ir_read_error(expr, "expected instruction tag");
+      return NULL;
+   }
+
+   ir_instruction *inst = NULL;
+   if (strcmp(tag->value(), "declare") == 0) {
+      inst = read_declaration(list);
+   } else if (strcmp(tag->value(), "assign") == 0) {
+      inst = read_assignment(list);
+   } else if (strcmp(tag->value(), "if") == 0) {
+      inst = read_if(list, loop_ctx);
+   } else if (strcmp(tag->value(), "loop") == 0) {
+      inst = read_loop(list);
+   } else if (strcmp(tag->value(), "return") == 0) {
+      inst = read_return(list);
+   } else if (strcmp(tag->value(), "function") == 0) {
+      inst = read_function(list, false);
+   } else {
+      inst = read_rvalue(list);
+      if (inst == NULL)
+	 ir_read_error(NULL, "when reading instruction");
+   }
+   return inst;
+}
+
+ir_variable *
+ir_reader::read_declaration(s_expression *expr)
+{
+   s_list *s_quals;
+   s_expression *s_type;
+   s_symbol *s_name;
+
+   s_pattern pat[] = { "declare", s_quals, s_type, s_name };
+   if (!MATCH(expr, pat)) {
+      ir_read_error(expr, "expected (declare (<qualifiers>) <type> <name>)");
+      return NULL;
+   }
+
+   const glsl_type *type = read_type(s_type);
+   if (type == NULL)
+      return NULL;
+
+   ir_variable *var = new(mem_ctx) ir_variable(type, s_name->value(),
+					       ir_var_auto);
+
+   foreach_iter(exec_list_iterator, it, s_quals->subexpressions) {
+      s_symbol *qualifier = SX_AS_SYMBOL(it.get());
+      if (qualifier == NULL) {
+	 ir_read_error(expr, "qualifier list must contain only symbols");
+	 return NULL;
+      }
+
+      // FINISHME: Check for duplicate/conflicting qualifiers.
+      if (strcmp(qualifier->value(), "centroid") == 0) {
+	 var->centroid = 1;
+      } else if (strcmp(qualifier->value(), "invariant") == 0) {
+	 var->invariant = 1;
+      } else if (strcmp(qualifier->value(), "uniform") == 0) {
+	 var->mode = ir_var_uniform;
+      } else if (strcmp(qualifier->value(), "auto") == 0) {
+	 var->mode = ir_var_auto;
+      } else if (strcmp(qualifier->value(), "in") == 0) {
+	 var->mode = ir_var_in;
+      } else if (strcmp(qualifier->value(), "const_in") == 0) {
+	 var->mode = ir_var_const_in;
+      } else if (strcmp(qualifier->value(), "out") == 0) {
+	 var->mode = ir_var_out;
+      } else if (strcmp(qualifier->value(), "inout") == 0) {
+	 var->mode = ir_var_inout;
+      } else if (strcmp(qualifier->value(), "smooth") == 0) {
+	 var->interpolation = ir_var_smooth;
+      } else if (strcmp(qualifier->value(), "flat") == 0) {
+	 var->interpolation = ir_var_flat;
+      } else if (strcmp(qualifier->value(), "noperspective") == 0) {
+	 var->interpolation = ir_var_noperspective;
+      } else {
+	 ir_read_error(expr, "unknown qualifier: %s", qualifier->value());
+	 return NULL;
+      }
+   }
+
+   // Add the variable to the symbol table
+   state->symbols->add_variable(var);
+
+   return var;
+}
+
+
+ir_if *
+ir_reader::read_if(s_expression *expr, ir_loop *loop_ctx)
+{
+   s_expression *s_cond;
+   s_expression *s_then;
+   s_expression *s_else;
+
+   s_pattern pat[] = { "if", s_cond, s_then, s_else };
+   if (!MATCH(expr, pat)) {
+      ir_read_error(expr, "expected (if <condition> (<then>...) (<else>...))");
+      return NULL;
+   }
+
+   ir_rvalue *condition = read_rvalue(s_cond);
+   if (condition == NULL) {
+      ir_read_error(NULL, "when reading condition of (if ...)");
+      return NULL;
+   }
+
+   ir_if *iff = new(mem_ctx) ir_if(condition);
+
+   read_instructions(&iff->then_instructions, s_then, loop_ctx);
+   read_instructions(&iff->else_instructions, s_else, loop_ctx);
+   if (state->error) {
+      delete iff;
+      iff = NULL;
+   }
+   return iff;
+}
+
+
+ir_loop *
+ir_reader::read_loop(s_expression *expr)
+{
+   s_expression *s_counter, *s_from, *s_to, *s_inc, *s_body;
+
+   s_pattern pat[] = { "loop", s_counter, s_from, s_to, s_inc, s_body };
+   if (!MATCH(expr, pat)) {
+      ir_read_error(expr, "expected (loop <counter> <from> <to> "
+			  "<increment> <body>)");
+      return NULL;
+   }
+
+   // FINISHME: actually read the count/from/to fields.
+
+   ir_loop *loop = new(mem_ctx) ir_loop;
+   read_instructions(&loop->body_instructions, s_body, loop);
+   if (state->error) {
+      delete loop;
+      loop = NULL;
+   }
+   return loop;
+}
+
+
+ir_return *
+ir_reader::read_return(s_expression *expr)
+{
+   s_expression *s_retval;
+
+   s_pattern return_value_pat[] = { "return", s_retval};
+   s_pattern return_void_pat[] = { "return" };
+   if (MATCH(expr, return_value_pat)) {
+      ir_rvalue *retval = read_rvalue(s_retval);
+      if (retval == NULL) {
+         ir_read_error(NULL, "when reading return value");
+         return NULL;
+      }
+      return new(mem_ctx) ir_return(retval);
+   } else if (MATCH(expr, return_void_pat)) {
+      return new(mem_ctx) ir_return;
+   } else {
+      ir_read_error(expr, "expected (return <rvalue>) or (return)");
+      return NULL;
+   }
+}
+
+
+ir_rvalue *
+ir_reader::read_rvalue(s_expression *expr)
+{
+   s_list *list = SX_AS_LIST(expr);
+   if (list == NULL || list->subexpressions.is_empty())
+      return NULL;
+
+   s_symbol *tag = SX_AS_SYMBOL(list->subexpressions.get_head());
+   if (tag == NULL) {
+      ir_read_error(expr, "expected rvalue tag");
+      return NULL;
+   }
+
+   ir_rvalue *rvalue = read_dereference(list);
+   if (rvalue != NULL || state->error)
+      return rvalue;
+   else if (strcmp(tag->value(), "swiz") == 0) {
+      rvalue = read_swizzle(list);
+   } else if (strcmp(tag->value(), "expression") == 0) {
+      rvalue = read_expression(list);
+   } else if (strcmp(tag->value(), "call") == 0) {
+      rvalue = read_call(list);
+   } else if (strcmp(tag->value(), "constant") == 0) {
+      rvalue = read_constant(list);
+   } else {
+      rvalue = read_texture(list);
+      if (rvalue == NULL && !state->error)
+	 ir_read_error(expr, "unrecognized rvalue tag: %s", tag->value());
+   }
+
+   return rvalue;
+}
+
+ir_assignment *
+ir_reader::read_assignment(s_expression *expr)
+{
+   s_expression *cond_expr = NULL;
+   s_expression *lhs_expr, *rhs_expr;
+   s_list       *mask_list;
+
+   s_pattern pat4[] = { "assign",            mask_list, lhs_expr, rhs_expr };
+   s_pattern pat5[] = { "assign", cond_expr, mask_list, lhs_expr, rhs_expr };
+   if (!MATCH(expr, pat4) && !MATCH(expr, pat5)) {
+      ir_read_error(expr, "expected (assign [<condition>] (<write mask>) "
+			  "<lhs> <rhs>)");
+      return NULL;
+   }
+
+   ir_rvalue *condition = NULL;
+   if (cond_expr != NULL) {
+      condition = read_rvalue(cond_expr);
+      if (condition == NULL) {
+	 ir_read_error(NULL, "when reading condition of assignment");
+	 return NULL;
+      }
+   }
+
+   unsigned mask = 0;
+
+   s_symbol *mask_symbol;
+   s_pattern mask_pat[] = { mask_symbol };
+   if (MATCH(mask_list, mask_pat)) {
+      const char *mask_str = mask_symbol->value();
+      unsigned mask_length = strlen(mask_str);
+      if (mask_length > 4) {
+	 ir_read_error(expr, "invalid write mask: %s", mask_str);
+	 return NULL;
+      }
+
+      const unsigned idx_map[] = { 3, 0, 1, 2 }; /* w=bit 3, x=0, y=1, z=2 */
+
+      for (unsigned i = 0; i < mask_length; i++) {
+	 if (mask_str[i] < 'w' || mask_str[i] > 'z') {
+	    ir_read_error(expr, "write mask contains invalid character: %c",
+			  mask_str[i]);
+	    return NULL;
+	 }
+	 mask |= 1 << idx_map[mask_str[i] - 'w'];
+      }
+   } else if (!mask_list->subexpressions.is_empty()) {
+      ir_read_error(mask_list, "expected () or (<write mask>)");
+      return NULL;
+   }
+
+   ir_dereference *lhs = read_dereference(lhs_expr);
+   if (lhs == NULL) {
+      ir_read_error(NULL, "when reading left-hand side of assignment");
+      return NULL;
+   }
+
+   ir_rvalue *rhs = read_rvalue(rhs_expr);
+   if (rhs == NULL) {
+      ir_read_error(NULL, "when reading right-hand side of assignment");
+      return NULL;
+   }
+
+   if (mask == 0 && (lhs->type->is_vector() || lhs->type->is_scalar())) {
+      ir_read_error(expr, "non-zero write mask required.");
+      return NULL;
+   }
+
+   return new(mem_ctx) ir_assignment(lhs, rhs, condition, mask);
+}
+
+ir_call *
+ir_reader::read_call(s_expression *expr)
+{
+   s_symbol *name;
+   s_list *params;
+
+   s_pattern pat[] = { "call", name, params };
+   if (!MATCH(expr, pat)) {
+      ir_read_error(expr, "expected (call <name> (<param> ...))");
+      return NULL;
+   }
+
+   exec_list parameters;
+
+   foreach_iter(exec_list_iterator, it, params->subexpressions) {
+      s_expression *expr = (s_expression*) it.get();
+      ir_rvalue *param = read_rvalue(expr);
+      if (param == NULL) {
+	 ir_read_error(expr, "when reading parameter to function call");
+	 return NULL;
+      }
+      parameters.push_tail(param);
+   }
+
+   ir_function *f = state->symbols->get_function(name->value());
+   if (f == NULL) {
+      ir_read_error(expr, "found call to undefined function %s",
+		    name->value());
+      return NULL;
+   }
+
+   ir_function_signature *callee = f->matching_signature(&parameters);
+   if (callee == NULL) {
+      ir_read_error(expr, "couldn't find matching signature for function "
+                    "%s", name->value());
+      return NULL;
+   }
+
+   return new(mem_ctx) ir_call(callee, &parameters);
+}
+
+ir_expression *
+ir_reader::read_expression(s_expression *expr)
+{
+   s_expression *s_type;
+   s_symbol *s_op;
+   s_expression *s_arg1;
+
+   s_pattern pat[] = { "expression", s_type, s_op, s_arg1 };
+   if (!PARTIAL_MATCH(expr, pat)) {
+      ir_read_error(expr, "expected (expression <type> <operator> "
+			  "<operand> [<operand>])");
+      return NULL;
+   }
+   s_expression *s_arg2 = (s_expression *) s_arg1->next; // may be tail sentinel
+
+   const glsl_type *type = read_type(s_type);
+   if (type == NULL)
+      return NULL;
+
+   /* Read the operator */
+   ir_expression_operation op = ir_expression::get_operator(s_op->value());
+   if (op == (ir_expression_operation) -1) {
+      ir_read_error(expr, "invalid operator: %s", s_op->value());
+      return NULL;
+   }
+    
+   unsigned num_operands = ir_expression::get_num_operands(op);
+   if (num_operands == 1 && !s_arg1->next->is_tail_sentinel()) {
+      ir_read_error(expr, "expected (expression <type> %s <operand>)",
+		    s_op->value());
+      return NULL;
+   }
+
+   ir_rvalue *arg1 = read_rvalue(s_arg1);
+   ir_rvalue *arg2 = NULL;
+   if (arg1 == NULL) {
+      ir_read_error(NULL, "when reading first operand of %s", s_op->value());
+      return NULL;
+   }
+
+   if (num_operands == 2) {
+      if (s_arg2->is_tail_sentinel() || !s_arg2->next->is_tail_sentinel()) {
+	 ir_read_error(expr, "expected (expression <type> %s <operand> "
+			     "<operand>)", s_op->value());
+	 return NULL;
+      }
+      arg2 = read_rvalue(s_arg2);
+      if (arg2 == NULL) {
+	 ir_read_error(NULL, "when reading second operand of %s",
+		       s_op->value());
+	 return NULL;
+      }
+   }
+
+   return new(mem_ctx) ir_expression(op, type, arg1, arg2);
+}
+
+ir_swizzle *
+ir_reader::read_swizzle(s_expression *expr)
+{
+   s_symbol *swiz;
+   s_expression *sub;
+
+   s_pattern pat[] = { "swiz", swiz, sub };
+   if (!MATCH(expr, pat)) {
+      ir_read_error(expr, "expected (swiz <swizzle> <rvalue>)");
+      return NULL;
+   }
+
+   if (strlen(swiz->value()) > 4) {
+      ir_read_error(expr, "expected a valid swizzle; found %s", swiz->value());
+      return NULL;
+   }
+
+   ir_rvalue *rvalue = read_rvalue(sub);
+   if (rvalue == NULL)
+      return NULL;
+
+   ir_swizzle *ir = ir_swizzle::create(rvalue, swiz->value(),
+				       rvalue->type->vector_elements);
+   if (ir == NULL)
+      ir_read_error(expr, "invalid swizzle");
+
+   return ir;
+}
+
+ir_constant *
+ir_reader::read_constant(s_expression *expr)
+{
+   s_expression *type_expr;
+   s_list *values;
+
+   s_pattern pat[] = { "constant", type_expr, values };
+   if (!MATCH(expr, pat)) {
+      ir_read_error(expr, "expected (constant <type> (...))");
+      return NULL;
+   }
+
+   const glsl_type *type = read_type(type_expr);
+   if (type == NULL)
+      return NULL;
+
+   if (values == NULL) {
+      ir_read_error(expr, "expected (constant <type> (...))");
+      return NULL;
+   }
+
+   if (type->is_array()) {
+      unsigned elements_supplied = 0;
+      exec_list elements;
+      foreach_iter(exec_list_iterator, it, values->subexpressions) {
+	 s_expression *elt = (s_expression *) it.get();
+	 ir_constant *ir_elt = read_constant(elt);
+	 if (ir_elt == NULL)
+	    return NULL;
+	 elements.push_tail(ir_elt);
+	 elements_supplied++;
+      }
+
+      if (elements_supplied != type->length) {
+	 ir_read_error(values, "expected exactly %u array elements, "
+		       "given %u", type->length, elements_supplied);
+	 return NULL;
+      }
+      return new(mem_ctx) ir_constant(type, &elements);
+   }
+
+   const glsl_type *const base_type = type->get_base_type();
+
+   ir_constant_data data = { { 0 } };
+
+   // Read in list of values (at most 16).
+   int k = 0;
+   foreach_iter(exec_list_iterator, it, values->subexpressions) {
+      if (k >= 16) {
+	 ir_read_error(values, "expected at most 16 numbers");
+	 return NULL;
+      }
+
+      s_expression *expr = (s_expression*) it.get();
+
+      if (base_type->base_type == GLSL_TYPE_FLOAT) {
+	 s_number *value = SX_AS_NUMBER(expr);
+	 if (value == NULL) {
+	    ir_read_error(values, "expected numbers");
+	    return NULL;
+	 }
+	 data.f[k] = value->fvalue();
+      } else {
+	 s_int *value = SX_AS_INT(expr);
+	 if (value == NULL) {
+	    ir_read_error(values, "expected integers");
+	    return NULL;
+	 }
+
+	 switch (base_type->base_type) {
+	 case GLSL_TYPE_UINT: {
+	    data.u[k] = value->value();
+	    break;
+	 }
+	 case GLSL_TYPE_INT: {
+	    data.i[k] = value->value();
+	    break;
+	 }
+	 case GLSL_TYPE_BOOL: {
+	    data.b[k] = value->value();
+	    break;
+	 }
+	 default:
+	    ir_read_error(values, "unsupported constant type");
+	    return NULL;
+	 }
+      }
+      ++k;
+   }
+
+   return new(mem_ctx) ir_constant(type, &data);
+}
+
+ir_dereference *
+ir_reader::read_dereference(s_expression *expr)
+{
+   s_symbol *s_var;
+   s_expression *s_subject;
+   s_expression *s_index;
+   s_symbol *s_field;
+
+   s_pattern var_pat[] = { "var_ref", s_var };
+   s_pattern array_pat[] = { "array_ref", s_subject, s_index };
+   s_pattern record_pat[] = { "record_ref", s_subject, s_field };
+
+   if (MATCH(expr, var_pat)) {
+      ir_variable *var = state->symbols->get_variable(s_var->value());
+      if (var == NULL) {
+	 ir_read_error(expr, "undeclared variable: %s", s_var->value());
+	 return NULL;
+      }
+      return new(mem_ctx) ir_dereference_variable(var);
+   } else if (MATCH(expr, array_pat)) {
+      ir_rvalue *subject = read_rvalue(s_subject);
+      if (subject == NULL) {
+	 ir_read_error(NULL, "when reading the subject of an array_ref");
+	 return NULL;
+      }
+
+      ir_rvalue *idx = read_rvalue(s_index);
+      if (subject == NULL) {
+	 ir_read_error(NULL, "when reading the index of an array_ref");
+	 return NULL;
+      }
+      return new(mem_ctx) ir_dereference_array(subject, idx);
+   } else if (MATCH(expr, record_pat)) {
+      ir_rvalue *subject = read_rvalue(s_subject);
+      if (subject == NULL) {
+	 ir_read_error(NULL, "when reading the subject of a record_ref");
+	 return NULL;
+      }
+      return new(mem_ctx) ir_dereference_record(subject, s_field->value());
+   }
+   return NULL;
+}
+
+ir_texture *
+ir_reader::read_texture(s_expression *expr)
+{
+   s_symbol *tag = NULL;
+   s_expression *s_type = NULL;
+   s_expression *s_sampler = NULL;
+   s_expression *s_coord = NULL;
+   s_expression *s_offset = NULL;
+   s_expression *s_proj = NULL;
+   s_list *s_shadow = NULL;
+   s_expression *s_lod = NULL;
+
+   ir_texture_opcode op = ir_tex; /* silence warning */
+
+   s_pattern tex_pattern[] =
+      { "tex", s_type, s_sampler, s_coord, s_offset, s_proj, s_shadow };
+   s_pattern txf_pattern[] =
+      { "txf", s_type, s_sampler, s_coord, s_offset, s_lod };
+   s_pattern other_pattern[] =
+      { tag, s_type, s_sampler, s_coord, s_offset, s_proj, s_shadow, s_lod };
+
+   if (MATCH(expr, tex_pattern)) {
+      op = ir_tex;
+   } else if (MATCH(expr, txf_pattern)) {
+      op = ir_txf;
+   } else if (MATCH(expr, other_pattern)) {
+      op = ir_texture::get_opcode(tag->value());
+      if (op == -1)
+	 return NULL;
+   } else {
+      ir_read_error(NULL, "unexpected texture pattern");
+      return NULL;
+   }
+
+   ir_texture *tex = new(mem_ctx) ir_texture(op);
+
+   // Read return type
+   const glsl_type *type = read_type(s_type);
+   if (type == NULL) {
+      ir_read_error(NULL, "when reading type in (%s ...)",
+		    tex->opcode_string());
+      return NULL;
+   }
+
+   // Read sampler (must be a deref)
+   ir_dereference *sampler = read_dereference(s_sampler);
+   if (sampler == NULL) {
+      ir_read_error(NULL, "when reading sampler in (%s ...)",
+		    tex->opcode_string());
+      return NULL;
+   }
+   tex->set_sampler(sampler, type);
+
+   // Read coordinate (any rvalue)
+   tex->coordinate = read_rvalue(s_coord);
+   if (tex->coordinate == NULL) {
+      ir_read_error(NULL, "when reading coordinate in (%s ...)",
+		    tex->opcode_string());
+      return NULL;
+   }
+
+   // Read texel offset - either 0 or an rvalue.
+   s_int *si_offset = SX_AS_INT(s_offset);
+   if (si_offset == NULL || si_offset->value() != 0) {
+      tex->offset = read_rvalue(s_offset);
+      if (tex->offset == NULL) {
+	 ir_read_error(s_offset, "expected 0 or an expression");
+	 return NULL;
+      }
+   }
+
+   if (op != ir_txf) {
+      s_int *proj_as_int = SX_AS_INT(s_proj);
+      if (proj_as_int && proj_as_int->value() == 1) {
+	 tex->projector = NULL;
+      } else {
+	 tex->projector = read_rvalue(s_proj);
+	 if (tex->projector == NULL) {
+	    ir_read_error(NULL, "when reading projective divide in (%s ..)",
+	                  tex->opcode_string());
+	    return NULL;
+	 }
+      }
+
+      if (s_shadow->subexpressions.is_empty()) {
+	 tex->shadow_comparitor = NULL;
+      } else {
+	 tex->shadow_comparitor = read_rvalue(s_shadow);
+	 if (tex->shadow_comparitor == NULL) {
+	    ir_read_error(NULL, "when reading shadow comparitor in (%s ..)",
+			  tex->opcode_string());
+	    return NULL;
+	 }
+      }
+   }
+
+   switch (op) {
+   case ir_txb:
+      tex->lod_info.bias = read_rvalue(s_lod);
+      if (tex->lod_info.bias == NULL) {
+	 ir_read_error(NULL, "when reading LOD bias in (txb ...)");
+	 return NULL;
+      }
+      break;
+   case ir_txl:
+   case ir_txf:
+      tex->lod_info.lod = read_rvalue(s_lod);
+      if (tex->lod_info.lod == NULL) {
+	 ir_read_error(NULL, "when reading LOD in (%s ...)",
+		       tex->opcode_string());
+	 return NULL;
+      }
+      break;
+   case ir_txd: {
+      s_expression *s_dx, *s_dy;
+      s_pattern dxdy_pat[] = { s_dx, s_dy };
+      if (!MATCH(s_lod, dxdy_pat)) {
+	 ir_read_error(s_lod, "expected (dPdx dPdy) in (txd ...)");
+	 return NULL;
+      }
+      tex->lod_info.grad.dPdx = read_rvalue(s_dx);
+      if (tex->lod_info.grad.dPdx == NULL) {
+	 ir_read_error(NULL, "when reading dPdx in (txd ...)");
+	 return NULL;
+      }
+      tex->lod_info.grad.dPdy = read_rvalue(s_dy);
+      if (tex->lod_info.grad.dPdy == NULL) {
+	 ir_read_error(NULL, "when reading dPdy in (txd ...)");
+	 return NULL;
+      }
+      break;
+   }
+   default:
+      // tex doesn't have any extra parameters.
+      break;
+   };
+   return tex;
+}
diff --git a/mesalib/src/glsl/linker.cpp b/mesalib/src/glsl/linker.cpp
index 265da84e5..34b64837a 100644
--- a/mesalib/src/glsl/linker.cpp
+++ b/mesalib/src/glsl/linker.cpp
@@ -1248,7 +1248,7 @@ assign_attribute_or_color_locations(gl_shader_program *prog,
     */
 
    const int generic_base = (target_index == MESA_SHADER_VERTEX)
-     ? VERT_ATTRIB_GENERIC0 : FRAG_RESULT_DATA0;
+      ? (int) VERT_ATTRIB_GENERIC0 : (int) FRAG_RESULT_DATA0;
 
    const enum ir_variable_mode direction =
       (target_index == MESA_SHADER_VERTEX) ? ir_var_in : ir_var_out;
diff --git a/mesalib/src/glsl/lower_jumps.cpp b/mesalib/src/glsl/lower_jumps.cpp
index 17be39f5b..61874990a 100644
--- a/mesalib/src/glsl/lower_jumps.cpp
+++ b/mesalib/src/glsl/lower_jumps.cpp
@@ -1,570 +1,1010 @@
-/*

- * Copyright © 2010 Luca Barbieri

- *

- * Permission is hereby granted, free of charge, to any person obtaining a

- * copy of this software and 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 lower_jumps.cpp

- *

- * This pass lowers jumps (break, continue, and return) to if/else structures.

- *

- * It can be asked to:

- * 1. Pull jumps out of ifs where possible

- * 2. Remove all "continue"s, replacing them with an "execute flag"

- * 3. Replace all "break" with a single conditional one at the end of the loop

- * 4. Replace all "return"s with a single return at the end of the function,

- *    for the main function and/or other functions

- *

- * Applying this pass gives several benefits:

- * 1. All functions can be inlined.

- * 2. nv40 and other pre-DX10 chips without "continue" can be supported

- * 3. nv30 and other pre-DX10 chips with no control flow at all are better

- *    supported

- *

- * Continues are lowered by adding a per-loop "execute flag", initialized to

- * true, that when cleared inhibits all execution until the end of the loop.

- *

- * Breaks are lowered to continues, plus setting a "break flag" that is checked

- * at the end of the loop, and trigger the unique "break".

- *

- * Returns are lowered to breaks/continues, plus adding a "return flag" that

- * causes loops to break again out of their enclosing loops until all the

- * loops are exited: then the "execute flag" logic will ignore everything

- * until the end of the function.

- *

- * Note that "continue" and "return" can also be implemented by adding

- * a dummy loop and using break.

- * However, this is bad for hardware with limited nesting depth, and

- * prevents further optimization, and thus is not currently performed.

- */

-

-#include "glsl_types.h"

-#include <string.h>

-#include "ir.h"

-

-enum jump_strength

-{

-   strength_none,

-   strength_always_clears_execute_flag,

-   strength_continue,

-   strength_break,

-   strength_return

-};

-

-struct block_record

-{

-   /* minimum jump strength (of lowered IR, not pre-lowering IR)

-    *

-    * If the block ends with a jump, must be the strength of the jump.

-    * Otherwise, the jump would be dead and have been deleted before)

-    *

-    * If the block doesn't end with a jump, it can be different than strength_none if all paths before it lead to some jump

-    * (e.g. an if with a return in one branch, and a break in the other, while not lowering them)

-    * Note that identical jumps are usually unified though.

-    */

-   jump_strength min_strength;

-

-   /* can anything clear the execute flag? */

-   bool may_clear_execute_flag;

-

-   block_record()

-   {

-      this->min_strength = strength_none;

-      this->may_clear_execute_flag = false;

-   }

-};

-

-struct loop_record

-{

-   ir_function_signature* signature;

-   ir_loop* loop;

-

-   /* used to avoid lowering the break used to represent lowered breaks */

-   unsigned nesting_depth;

-   bool in_if_at_the_end_of_the_loop;

-

-   bool may_set_return_flag;

-

-   ir_variable* break_flag;

-   ir_variable* execute_flag; /* cleared to emulate continue */

-

-   loop_record(ir_function_signature* p_signature = 0, ir_loop* p_loop = 0)

-   {

-      this->signature = p_signature;

-      this->loop = p_loop;

-      this->nesting_depth = 0;

-      this->in_if_at_the_end_of_the_loop = false;

-      this->may_set_return_flag = false;

-      this->break_flag = 0;

-      this->execute_flag = 0;

-   }

-

-   ir_variable* get_execute_flag()

-   {

-      /* also supported for the "function loop" */

-      if(!this->execute_flag) {

-         exec_list& list = this->loop ? this->loop->body_instructions : signature->body;

-         this->execute_flag = new(this->signature) ir_variable(glsl_type::bool_type, "execute_flag", ir_var_temporary);

-         list.push_head(new(this->signature) ir_assignment(new(this->signature) ir_dereference_variable(execute_flag), new(this->signature) ir_constant(true), 0));

-         list.push_head(this->execute_flag);

-      }

-      return this->execute_flag;

-   }

-

-   ir_variable* get_break_flag()

-   {

-      assert(this->loop);

-      if(!this->break_flag) {

-         this->break_flag = new(this->signature) ir_variable(glsl_type::bool_type, "break_flag", ir_var_temporary);

-         this->loop->insert_before(this->break_flag);

-         this->loop->insert_before(new(this->signature) ir_assignment(new(this->signature) ir_dereference_variable(break_flag), new(this->signature) ir_constant(false), 0));

-      }

-      return this->break_flag;

-   }

-};

-

-struct function_record

-{

-   ir_function_signature* signature;

-   ir_variable* return_flag; /* used to break out of all loops and then jump to the return instruction */

-   ir_variable* return_value;

-   bool is_main;

-   unsigned nesting_depth;

-

-   function_record(ir_function_signature* p_signature = 0)

-   {

-      this->signature = p_signature;

-      this->return_flag = 0;

-      this->return_value = 0;

-      this->nesting_depth = 0;

-      this->is_main = this->signature && (strcmp(this->signature->function_name(), "main") == 0);

-   }

-

-   ir_variable* get_return_flag()

-   {

-      if(!this->return_flag) {

-         this->return_flag = new(this->signature) ir_variable(glsl_type::bool_type, "return_flag", ir_var_temporary);

-         this->signature->body.push_head(new(this->signature) ir_assignment(new(this->signature) ir_dereference_variable(return_flag), new(this->signature) ir_constant(false), 0));

-         this->signature->body.push_head(this->return_flag);

-      }

-      return this->return_flag;

-   }

-

-   ir_variable* get_return_value()

-   {

-      if(!this->return_value) {

-         assert(!this->signature->return_type->is_void());

-         return_value = new(this->signature) ir_variable(this->signature->return_type, "return_value", ir_var_temporary);

-         this->signature->body.push_head(this->return_value);

-      }

-      return this->return_value;

-   }

-};

-

-struct ir_lower_jumps_visitor : public ir_control_flow_visitor {

-   bool progress;

-

-   struct function_record function;

-   struct loop_record loop;

-   struct block_record block;

-

-   bool pull_out_jumps;

-   bool lower_continue;

-   bool lower_break;

-   bool lower_sub_return;

-   bool lower_main_return;

-

-   ir_lower_jumps_visitor()

-   {

-      this->progress = false;

-   }

-

-   void truncate_after_instruction(exec_node *ir)

-   {

-      if (!ir)

-         return;

-

-      while (!ir->get_next()->is_tail_sentinel()) {

-         ((ir_instruction *)ir->get_next())->remove();

-         this->progress = true;

-      }

-   }

-

-   void move_outer_block_inside(ir_instruction *ir, exec_list *inner_block)

-   {

-      while (!ir->get_next()->is_tail_sentinel()) {

-         ir_instruction *move_ir = (ir_instruction *)ir->get_next();

-

-         move_ir->remove();

-         inner_block->push_tail(move_ir);

-      }

-   }

-

-   virtual void visit(class ir_loop_jump * ir)

-   {

-      truncate_after_instruction(ir);

-      this->block.min_strength = ir->is_break() ? strength_break : strength_continue;

-   }

-

-   virtual void visit(class ir_return * ir)

-   {

-      truncate_after_instruction(ir);

-      this->block.min_strength = strength_return;

-   }

-

-   virtual void visit(class ir_discard * ir)

-   {

-   }

-

-   enum jump_strength get_jump_strength(ir_instruction* ir)

-   {

-      if(!ir)

-         return strength_none;

-      else if(ir->ir_type == ir_type_loop_jump) {

-         if(((ir_loop_jump*)ir)->is_break())

-            return strength_break;

-         else

-            return strength_continue;

-      } else if(ir->ir_type == ir_type_return)

-         return strength_return;

-      else

-         return strength_none;

-   }

-

-   bool should_lower_jump(ir_jump* ir)

-   {

-      unsigned strength = get_jump_strength(ir);

-      bool lower;

-      switch(strength)

-      {

-      case strength_none:

-         lower = false; /* don't change this, code relies on it */

-         break;

-      case strength_continue:

-         lower = lower_continue;

-         break;

-      case strength_break:

-         assert(this->loop.loop);

-         /* never lower "canonical break" */

-         if(ir->get_next()->is_tail_sentinel() && (this->loop.nesting_depth == 0

-               || (this->loop.nesting_depth == 1 && this->loop.in_if_at_the_end_of_the_loop)))

-            lower = false;

-         else

-            lower = lower_break;

-         break;

-      case strength_return:

-         /* never lower return at the end of a this->function */

-         if(this->function.nesting_depth == 0 && ir->get_next()->is_tail_sentinel())

-            lower = false;

-         else if (this->function.is_main)

-            lower = lower_main_return;

-         else

-            lower = lower_sub_return;

-         break;

-      }

-      return lower;

-   }

-

-   block_record visit_block(exec_list* list)

-   {

-      block_record saved_block = this->block;

-      this->block = block_record();

-      visit_exec_list(list, this);

-      block_record ret = this->block;

-      this->block = saved_block;

-      return ret;

-   }

-

-   virtual void visit(ir_if *ir)

-   {

-      if(this->loop.nesting_depth == 0 && ir->get_next()->is_tail_sentinel())

-         this->loop.in_if_at_the_end_of_the_loop = true;

-

-      ++this->function.nesting_depth;

-      ++this->loop.nesting_depth;

-

-      block_record block_records[2];

-      ir_jump* jumps[2];

-

-      block_records[0] = visit_block(&ir->then_instructions);

-      block_records[1] = visit_block(&ir->else_instructions);

-

-retry: /* we get here if we put code after the if inside a branch */

-   for(unsigned i = 0; i < 2; ++i) {

-      exec_list& list = i ? ir->else_instructions : ir->then_instructions;

-      jumps[i] = 0;

-      if(!list.is_empty() && get_jump_strength((ir_instruction*)list.get_tail()))

-         jumps[i] = (ir_jump*)list.get_tail();

-   }

-

-      for(;;) {

-         jump_strength jump_strengths[2];

-

-         for(unsigned i = 0; i < 2; ++i) {

-            if(jumps[i]) {

-               jump_strengths[i] = block_records[i].min_strength;

-               assert(jump_strengths[i] == get_jump_strength(jumps[i]));

-            } else

-               jump_strengths[i] = strength_none;

-         }

-

-         /* move both jumps out if possible */

-         if(pull_out_jumps && jump_strengths[0] == jump_strengths[1]) {

-            bool unify = true;

-            if(jump_strengths[0] == strength_continue)

-               ir->insert_after(new(ir) ir_loop_jump(ir_loop_jump::jump_continue));

-            else if(jump_strengths[0] == strength_break)

-               ir->insert_after(new(ir) ir_loop_jump(ir_loop_jump::jump_break));

-            /* FINISHME: unify returns with identical expressions */

-            else if(jump_strengths[0] == strength_return && this->function.signature->return_type->is_void())

-               ir->insert_after(new(ir) ir_return(NULL));

-	    else

-	       unify = false;

-

-            if(unify) {

-               jumps[0]->remove();

-               jumps[1]->remove();

-               this->progress = true;

-

-               jumps[0] = 0;

-               jumps[1] = 0;

-               block_records[0].min_strength = strength_none;

-               block_records[1].min_strength = strength_none;

-               break;

-            }

-         }

-

-         /* lower a jump: if both need to lowered, start with the strongest one, so that

-          * we might later unify the lowered version with the other one

-          */

-         bool should_lower[2];

-         for(unsigned i = 0; i < 2; ++i)

-            should_lower[i] = should_lower_jump(jumps[i]);

-

-         int lower;

-         if(should_lower[1] && should_lower[0])

-            lower = jump_strengths[1] > jump_strengths[0];

-         else if(should_lower[0])

-            lower = 0;

-         else if(should_lower[1])

-            lower = 1;

-         else

-            break;

-

-         if(jump_strengths[lower] == strength_return) {

-            ir_variable* return_flag = this->function.get_return_flag();

-            if(!this->function.signature->return_type->is_void()) {

-               ir_variable* return_value = this->function.get_return_value();

-               jumps[lower]->insert_before(new(ir) ir_assignment(new (ir) ir_dereference_variable(return_value), ((ir_return*)jumps[lower])->value, NULL));

-            }

-            jumps[lower]->insert_before(new(ir) ir_assignment(new (ir) ir_dereference_variable(return_flag), new (ir) ir_constant(true), NULL));

-            this->loop.may_set_return_flag = true;

-            if(this->loop.loop) {

-               ir_loop_jump* lowered = 0;

-               lowered = new(ir) ir_loop_jump(ir_loop_jump::jump_break);

-               block_records[lower].min_strength = strength_break;

-               jumps[lower]->replace_with(lowered);

-               jumps[lower] = lowered;

-            } else

-               goto lower_continue;

-            this->progress = true;

-         } else if(jump_strengths[lower] == strength_break) {

-            /* We can't lower to an actual continue because that would execute the increment.

-             *

-             * In the lowered code, we instead put the break check between the this->loop body and the increment,

-             * which is impossible with a real continue as defined by the GLSL IR currently.

-             *

-             * Smarter options (such as undoing the increment) are possible but it's not worth implementing them,

-             * because if break is lowered, continue is almost surely lowered too.

-             */

-            jumps[lower]->insert_before(new(ir) ir_assignment(new (ir) ir_dereference_variable(this->loop.get_break_flag()), new (ir) ir_constant(true), 0));

-            goto lower_continue;

-         } else if(jump_strengths[lower] == strength_continue) {

-lower_continue:

-            ir_variable* execute_flag = this->loop.get_execute_flag();

-            jumps[lower]->replace_with(new(ir) ir_assignment(new (ir) ir_dereference_variable(execute_flag), new (ir) ir_constant(false), 0));

-            jumps[lower] = 0;

-            block_records[lower].min_strength = strength_always_clears_execute_flag;

-            block_records[lower].may_clear_execute_flag = true;

-            this->progress = true;

-            break;

-         }

-      }

-

-      /* move out a jump out if possible */

-      if(pull_out_jumps) {

-         int move_out = -1;

-         if(jumps[0] && block_records[1].min_strength >= strength_continue)

-            move_out = 0;

-         else if(jumps[1] && block_records[0].min_strength >= strength_continue)

-            move_out = 1;

-

-         if(move_out >= 0)

-         {

-            jumps[move_out]->remove();

-            ir->insert_after(jumps[move_out]);

-            jumps[move_out] = 0;

-            block_records[move_out].min_strength = strength_none;

-            this->progress = true;

-         }

-      }

-

-      if(block_records[0].min_strength < block_records[1].min_strength)

-         this->block.min_strength = block_records[0].min_strength;

-      else

-         this->block.min_strength = block_records[1].min_strength;

-      this->block.may_clear_execute_flag = this->block.may_clear_execute_flag || block_records[0].may_clear_execute_flag || block_records[1].may_clear_execute_flag;

-

-      if(this->block.min_strength)

-         truncate_after_instruction(ir);

-      else if(this->block.may_clear_execute_flag)

-      {

-         int move_into = -1;

-         if(block_records[0].min_strength && !block_records[1].may_clear_execute_flag)

-            move_into = 1;

-         else if(block_records[1].min_strength && !block_records[0].may_clear_execute_flag)

-            move_into = 0;

-

-         if(move_into >= 0) {

-            assert(!block_records[move_into].min_strength && !block_records[move_into].may_clear_execute_flag); /* otherwise, we just truncated */

-

-            exec_list* list = move_into ? &ir->else_instructions : &ir->then_instructions;

-            exec_node* next = ir->get_next();

-            if(!next->is_tail_sentinel()) {

-               move_outer_block_inside(ir, list);

-

-               exec_list list;

-               list.head = next;

-               block_records[move_into] = visit_block(&list);

-

-               this->progress = true;

-               goto retry;

-            }

-         } else {

-            ir_instruction* ir_after;

-            for(ir_after = (ir_instruction*)ir->get_next(); !ir_after->is_tail_sentinel();)

-            {

-               ir_if* ir_if = ir_after->as_if();

-               if(ir_if && ir_if->else_instructions.is_empty()) {

-                  ir_dereference_variable* ir_if_cond_deref = ir_if->condition->as_dereference_variable();

-                  if(ir_if_cond_deref && ir_if_cond_deref->var == this->loop.execute_flag) {

-                     ir_instruction* ir_next = (ir_instruction*)ir_after->get_next();

-                     ir_after->insert_before(&ir_if->then_instructions);

-                     ir_after->remove();

-                     ir_after = ir_next;

-                     continue;

-                  }

-               }

-               ir_after = (ir_instruction*)ir_after->get_next();

-

-               /* only set this if we find any unprotected instruction */

-               this->progress = true;

-            }

-

-            if(!ir->get_next()->is_tail_sentinel()) {

-               assert(this->loop.execute_flag);

-               ir_if* if_execute = new(ir) ir_if(new(ir) ir_dereference_variable(this->loop.execute_flag));

-               move_outer_block_inside(ir, &if_execute->then_instructions);

-               ir->insert_after(if_execute);

-            }

-         }

-      }

-      --this->loop.nesting_depth;

-      --this->function.nesting_depth;

-   }

-

-   virtual void visit(ir_loop *ir)

-   {

-      ++this->function.nesting_depth;

-      loop_record saved_loop = this->loop;

-      this->loop = loop_record(this->function.signature, ir);

-

-      block_record body = visit_block(&ir->body_instructions);

-

-      if(body.min_strength >= strength_break) {

-         /* FINISHME: turn the this->loop into an if, or replace it with its body */

-      }

-

-      if(this->loop.break_flag) {

-         ir_if* break_if = new(ir) ir_if(new(ir) ir_dereference_variable(this->loop.break_flag));

-         break_if->then_instructions.push_tail(new(ir) ir_loop_jump(ir_loop_jump::jump_break));

-         ir->body_instructions.push_tail(break_if);

-      }

-

-      if(this->loop.may_set_return_flag) {

-         assert(this->function.return_flag);

-         ir_if* return_if = new(ir) ir_if(new(ir) ir_dereference_variable(this->function.return_flag));

-         saved_loop.may_set_return_flag = true;

-         if(saved_loop.loop)

-            return_if->then_instructions.push_tail(new(ir) ir_loop_jump(ir_loop_jump::jump_break));

-         else

-            move_outer_block_inside(ir, &return_if->else_instructions);

-         ir->insert_after(return_if);

-      }

-

-      this->loop = saved_loop;

-      --this->function.nesting_depth;

-   }

-

-   virtual void visit(ir_function_signature *ir)

-   {

-      /* these are not strictly necessary */

-      assert(!this->function.signature);

-      assert(!this->loop.loop);

-

-      function_record saved_function = this->function;

-      loop_record saved_loop = this->loop;

-      this->function = function_record(ir);

-      this->loop = loop_record(ir);

-

-      assert(!this->loop.loop);

-      visit_block(&ir->body);

-

-      if(this->function.return_value)

-         ir->body.push_tail(new(ir) ir_return(new (ir) ir_dereference_variable(this->function.return_value)));

-

-      this->loop = saved_loop;

-      this->function = saved_function;

-   }

-

-   virtual void visit(class ir_function * ir)

-   {

-      visit_block(&ir->signatures);

-   }

-};

-

-bool

-do_lower_jumps(exec_list *instructions, bool pull_out_jumps, bool lower_sub_return, bool lower_main_return, bool lower_continue, bool lower_break)

-{

-   ir_lower_jumps_visitor v;

-   v.pull_out_jumps = pull_out_jumps;

-   v.lower_continue = lower_continue;

-   v.lower_break = lower_break;

-   v.lower_sub_return = lower_sub_return;

-   v.lower_main_return = lower_main_return;

-

-   do {

-      v.progress = false;

-      visit_exec_list(instructions, &v);

-   } while (v.progress);

-

-   return v.progress;

-}

+/*
+ * Copyright © 2010 Luca Barbieri
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and 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 lower_jumps.cpp
+ *
+ * This pass lowers jumps (break, continue, and return) to if/else structures.
+ *
+ * It can be asked to:
+ * 1. Pull jumps out of ifs where possible
+ * 2. Remove all "continue"s, replacing them with an "execute flag"
+ * 3. Replace all "break" with a single conditional one at the end of the loop
+ * 4. Replace all "return"s with a single return at the end of the function,
+ *    for the main function and/or other functions
+ *
+ * Applying this pass gives several benefits:
+ * 1. All functions can be inlined.
+ * 2. nv40 and other pre-DX10 chips without "continue" can be supported
+ * 3. nv30 and other pre-DX10 chips with no control flow at all are better
+ *    supported
+ *
+ * Continues are lowered by adding a per-loop "execute flag", initialized to
+ * true, that when cleared inhibits all execution until the end of the loop.
+ *
+ * Breaks are lowered to continues, plus setting a "break flag" that is checked
+ * at the end of the loop, and trigger the unique "break".
+ *
+ * Returns are lowered to breaks/continues, plus adding a "return flag" that
+ * causes loops to break again out of their enclosing loops until all the
+ * loops are exited: then the "execute flag" logic will ignore everything
+ * until the end of the function.
+ *
+ * Note that "continue" and "return" can also be implemented by adding
+ * a dummy loop and using break.
+ * However, this is bad for hardware with limited nesting depth, and
+ * prevents further optimization, and thus is not currently performed.
+ */
+
+#include "glsl_types.h"
+#include <string.h>
+#include "ir.h"
+
+/**
+ * Enum recording the result of analyzing how control flow might exit
+ * an IR node.
+ *
+ * Each possible value of jump_strength indicates a strictly stronger
+ * guarantee on control flow than the previous value.
+ *
+ * The ordering of strengths roughly reflects the way jumps are
+ * lowered: jumps with higher strength tend to be lowered to jumps of
+ * lower strength.  Accordingly, strength is used as a heuristic to
+ * determine which lowering to perform first.
+ *
+ * This enum is also used by get_jump_strength() to categorize
+ * instructions as either break, continue, return, or other.  When
+ * used in this fashion, strength_always_clears_execute_flag is not
+ * used.
+ *
+ * The control flow analysis made by this optimization pass makes two
+ * simplifying assumptions:
+ *
+ * - It ignores discard instructions, since they are lowered by a
+ *   separate pass (lower_discard.cpp).
+ *
+ * - It assumes it is always possible for control to flow from a loop
+ *   to the instruction immediately following it.  Technically, this
+ *   is not true (since all execution paths through the loop might
+ *   jump back to the top, or return from the function).
+ *
+ * Both of these simplifying assumtions are safe, since they can never
+ * cause reachable code to be incorrectly classified as unreachable;
+ * they can only do the opposite.
+ */
+enum jump_strength
+{
+   /**
+    * Analysis has produced no guarantee on how control flow might
+    * exit this IR node.  It might fall out the bottom (with or
+    * without clearing the execute flag, if present), or it might
+    * continue to the top of the innermost enclosing loop, break out
+    * of it, or return from the function.
+    */
+   strength_none,
+
+   /**
+    * The only way control can fall out the bottom of this node is
+    * through a code path that clears the execute flag.  It might also
+    * continue to the top of the innermost enclosing loop, break out
+    * of it, or return from the function.
+    */
+   strength_always_clears_execute_flag,
+
+   /**
+    * Control cannot fall out the bottom of this node.  It might
+    * continue to the top of the innermost enclosing loop, break out
+    * of it, or return from the function.
+    */
+   strength_continue,
+
+   /**
+    * Control cannot fall out the bottom of this node, or continue the
+    * top of the innermost enclosing loop.  It can only break out of
+    * it or return from the function.
+    */
+   strength_break,
+
+   /**
+    * Control cannot fall out the bottom of this node, continue to the
+    * top of the innermost enclosing loop, or break out of it.  It can
+    * only return from the function.
+    */
+   strength_return
+};
+
+struct block_record
+{
+   /* minimum jump strength (of lowered IR, not pre-lowering IR)
+    *
+    * If the block ends with a jump, must be the strength of the jump.
+    * Otherwise, the jump would be dead and have been deleted before)
+    *
+    * If the block doesn't end with a jump, it can be different than strength_none if all paths before it lead to some jump
+    * (e.g. an if with a return in one branch, and a break in the other, while not lowering them)
+    * Note that identical jumps are usually unified though.
+    */
+   jump_strength min_strength;
+
+   /* can anything clear the execute flag? */
+   bool may_clear_execute_flag;
+
+   block_record()
+   {
+      this->min_strength = strength_none;
+      this->may_clear_execute_flag = false;
+   }
+};
+
+struct loop_record
+{
+   ir_function_signature* signature;
+   ir_loop* loop;
+
+   /* used to avoid lowering the break used to represent lowered breaks */
+   unsigned nesting_depth;
+   bool in_if_at_the_end_of_the_loop;
+
+   bool may_set_return_flag;
+
+   ir_variable* break_flag;
+   ir_variable* execute_flag; /* cleared to emulate continue */
+
+   loop_record(ir_function_signature* p_signature = 0, ir_loop* p_loop = 0)
+   {
+      this->signature = p_signature;
+      this->loop = p_loop;
+      this->nesting_depth = 0;
+      this->in_if_at_the_end_of_the_loop = false;
+      this->may_set_return_flag = false;
+      this->break_flag = 0;
+      this->execute_flag = 0;
+   }
+
+   ir_variable* get_execute_flag()
+   {
+      /* also supported for the "function loop" */
+      if(!this->execute_flag) {
+         exec_list& list = this->loop ? this->loop->body_instructions : signature->body;
+         this->execute_flag = new(this->signature) ir_variable(glsl_type::bool_type, "execute_flag", ir_var_temporary);
+         list.push_head(new(this->signature) ir_assignment(new(this->signature) ir_dereference_variable(execute_flag), new(this->signature) ir_constant(true), 0));
+         list.push_head(this->execute_flag);
+      }
+      return this->execute_flag;
+   }
+
+   ir_variable* get_break_flag()
+   {
+      assert(this->loop);
+      if(!this->break_flag) {
+         this->break_flag = new(this->signature) ir_variable(glsl_type::bool_type, "break_flag", ir_var_temporary);
+         this->loop->insert_before(this->break_flag);
+         this->loop->insert_before(new(this->signature) ir_assignment(new(this->signature) ir_dereference_variable(break_flag), new(this->signature) ir_constant(false), 0));
+      }
+      return this->break_flag;
+   }
+};
+
+struct function_record
+{
+   ir_function_signature* signature;
+   ir_variable* return_flag; /* used to break out of all loops and then jump to the return instruction */
+   ir_variable* return_value;
+   bool lower_return;
+   unsigned nesting_depth;
+
+   function_record(ir_function_signature* p_signature = 0,
+                   bool lower_return = false)
+   {
+      this->signature = p_signature;
+      this->return_flag = 0;
+      this->return_value = 0;
+      this->nesting_depth = 0;
+      this->lower_return = lower_return;
+   }
+
+   ir_variable* get_return_flag()
+   {
+      if(!this->return_flag) {
+         this->return_flag = new(this->signature) ir_variable(glsl_type::bool_type, "return_flag", ir_var_temporary);
+         this->signature->body.push_head(new(this->signature) ir_assignment(new(this->signature) ir_dereference_variable(return_flag), new(this->signature) ir_constant(false), 0));
+         this->signature->body.push_head(this->return_flag);
+      }
+      return this->return_flag;
+   }
+
+   ir_variable* get_return_value()
+   {
+      if(!this->return_value) {
+         assert(!this->signature->return_type->is_void());
+         return_value = new(this->signature) ir_variable(this->signature->return_type, "return_value", ir_var_temporary);
+         this->signature->body.push_head(this->return_value);
+      }
+      return this->return_value;
+   }
+};
+
+struct ir_lower_jumps_visitor : public ir_control_flow_visitor {
+   /* Postconditions: on exit of any visit() function:
+    *
+    * ANALYSIS: this->block.min_strength,
+    * this->block.may_clear_execute_flag, and
+    * this->loop.may_set_return_flag are updated to reflect the
+    * characteristics of the visited statement.
+    *
+    * DEAD_CODE_ELIMINATION: If this->block.min_strength is not
+    * strength_none, the visited node is at the end of its exec_list.
+    * In other words, any unreachable statements that follow the
+    * visited statement in its exec_list have been removed.
+    *
+    * CONTAINED_JUMPS_LOWERED: If the visited statement contains other
+    * statements, then should_lower_jump() is false for all of the
+    * return, break, or continue statements it contains.
+    *
+    * Note that visiting a jump does not lower it.  That is the
+    * responsibility of the statement (or function signature) that
+    * contains the jump.
+    */
+
+   bool progress;
+
+   struct function_record function;
+   struct loop_record loop;
+   struct block_record block;
+
+   bool pull_out_jumps;
+   bool lower_continue;
+   bool lower_break;
+   bool lower_sub_return;
+   bool lower_main_return;
+
+   ir_lower_jumps_visitor()
+   {
+      this->progress = false;
+   }
+
+   void truncate_after_instruction(exec_node *ir)
+   {
+      if (!ir)
+         return;
+
+      while (!ir->get_next()->is_tail_sentinel()) {
+         ((ir_instruction *)ir->get_next())->remove();
+         this->progress = true;
+      }
+   }
+
+   void move_outer_block_inside(ir_instruction *ir, exec_list *inner_block)
+   {
+      while (!ir->get_next()->is_tail_sentinel()) {
+         ir_instruction *move_ir = (ir_instruction *)ir->get_next();
+
+         move_ir->remove();
+         inner_block->push_tail(move_ir);
+      }
+   }
+
+   /**
+    * Insert the instructions necessary to lower a return statement,
+    * before the given return instruction.
+    */
+   void insert_lowered_return(ir_return *ir)
+   {
+      ir_variable* return_flag = this->function.get_return_flag();
+      if(!this->function.signature->return_type->is_void()) {
+         ir_variable* return_value = this->function.get_return_value();
+         ir->insert_before(
+            new(ir) ir_assignment(
+               new (ir) ir_dereference_variable(return_value),
+               ir->value));
+      }
+      ir->insert_before(
+         new(ir) ir_assignment(
+            new (ir) ir_dereference_variable(return_flag),
+            new (ir) ir_constant(true)));
+      this->loop.may_set_return_flag = true;
+   }
+
+   /**
+    * If the given instruction is a return, lower it to instructions
+    * that store the return value (if there is one), set the return
+    * flag, and then break.
+    *
+    * It is safe to pass NULL to this function.
+    */
+   void lower_return_unconditionally(ir_instruction *ir)
+   {
+      if (get_jump_strength(ir) != strength_return) {
+         return;
+      }
+      insert_lowered_return((ir_return*)ir);
+      ir->replace_with(new(ir) ir_loop_jump(ir_loop_jump::jump_break));
+   }
+
+   /**
+    * Create the necessary instruction to replace a break instruction.
+    */
+   ir_instruction *create_lowered_break()
+   {
+      void *ctx = this->function.signature;
+      return new(ctx) ir_assignment(
+          new(ctx) ir_dereference_variable(this->loop.get_break_flag()),
+          new(ctx) ir_constant(true),
+          0);
+   }
+
+   /**
+    * If the given instruction is a break, lower it to an instruction
+    * that sets the break flag, without consulting
+    * should_lower_jump().
+    *
+    * It is safe to pass NULL to this function.
+    */
+   void lower_break_unconditionally(ir_instruction *ir)
+   {
+      if (get_jump_strength(ir) != strength_break) {
+         return;
+      }
+      ir->replace_with(create_lowered_break());
+   }
+
+   /**
+    * If the block ends in a conditional or unconditional break, lower
+    * it, even though should_lower_jump() says it needn't be lowered.
+    */
+   void lower_final_breaks(exec_list *block)
+   {
+      ir_instruction *ir = (ir_instruction *) block->get_tail();
+      lower_break_unconditionally(ir);
+      ir_if *ir_if = ir->as_if();
+      if (ir_if) {
+          lower_break_unconditionally(
+              (ir_instruction *) ir_if->then_instructions.get_tail());
+          lower_break_unconditionally(
+              (ir_instruction *) ir_if->else_instructions.get_tail());
+      }
+   }
+
+   virtual void visit(class ir_loop_jump * ir)
+   {
+      /* Eliminate all instructions after each one, since they are
+       * unreachable.  This satisfies the DEAD_CODE_ELIMINATION
+       * postcondition.
+       */
+      truncate_after_instruction(ir);
+
+      /* Set this->block.min_strength based on this instruction.  This
+       * satisfies the ANALYSIS postcondition.  It is not necessary to
+       * update this->block.may_clear_execute_flag or
+       * this->loop.may_set_return_flag, because an unlowered jump
+       * instruction can't change any flags.
+       */
+      this->block.min_strength = ir->is_break() ? strength_break : strength_continue;
+
+      /* The CONTAINED_JUMPS_LOWERED postcondition is already
+       * satisfied, because jump statements can't contain other
+       * statements.
+       */
+   }
+
+   virtual void visit(class ir_return * ir)
+   {
+      /* Eliminate all instructions after each one, since they are
+       * unreachable.  This satisfies the DEAD_CODE_ELIMINATION
+       * postcondition.
+       */
+      truncate_after_instruction(ir);
+
+      /* Set this->block.min_strength based on this instruction.  This
+       * satisfies the ANALYSIS postcondition.  It is not necessary to
+       * update this->block.may_clear_execute_flag or
+       * this->loop.may_set_return_flag, because an unlowered return
+       * instruction can't change any flags.
+       */
+      this->block.min_strength = strength_return;
+
+      /* The CONTAINED_JUMPS_LOWERED postcondition is already
+       * satisfied, because jump statements can't contain other
+       * statements.
+       */
+   }
+
+   virtual void visit(class ir_discard * ir)
+   {
+      /* Nothing needs to be done.  The ANALYSIS and
+       * DEAD_CODE_ELIMINATION postconditions are already satisfied,
+       * because discard statements are ignored by this optimization
+       * pass.  The CONTAINED_JUMPS_LOWERED postcondition is already
+       * satisfied, because discard statements can't contain other
+       * statements.
+       */
+   }
+
+   enum jump_strength get_jump_strength(ir_instruction* ir)
+   {
+      if(!ir)
+         return strength_none;
+      else if(ir->ir_type == ir_type_loop_jump) {
+         if(((ir_loop_jump*)ir)->is_break())
+            return strength_break;
+         else
+            return strength_continue;
+      } else if(ir->ir_type == ir_type_return)
+         return strength_return;
+      else
+         return strength_none;
+   }
+
+   bool should_lower_jump(ir_jump* ir)
+   {
+      unsigned strength = get_jump_strength(ir);
+      bool lower;
+      switch(strength)
+      {
+      case strength_none:
+         lower = false; /* don't change this, code relies on it */
+         break;
+      case strength_continue:
+         lower = lower_continue;
+         break;
+      case strength_break:
+         assert(this->loop.loop);
+         /* never lower "canonical break" */
+         if(ir->get_next()->is_tail_sentinel() && (this->loop.nesting_depth == 0
+               || (this->loop.nesting_depth == 1 && this->loop.in_if_at_the_end_of_the_loop)))
+            lower = false;
+         else
+            lower = lower_break;
+         break;
+      case strength_return:
+         /* never lower return at the end of a this->function */
+         if(this->function.nesting_depth == 0 && ir->get_next()->is_tail_sentinel())
+            lower = false;
+         else
+            lower = this->function.lower_return;
+         break;
+      }
+      return lower;
+   }
+
+   block_record visit_block(exec_list* list)
+   {
+      /* Note: since visiting a node may change that node's next
+       * pointer, we can't use visit_exec_list(), because
+       * visit_exec_list() caches the node's next pointer before
+       * visiting it.  So we use foreach_list() instead.
+       *
+       * foreach_list() isn't safe if the node being visited gets
+       * removed, but fortunately this visitor doesn't do that.
+       */
+
+      block_record saved_block = this->block;
+      this->block = block_record();
+      foreach_list(node, list) {
+         ((ir_instruction *) node)->accept(this);
+      }
+      block_record ret = this->block;
+      this->block = saved_block;
+      return ret;
+   }
+
+   virtual void visit(ir_if *ir)
+   {
+      if(this->loop.nesting_depth == 0 && ir->get_next()->is_tail_sentinel())
+         this->loop.in_if_at_the_end_of_the_loop = true;
+
+      ++this->function.nesting_depth;
+      ++this->loop.nesting_depth;
+
+      block_record block_records[2];
+      ir_jump* jumps[2];
+
+      /* Recursively lower nested jumps.  This satisfies the
+       * CONTAINED_JUMPS_LOWERED postcondition, except in the case of
+       * unconditional jumps at the end of ir->then_instructions and
+       * ir->else_instructions, which are handled below.
+       */
+      block_records[0] = visit_block(&ir->then_instructions);
+      block_records[1] = visit_block(&ir->else_instructions);
+
+retry: /* we get here if we put code after the if inside a branch */
+
+      /* Determine which of ir->then_instructions and
+       * ir->else_instructions end with an unconditional jump.
+       */
+      for(unsigned i = 0; i < 2; ++i) {
+         exec_list& list = i ? ir->else_instructions : ir->then_instructions;
+         jumps[i] = 0;
+         if(!list.is_empty() && get_jump_strength((ir_instruction*)list.get_tail()))
+            jumps[i] = (ir_jump*)list.get_tail();
+      }
+
+      /* Loop until we have satisfied the CONTAINED_JUMPS_LOWERED
+       * postcondition by lowering jumps in both then_instructions and
+       * else_instructions.
+       */
+      for(;;) {
+         /* Determine the types of the jumps that terminate
+          * ir->then_instructions and ir->else_instructions.
+          */
+         jump_strength jump_strengths[2];
+
+         for(unsigned i = 0; i < 2; ++i) {
+            if(jumps[i]) {
+               jump_strengths[i] = block_records[i].min_strength;
+               assert(jump_strengths[i] == get_jump_strength(jumps[i]));
+            } else
+               jump_strengths[i] = strength_none;
+         }
+
+         /* If both code paths end in a jump, and the jumps are the
+          * same, and we are pulling out jumps, replace them with a
+          * single jump that comes after the if instruction.  The new
+          * jump will be visited next, and it will be lowered if
+          * necessary by the loop or conditional that encloses it.
+          */
+         if(pull_out_jumps && jump_strengths[0] == jump_strengths[1]) {
+            bool unify = true;
+            if(jump_strengths[0] == strength_continue)
+               ir->insert_after(new(ir) ir_loop_jump(ir_loop_jump::jump_continue));
+            else if(jump_strengths[0] == strength_break)
+               ir->insert_after(new(ir) ir_loop_jump(ir_loop_jump::jump_break));
+            /* FINISHME: unify returns with identical expressions */
+            else if(jump_strengths[0] == strength_return && this->function.signature->return_type->is_void())
+               ir->insert_after(new(ir) ir_return(NULL));
+	    else
+	       unify = false;
+
+            if(unify) {
+               jumps[0]->remove();
+               jumps[1]->remove();
+               this->progress = true;
+
+               /* Update jumps[] to reflect the fact that the jumps
+                * are gone, and update block_records[] to reflect the
+                * fact that control can now flow to the next
+                * instruction.
+                */
+               jumps[0] = 0;
+               jumps[1] = 0;
+               block_records[0].min_strength = strength_none;
+               block_records[1].min_strength = strength_none;
+
+               /* The CONTAINED_JUMPS_LOWERED postcondition is now
+                * satisfied, so we can break out of the loop.
+                */
+               break;
+            }
+         }
+
+         /* lower a jump: if both need to lowered, start with the strongest one, so that
+          * we might later unify the lowered version with the other one
+          */
+         bool should_lower[2];
+         for(unsigned i = 0; i < 2; ++i)
+            should_lower[i] = should_lower_jump(jumps[i]);
+
+         int lower;
+         if(should_lower[1] && should_lower[0])
+            lower = jump_strengths[1] > jump_strengths[0];
+         else if(should_lower[0])
+            lower = 0;
+         else if(should_lower[1])
+            lower = 1;
+         else
+            /* Neither code path ends in a jump that needs to be
+             * lowered, so the CONTAINED_JUMPS_LOWERED postcondition
+             * is satisfied and we can break out of the loop.
+             */
+            break;
+
+         if(jump_strengths[lower] == strength_return) {
+            /* To lower a return, we create a return flag (if the
+             * function doesn't have one already) and add instructions
+             * that: 1. store the return value (if this function has a
+             * non-void return) and 2. set the return flag
+             */
+            insert_lowered_return((ir_return*)jumps[lower]);
+            if(this->loop.loop) {
+               /* If we are in a loop, replace the return instruction
+                * with a break instruction, and then loop so that the
+                * break instruction can be lowered if necessary.
+                */
+               ir_loop_jump* lowered = 0;
+               lowered = new(ir) ir_loop_jump(ir_loop_jump::jump_break);
+               /* Note: we must update block_records and jumps to
+                * reflect the fact that the control path has been
+                * altered from a return to a break.
+                */
+               block_records[lower].min_strength = strength_break;
+               jumps[lower]->replace_with(lowered);
+               jumps[lower] = lowered;
+            } else {
+               /* If we are not in a loop, we then proceed as we would
+                * for a continue statement (set the execute flag to
+                * false to prevent the rest of the function from
+                * executing).
+                */
+               goto lower_continue;
+            }
+            this->progress = true;
+         } else if(jump_strengths[lower] == strength_break) {
+            /* To lower a break, we create a break flag (if the loop
+             * doesn't have one already) and add an instruction that
+             * sets it.
+             *
+             * Then we proceed as we would for a continue statement
+             * (set the execute flag to false to prevent the rest of
+             * the loop body from executing).
+             *
+             * The visit() function for the loop will ensure that the
+             * break flag is checked after executing the loop body.
+             */
+            jumps[lower]->insert_before(create_lowered_break());
+            goto lower_continue;
+         } else if(jump_strengths[lower] == strength_continue) {
+lower_continue:
+            /* To lower a continue, we create an execute flag (if the
+             * loop doesn't have one already) and replace the continue
+             * with an instruction that clears it.
+             *
+             * Note that this code path gets exercised when lowering
+             * return statements that are not inside a loop, so
+             * this->loop must be initialized even outside of loops.
+             */
+            ir_variable* execute_flag = this->loop.get_execute_flag();
+            jumps[lower]->replace_with(new(ir) ir_assignment(new (ir) ir_dereference_variable(execute_flag), new (ir) ir_constant(false), 0));
+            /* Note: we must update block_records and jumps to reflect
+             * the fact that the control path has been altered to an
+             * instruction that clears the execute flag.
+             */
+            jumps[lower] = 0;
+            block_records[lower].min_strength = strength_always_clears_execute_flag;
+            block_records[lower].may_clear_execute_flag = true;
+            this->progress = true;
+
+            /* Let the loop run again, in case the other branch of the
+             * if needs to be lowered too.
+             */
+         }
+      }
+
+      /* move out a jump out if possible */
+      if(pull_out_jumps) {
+         /* If one of the branches ends in a jump, and control cannot
+          * fall out the bottom of the other branch, then we can move
+          * the jump after the if.
+          *
+          * Set move_out to the branch we are moving a jump out of.
+          */
+         int move_out = -1;
+         if(jumps[0] && block_records[1].min_strength >= strength_continue)
+            move_out = 0;
+         else if(jumps[1] && block_records[0].min_strength >= strength_continue)
+            move_out = 1;
+
+         if(move_out >= 0)
+         {
+            jumps[move_out]->remove();
+            ir->insert_after(jumps[move_out]);
+            /* Note: we must update block_records and jumps to reflect
+             * the fact that the jump has been moved out of the if.
+             */
+            jumps[move_out] = 0;
+            block_records[move_out].min_strength = strength_none;
+            this->progress = true;
+         }
+      }
+
+      /* Now satisfy the ANALYSIS postcondition by setting
+       * this->block.min_strength and
+       * this->block.may_clear_execute_flag based on the
+       * characteristics of the two branches.
+       */
+      if(block_records[0].min_strength < block_records[1].min_strength)
+         this->block.min_strength = block_records[0].min_strength;
+      else
+         this->block.min_strength = block_records[1].min_strength;
+      this->block.may_clear_execute_flag = this->block.may_clear_execute_flag || block_records[0].may_clear_execute_flag || block_records[1].may_clear_execute_flag;
+
+      /* Now we need to clean up the instructions that follow the
+       * if.
+       *
+       * If those instructions are unreachable, then satisfy the
+       * DEAD_CODE_ELIMINATION postcondition by eliminating them.
+       * Otherwise that postcondition is already satisfied.
+       */
+      if(this->block.min_strength)
+         truncate_after_instruction(ir);
+      else if(this->block.may_clear_execute_flag)
+      {
+         /* If the "if" instruction might clear the execute flag, then
+          * we need to guard any instructions that follow so that they
+          * are only executed if the execute flag is set.
+          *
+          * If one of the branches of the "if" always clears the
+          * execute flag, and the other branch never clears it, then
+          * this is easy: just move all the instructions following the
+          * "if" into the branch that never clears it.
+          */
+         int move_into = -1;
+         if(block_records[0].min_strength && !block_records[1].may_clear_execute_flag)
+            move_into = 1;
+         else if(block_records[1].min_strength && !block_records[0].may_clear_execute_flag)
+            move_into = 0;
+
+         if(move_into >= 0) {
+            assert(!block_records[move_into].min_strength && !block_records[move_into].may_clear_execute_flag); /* otherwise, we just truncated */
+
+            exec_list* list = move_into ? &ir->else_instructions : &ir->then_instructions;
+            exec_node* next = ir->get_next();
+            if(!next->is_tail_sentinel()) {
+               move_outer_block_inside(ir, list);
+
+               /* If any instructions moved, then we need to visit
+                * them (since they are now inside the "if").  Since
+                * block_records[move_into] is in its default state
+                * (see assertion above), we can safely replace
+                * block_records[move_into] with the result of this
+                * analysis.
+                */
+               exec_list list;
+               list.head = next;
+               block_records[move_into] = visit_block(&list);
+
+               /*
+                * Then we need to re-start our jump lowering, since one
+                * of the instructions we moved might be a jump that
+                * needs to be lowered.
+                */
+               this->progress = true;
+               goto retry;
+            }
+         } else {
+            /* If we get here, then the simple case didn't apply; we
+             * need to actually guard the instructions that follow.
+             *
+             * To avoid creating unnecessarily-deep nesting, first
+             * look through the instructions that follow and unwrap
+             * any instructions that that are already wrapped in the
+             * appropriate guard.
+             */
+            ir_instruction* ir_after;
+            for(ir_after = (ir_instruction*)ir->get_next(); !ir_after->is_tail_sentinel();)
+            {
+               ir_if* ir_if = ir_after->as_if();
+               if(ir_if && ir_if->else_instructions.is_empty()) {
+                  ir_dereference_variable* ir_if_cond_deref = ir_if->condition->as_dereference_variable();
+                  if(ir_if_cond_deref && ir_if_cond_deref->var == this->loop.execute_flag) {
+                     ir_instruction* ir_next = (ir_instruction*)ir_after->get_next();
+                     ir_after->insert_before(&ir_if->then_instructions);
+                     ir_after->remove();
+                     ir_after = ir_next;
+                     continue;
+                  }
+               }
+               ir_after = (ir_instruction*)ir_after->get_next();
+
+               /* only set this if we find any unprotected instruction */
+               this->progress = true;
+            }
+
+            /* Then, wrap all the instructions that follow in a single
+             * guard.
+             */
+            if(!ir->get_next()->is_tail_sentinel()) {
+               assert(this->loop.execute_flag);
+               ir_if* if_execute = new(ir) ir_if(new(ir) ir_dereference_variable(this->loop.execute_flag));
+               move_outer_block_inside(ir, &if_execute->then_instructions);
+               ir->insert_after(if_execute);
+            }
+         }
+      }
+      --this->loop.nesting_depth;
+      --this->function.nesting_depth;
+   }
+
+   virtual void visit(ir_loop *ir)
+   {
+      /* Visit the body of the loop, with a fresh data structure in
+       * this->loop so that the analysis we do here won't bleed into
+       * enclosing loops.
+       *
+       * We assume that all code after a loop is reachable from the
+       * loop (see comments on enum jump_strength), so the
+       * DEAD_CODE_ELIMINATION postcondition is automatically
+       * satisfied, as is the block.min_strength portion of the
+       * ANALYSIS postcondition.
+       *
+       * The block.may_clear_execute_flag portion of the ANALYSIS
+       * postcondition is automatically satisfied because execute
+       * flags do not propagate outside of loops.
+       *
+       * The loop.may_set_return_flag portion of the ANALYSIS
+       * postcondition is handled below.
+       */
+      ++this->function.nesting_depth;
+      loop_record saved_loop = this->loop;
+      this->loop = loop_record(this->function.signature, ir);
+
+      /* Recursively lower nested jumps.  This satisfies the
+       * CONTAINED_JUMPS_LOWERED postcondition, except in the case of
+       * an unconditional continue or return at the bottom of the
+       * loop, which are handled below.
+       */
+      block_record body = visit_block(&ir->body_instructions);
+
+      /* If the loop ends in an unconditional continue, eliminate it
+       * because it is redundant.
+       */
+      ir_instruction *ir_last
+         = (ir_instruction *) ir->body_instructions.get_tail();
+      if (get_jump_strength(ir_last) == strength_continue) {
+         ir_last->remove();
+      }
+
+      /* If the loop ends in an unconditional return, and we are
+       * lowering returns, lower it.
+       */
+      if (this->function.lower_return)
+         lower_return_unconditionally(ir_last);
+
+      if(body.min_strength >= strength_break) {
+         /* FINISHME: If the min_strength of the loop body is
+          * strength_break or strength_return, that means that it
+          * isn't a loop at all, since control flow always leaves the
+          * body of the loop via break or return.  In principle the
+          * loop could be eliminated in this case.  This optimization
+          * is not implemented yet.
+          */
+      }
+
+      if(this->loop.break_flag) {
+         /* We only get here if we are lowering breaks */
+         assert (lower_break);
+
+         /* If a break flag was generated while visiting the body of
+          * the loop, then at least one break was lowered, so we need
+          * to generate an if statement at the end of the loop that
+          * does a "break" if the break flag is set.  The break we
+          * generate won't violate the CONTAINED_JUMPS_LOWERED
+          * postcondition, because should_lower_jump() always returns
+          * false for a break that happens at the end of a loop.
+          *
+          * However, if the loop already ends in a conditional or
+          * unconditional break, then we need to lower that break,
+          * because it won't be at the end of the loop anymore.
+          */
+         lower_final_breaks(&ir->body_instructions);
+
+         ir_if* break_if = new(ir) ir_if(new(ir) ir_dereference_variable(this->loop.break_flag));
+         break_if->then_instructions.push_tail(new(ir) ir_loop_jump(ir_loop_jump::jump_break));
+         ir->body_instructions.push_tail(break_if);
+      }
+
+      /* If the body of the loop may set the return flag, then at
+       * least one return was lowered to a break, so we need to ensure
+       * that the return flag is checked after the body of the loop is
+       * executed.
+       */
+      if(this->loop.may_set_return_flag) {
+         assert(this->function.return_flag);
+         /* Generate the if statement to check the return flag */
+         ir_if* return_if = new(ir) ir_if(new(ir) ir_dereference_variable(this->function.return_flag));
+         /* Note: we also need to propagate the knowledge that the
+          * return flag may get set to the outer context.  This
+          * satisfies the loop.may_set_return_flag part of the
+          * ANALYSIS postcondition.
+          */
+         saved_loop.may_set_return_flag = true;
+         if(saved_loop.loop)
+            /* If this loop is nested inside another one, then the if
+             * statement that we generated should break out of that
+             * loop if the return flag is set.  Caller will lower that
+             * break statement if necessary.
+             */
+            return_if->then_instructions.push_tail(new(ir) ir_loop_jump(ir_loop_jump::jump_break));
+         else
+            /* Otherwise, all we need to do is ensure that the
+             * instructions that follow are only executed if the
+             * return flag is clear.  We can do that by moving those
+             * instructions into the else clause of the generated if
+             * statement.
+             */
+            move_outer_block_inside(ir, &return_if->else_instructions);
+         ir->insert_after(return_if);
+      }
+
+      this->loop = saved_loop;
+      --this->function.nesting_depth;
+   }
+
+   virtual void visit(ir_function_signature *ir)
+   {
+      /* these are not strictly necessary */
+      assert(!this->function.signature);
+      assert(!this->loop.loop);
+
+      bool lower_return;
+      if (strcmp(ir->function_name(), "main") == 0)
+         lower_return = lower_main_return;
+      else
+         lower_return = lower_sub_return;
+
+      function_record saved_function = this->function;
+      loop_record saved_loop = this->loop;
+      this->function = function_record(ir, lower_return);
+      this->loop = loop_record(ir);
+
+      assert(!this->loop.loop);
+
+      /* Visit the body of the function to lower any jumps that occur
+       * in it, except possibly an unconditional return statement at
+       * the end of it.
+       */
+      visit_block(&ir->body);
+
+      /* If the body ended in an unconditional return of non-void,
+       * then we don't need to lower it because it's the one canonical
+       * return.
+       *
+       * If the body ended in a return of void, eliminate it because
+       * it is redundant.
+       */
+      if (ir->return_type->is_void() &&
+          get_jump_strength((ir_instruction *) ir->body.get_tail())) {
+         ir_jump *jump = (ir_jump *) ir->body.get_tail();
+         assert (jump->ir_type == ir_type_return);
+         jump->remove();
+      }
+
+      if(this->function.return_value)
+         ir->body.push_tail(new(ir) ir_return(new (ir) ir_dereference_variable(this->function.return_value)));
+
+      this->loop = saved_loop;
+      this->function = saved_function;
+   }
+
+   virtual void visit(class ir_function * ir)
+   {
+      visit_block(&ir->signatures);
+   }
+};
+
+bool
+do_lower_jumps(exec_list *instructions, bool pull_out_jumps, bool lower_sub_return, bool lower_main_return, bool lower_continue, bool lower_break)
+{
+   ir_lower_jumps_visitor v;
+   v.pull_out_jumps = pull_out_jumps;
+   v.lower_continue = lower_continue;
+   v.lower_break = lower_break;
+   v.lower_sub_return = lower_sub_return;
+   v.lower_main_return = lower_main_return;
+
+   do {
+      v.progress = false;
+      visit_exec_list(instructions, &v);
+   } while (v.progress);
+
+   return v.progress;
+}
diff --git a/mesalib/src/mesa/main/depthstencil.c b/mesalib/src/mesa/main/depthstencil.c
index ab62c97fe..4d0600050 100644
--- a/mesalib/src/mesa/main/depthstencil.c
+++ b/mesalib/src/mesa/main/depthstencil.c
@@ -393,6 +393,217 @@ _mesa_new_z24_renderbuffer_wrapper(struct gl_context *ctx,
 }
 
 
+static void
+get_row_z32f(struct gl_context *ctx, struct gl_renderbuffer *z32frb, GLuint count,
+             GLint x, GLint y, void *values)
+{
+   struct gl_renderbuffer *dsrb = z32frb->Wrapped;
+   GLfloat temp[MAX_WIDTH*2];
+   GLfloat *dst = (GLfloat *) values;
+   const GLfloat *src = (const GLfloat *) dsrb->GetPointer(ctx, dsrb, x, y);
+   GLuint i;
+   ASSERT(z32frb->DataType == GL_FLOAT);
+   ASSERT(dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+   ASSERT(dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   if (!src) {
+      dsrb->GetRow(ctx, dsrb, count, x, y, temp);
+      src = temp;
+   }
+   for (i = 0; i < count; i++) {
+      dst[i] = src[i*2];
+   }
+}
+
+static void
+get_values_z32f(struct gl_context *ctx, struct gl_renderbuffer *z32frb, GLuint count,
+                const GLint x[], const GLint y[], void *values)
+{
+   struct gl_renderbuffer *dsrb = z32frb->Wrapped;
+   GLfloat temp[MAX_WIDTH*2];
+   GLfloat *dst = (GLfloat *) values;
+   GLuint i;
+   ASSERT(z32frb->DataType == GL_FLOAT);
+   ASSERT(dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+   ASSERT(dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   ASSERT(count <= MAX_WIDTH);
+   /* don't bother trying direct access */
+   dsrb->GetValues(ctx, dsrb, count, x, y, temp);
+   for (i = 0; i < count; i++) {
+      dst[i] = temp[i*2];
+   }
+}
+
+static void
+put_row_z32f(struct gl_context *ctx, struct gl_renderbuffer *z32frb, GLuint count,
+             GLint x, GLint y, const void *values, const GLubyte *mask)
+{
+   struct gl_renderbuffer *dsrb = z32frb->Wrapped;
+   const GLfloat *src = (const GLfloat *) values;
+   GLfloat *dst = (GLfloat *) dsrb->GetPointer(ctx, dsrb, x, y);
+   ASSERT(z32frb->DataType == GL_FLOAT);
+   ASSERT(dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+   ASSERT(dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   if (dst) {
+      /* direct access */
+      GLuint i;
+      for (i = 0; i < count; i++) {
+         if (!mask || mask[i]) {
+            dst[i*2] = src[i];
+         }
+      }
+   }
+   else {
+      /* get, modify, put */
+      GLfloat temp[MAX_WIDTH*2];
+      GLuint i;
+      dsrb->GetRow(ctx, dsrb, count, x, y, temp);
+      for (i = 0; i < count; i++) {
+         if (!mask || mask[i]) {
+            temp[i*2] = src[i];
+         }
+      }
+      dsrb->PutRow(ctx, dsrb, count, x, y, temp, mask);
+   }
+}
+
+static void
+put_mono_row_z32f(struct gl_context *ctx, struct gl_renderbuffer *z32frb, GLuint count,
+                  GLint x, GLint y, const void *value, const GLubyte *mask)
+{
+   struct gl_renderbuffer *dsrb = z32frb->Wrapped;
+   GLfloat *dst = (GLfloat *) dsrb->GetPointer(ctx, dsrb, x, y);
+   ASSERT(z32frb->DataType == GL_FLOAT);
+   ASSERT(dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+   ASSERT(dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   if (dst) {
+      /* direct access */
+      GLuint i;
+      const GLfloat val = *(GLfloat*)value;
+      for (i = 0; i < count; i++) {
+         if (!mask || mask[i]) {
+            dst[i*2] = val;
+         }
+      }
+   }
+   else {
+      /* get, modify, put */
+      GLfloat temp[MAX_WIDTH*2];
+      GLuint i;
+      const GLfloat val = *(GLfloat *)value;
+      dsrb->GetRow(ctx, dsrb, count, x, y, temp);
+      for (i = 0; i < count; i++) {
+         if (!mask || mask[i]) {
+            temp[i*2] = val;
+         }
+      }
+      dsrb->PutRow(ctx, dsrb, count, x, y, temp, mask);
+   }
+}
+
+static void
+put_values_z32f(struct gl_context *ctx, struct gl_renderbuffer *z32frb, GLuint count,
+                const GLint x[], const GLint y[],
+                const void *values, const GLubyte *mask)
+{
+   struct gl_renderbuffer *dsrb = z32frb->Wrapped;
+   const GLfloat *src = (const GLfloat *) values;
+   ASSERT(z32frb->DataType == GL_FLOAT);
+   ASSERT(dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+   ASSERT(dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   if (dsrb->GetPointer(ctx, dsrb, 0, 0)) {
+      /* direct access */
+      GLuint i;
+      for (i = 0; i < count; i++) {
+         if (!mask || mask[i]) {
+            GLfloat *dst = (GLfloat *) dsrb->GetPointer(ctx, dsrb, x[i], y[i]);
+            *dst = src[i];
+         }
+      }
+   }
+   else {
+      /* get, modify, put */
+      GLfloat temp[MAX_WIDTH*2];
+      GLuint i;
+      dsrb->GetValues(ctx, dsrb, count, x, y, temp);
+      for (i = 0; i < count; i++) {
+         if (!mask || mask[i]) {
+            temp[i*2] = src[i];
+         }
+      }
+      dsrb->PutValues(ctx, dsrb, count, x, y, temp, mask);
+   }
+}
+
+static void
+put_mono_values_z32f(struct gl_context *ctx, struct gl_renderbuffer *z32frb,
+                     GLuint count, const GLint x[], const GLint y[],
+                     const void *value, const GLubyte *mask)
+{
+   struct gl_renderbuffer *dsrb = z32frb->Wrapped;
+   GLfloat temp[MAX_WIDTH*2];
+   GLuint i;
+   const GLfloat val = *(GLfloat *)value;
+   ASSERT(z32frb->DataType == GL_FLOAT);
+   ASSERT(dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+   ASSERT(dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   /* get, modify, put */
+   dsrb->GetValues(ctx, dsrb, count, x, y, temp);
+   for (i = 0; i < count; i++) {
+      if (!mask || mask[i]) {
+         temp[i*2] = val;
+      }
+   }
+   dsrb->PutValues(ctx, dsrb, count, x, y, temp, mask);
+}
+
+
+/**
+ * Wrap the given GL_DEPTH_STENCIL renderbuffer so that it acts like
+ * a depth renderbuffer.
+ * \return new depth renderbuffer
+ */
+struct gl_renderbuffer *
+_mesa_new_z32f_renderbuffer_wrapper(struct gl_context *ctx,
+                                    struct gl_renderbuffer *dsrb)
+{
+   struct gl_renderbuffer *z32frb;
+
+   ASSERT(dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   ASSERT(dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+
+   z32frb = ctx->Driver.NewRenderbuffer(ctx, 0);
+   if (!z32frb)
+      return NULL;
+
+   /* NOTE: need to do manual refcounting here */
+   z32frb->Wrapped = dsrb;
+   dsrb->RefCount++;
+
+   z32frb->Name = dsrb->Name;
+   z32frb->RefCount = 0;
+   z32frb->Width = dsrb->Width;
+   z32frb->Height = dsrb->Height;
+   z32frb->RowStride = dsrb->RowStride;
+   z32frb->InternalFormat = GL_DEPTH_COMPONENT32F;
+   z32frb->Format = MESA_FORMAT_Z32_FLOAT;
+   z32frb->_BaseFormat = GL_DEPTH_COMPONENT;
+   z32frb->DataType = GL_FLOAT;
+   z32frb->Data = NULL;
+   z32frb->Delete = delete_wrapper;
+   z32frb->AllocStorage = alloc_wrapper_storage;
+   z32frb->GetPointer = nop_get_pointer;
+   z32frb->GetRow = get_row_z32f;
+   z32frb->GetValues = get_values_z32f;
+   z32frb->PutRow = put_row_z32f;
+   z32frb->PutRowRGB = NULL;
+   z32frb->PutMonoRow = put_mono_row_z32f;
+   z32frb->PutValues = put_values_z32f;
+   z32frb->PutMonoValues = put_mono_values_z32f;
+
+   return z32frb;
+}
+
+
 /*======================================================================
  * Stencil wrapper around depth/stencil renderbuffer
  */
@@ -402,16 +613,22 @@ get_row_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint count,
            GLint x, GLint y, void *values)
 {
    struct gl_renderbuffer *dsrb = s8rb->Wrapped;
-   GLuint temp[MAX_WIDTH], i;
+   GLuint temp[MAX_WIDTH*2], i;
    GLubyte *dst = (GLubyte *) values;
    const GLuint *src = (const GLuint *) dsrb->GetPointer(ctx, dsrb, x, y);
    ASSERT(s8rb->DataType == GL_UNSIGNED_BYTE);
-   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT);
+   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT ||
+          dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
    if (!src) {
       dsrb->GetRow(ctx, dsrb, count, x, y, temp);
       src = temp;
    }
-   if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+   if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+      for (i = 0; i < count; i++) {
+         dst[i] = src[i*2+1] & 0xff;
+      }
+   }
+   else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
       for (i = 0; i < count; i++) {
          dst[i] = src[i] & 0xff;
       }
@@ -429,14 +646,20 @@ get_values_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint count
               const GLint x[], const GLint y[], void *values)
 {
    struct gl_renderbuffer *dsrb = s8rb->Wrapped;
-   GLuint temp[MAX_WIDTH], i;
+   GLuint temp[MAX_WIDTH*2], i;
    GLubyte *dst = (GLubyte *) values;
    ASSERT(s8rb->DataType == GL_UNSIGNED_BYTE);
-   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT);
+   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT ||
+          dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
    ASSERT(count <= MAX_WIDTH);
    /* don't bother trying direct access */
    dsrb->GetValues(ctx, dsrb, count, x, y, temp);
-   if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+   if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+      for (i = 0; i < count; i++) {
+         dst[i] = temp[i*2+1] & 0xff;
+      }
+   }
+   else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
       for (i = 0; i < count; i++) {
          dst[i] = temp[i] & 0xff;
       }
@@ -457,11 +680,19 @@ put_row_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint count,
    const GLubyte *src = (const GLubyte *) values;
    GLuint *dst = (GLuint *) dsrb->GetPointer(ctx, dsrb, x, y);
    ASSERT(s8rb->DataType == GL_UNSIGNED_BYTE);
-   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT);
+   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT ||
+          dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
    if (dst) {
       /* direct access */
       GLuint i;
-      if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+      if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               dst[i*2+1] = src[i];
+            }
+         }
+      }
+      else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                dst[i] = (dst[i] & 0xffffff00) | src[i];
@@ -479,9 +710,16 @@ put_row_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint count,
    }
    else {
       /* get, modify, put */
-      GLuint temp[MAX_WIDTH], i;
+      GLuint temp[MAX_WIDTH*2], i;
       dsrb->GetRow(ctx, dsrb, count, x, y, temp);
-      if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+      if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               temp[i*2+1] = src[i];
+            }
+         }
+      }
+      else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                temp[i] = (temp[i] & 0xffffff00) | src[i];
@@ -508,11 +746,19 @@ put_mono_row_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint cou
    const GLubyte val = *((GLubyte *) value);
    GLuint *dst = (GLuint *) dsrb->GetPointer(ctx, dsrb, x, y);
    ASSERT(s8rb->DataType == GL_UNSIGNED_BYTE);
-   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT);
+   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT ||
+          dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
    if (dst) {
       /* direct access */
       GLuint i;
-      if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+      if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               dst[i*2+1] = val;
+            }
+         }
+      }
+      else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                dst[i] = (dst[i] & 0xffffff00) | val;
@@ -530,9 +776,16 @@ put_mono_row_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint cou
    }
    else {
       /* get, modify, put */
-      GLuint temp[MAX_WIDTH], i;
+      GLuint temp[MAX_WIDTH*2], i;
       dsrb->GetRow(ctx, dsrb, count, x, y, temp);
-      if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+      if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               temp[i*2+1] = val;
+            }
+         }
+      }
+      else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                temp[i] = (temp[i] & 0xffffff00) | val;
@@ -559,11 +812,20 @@ put_values_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint count
    struct gl_renderbuffer *dsrb = s8rb->Wrapped;
    const GLubyte *src = (const GLubyte *) values;
    ASSERT(s8rb->DataType == GL_UNSIGNED_BYTE);
-   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT);
+   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT ||
+          dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
    if (dsrb->GetPointer(ctx, dsrb, 0, 0)) {
       /* direct access */
       GLuint i;
-      if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+      if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               GLuint *dst = (GLuint *) dsrb->GetPointer(ctx, dsrb, x[i], y[i]);
+               dst[1] = src[i];
+            }
+         }
+      }
+      else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                GLuint *dst = (GLuint *) dsrb->GetPointer(ctx, dsrb, x[i], y[i]);
@@ -583,9 +845,16 @@ put_values_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint count
    }
    else {
       /* get, modify, put */
-      GLuint temp[MAX_WIDTH], i;
+      GLuint temp[MAX_WIDTH*2], i;
       dsrb->GetValues(ctx, dsrb, count, x, y, temp);
-      if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+      if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               temp[i*2+1] = src[i];
+            }
+         }
+      }
+      else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                temp[i] = (temp[i] & 0xffffff00) | src[i];
@@ -610,11 +879,18 @@ put_mono_values_s8(struct gl_context *ctx, struct gl_renderbuffer *s8rb, GLuint
                    const void *value, const GLubyte *mask)
 {
    struct gl_renderbuffer *dsrb = s8rb->Wrapped;
-   GLuint temp[MAX_WIDTH], i;
+   GLuint temp[MAX_WIDTH*2], i;
    const GLubyte val = *((GLubyte *) value);
    /* get, modify, put */
    dsrb->GetValues(ctx, dsrb, count, x, y, temp);
-   if (dsrb->Format == MESA_FORMAT_Z24_S8) {
+   if (dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+      for (i = 0; i < count; i++) {
+         if (!mask || mask[i]) {
+            temp[i*2+1] = val;
+         }
+      }
+   }
+   else if (dsrb->Format == MESA_FORMAT_Z24_S8) {
       for (i = 0; i < count; i++) {
          if (!mask || mask[i]) {
             temp[i] = (temp[i] & 0xffffff00) | val;
@@ -644,8 +920,10 @@ _mesa_new_s8_renderbuffer_wrapper(struct gl_context *ctx, struct gl_renderbuffer
    struct gl_renderbuffer *s8rb;
 
    ASSERT(dsrb->Format == MESA_FORMAT_Z24_S8 ||
-          dsrb->Format == MESA_FORMAT_S8_Z24);
-   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT);
+          dsrb->Format == MESA_FORMAT_S8_Z24 ||
+          dsrb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
+   ASSERT(dsrb->DataType == GL_UNSIGNED_INT_24_8_EXT ||
+          dsrb->DataType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
    s8rb = ctx->Driver.NewRenderbuffer(ctx, 0);
    if (!s8rb)
diff --git a/mesalib/src/mesa/main/depthstencil.h b/mesalib/src/mesa/main/depthstencil.h
index a86f54d96..b47a2e482 100644
--- a/mesalib/src/mesa/main/depthstencil.h
+++ b/mesalib/src/mesa/main/depthstencil.h
@@ -1,57 +1,62 @@
-/*

- * 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.

- */

-

-

-#ifndef DEPTHSTENCIL_H

-#define DEPTHSTENCIL_H

-

-struct gl_context;

-

-extern struct gl_renderbuffer *

-_mesa_new_z24_renderbuffer_wrapper(struct gl_context *ctx,

-                                   struct gl_renderbuffer *dsrb);

-

-

-extern struct gl_renderbuffer *

-_mesa_new_s8_renderbuffer_wrapper(struct gl_context *ctx,

-                                  struct gl_renderbuffer *dsrb);

-

-

-extern void

-_mesa_extract_stencil(struct gl_context *ctx,

-                      struct gl_renderbuffer *dsRb,

-                      struct gl_renderbuffer *stencilRb);

-

-

-extern void

-_mesa_insert_stencil(struct gl_context *ctx,

-                     struct gl_renderbuffer *dsRb,

-                     struct gl_renderbuffer *stencilRb);

-

-

-extern void

-_mesa_promote_stencil(struct gl_context *ctx, struct gl_renderbuffer *stencilRb);

-

-

-#endif /* DEPTHSTENCIL_H */

+/*
+ * 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.
+ */
+
+
+#ifndef DEPTHSTENCIL_H
+#define DEPTHSTENCIL_H
+
+struct gl_context;
+
+extern struct gl_renderbuffer *
+_mesa_new_z24_renderbuffer_wrapper(struct gl_context *ctx,
+                                   struct gl_renderbuffer *dsrb);
+
+
+extern struct gl_renderbuffer *
+_mesa_new_z32f_renderbuffer_wrapper(struct gl_context *ctx,
+                                    struct gl_renderbuffer *dsrb);
+
+
+extern struct gl_renderbuffer *
+_mesa_new_s8_renderbuffer_wrapper(struct gl_context *ctx,
+                                  struct gl_renderbuffer *dsrb);
+
+
+extern void
+_mesa_extract_stencil(struct gl_context *ctx,
+                      struct gl_renderbuffer *dsRb,
+                      struct gl_renderbuffer *stencilRb);
+
+
+extern void
+_mesa_insert_stencil(struct gl_context *ctx,
+                     struct gl_renderbuffer *dsRb,
+                     struct gl_renderbuffer *stencilRb);
+
+
+extern void
+_mesa_promote_stencil(struct gl_context *ctx, struct gl_renderbuffer *stencilRb);
+
+
+#endif /* DEPTHSTENCIL_H */
diff --git a/mesalib/src/mesa/main/fbobject.c b/mesalib/src/mesa/main/fbobject.c
index 8cc3fd49a..84969360d 100644
--- a/mesalib/src/mesa/main/fbobject.c
+++ b/mesalib/src/mesa/main/fbobject.c
@@ -1131,6 +1131,16 @@ _mesa_base_fbo_format(struct gl_context *ctx, GLenum internalFormat)
          return GL_DEPTH_STENCIL_EXT;
       else
          return 0;
+   case GL_DEPTH_COMPONENT32F:
+      if (ctx->Extensions.ARB_depth_buffer_float)
+         return GL_DEPTH_COMPONENT;
+      else
+         return 0;
+   case GL_DEPTH32F_STENCIL8:
+      if (ctx->Extensions.ARB_depth_buffer_float)
+         return GL_DEPTH_STENCIL;
+      else
+         return 0;
    case GL_RED:
    case GL_R8:
    case GL_R16:
@@ -2266,6 +2276,15 @@ _mesa_GetFramebufferAttachmentParameterivEXT(GLenum target, GLenum attachment,
             /* special cases */
             *params = GL_INDEX;
          }
+         else if (format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+            /* depends on the attachment parameter */
+            if (attachment == GL_STENCIL_ATTACHMENT) {
+               *params = GL_INDEX;
+            }
+            else {
+               *params = GL_FLOAT;
+            }
+         }
          else {
             *params = _mesa_get_format_datatype(format);
          }
@@ -2584,6 +2603,10 @@ _mesa_BlitFramebufferEXT(GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
       }
    }
 
+   if (!mask) {
+      return;
+   }
+
    ASSERT(ctx->Driver.BlitFramebuffer);
    ctx->Driver.BlitFramebuffer(ctx,
                                srcX0, srcY0, srcX1, srcY1,
diff --git a/mesalib/src/mesa/main/formats.c b/mesalib/src/mesa/main/formats.c
index e88ba4397..f58b19756 100644
--- a/mesalib/src/mesa/main/formats.c
+++ b/mesalib/src/mesa/main/formats.c
@@ -1091,6 +1091,25 @@ static struct gl_format_info format_info[MESA_FORMAT_COUNT] =
       0, 0, 0, 0, 0,
       1, 1, 4
    },
+   /* ARB_depth_buffer_float */
+   {
+      MESA_FORMAT_Z32_FLOAT,       /* Name */
+      "MESA_FORMAT_Z32_FLOAT",     /* StrName */
+      GL_DEPTH_COMPONENT,          /* BaseFormat */
+      GL_FLOAT,                    /* DataType */
+      0, 0, 0, 0,                  /* Red/Green/Blue/AlphaBits */
+      0, 0, 0, 32, 0,              /* Lum/Int/Index/Depth/StencilBits */
+      1, 1, 4                      /* BlockWidth/Height,Bytes */
+   },
+   {
+      MESA_FORMAT_Z32_FLOAT_X24S8, /* Name */
+      "MESA_FORMAT_Z32_FLOAT_X24S8", /* StrName */
+      GL_DEPTH_STENCIL,            /* BaseFormat */
+      GL_NONE /* XXX */,           /* DataType */
+      0, 0, 0, 0,                  /* Red/Green/Blue/AlphaBits */
+      0, 0, 0, 32, 8,              /* Lum/Int/Index/Depth/StencilBits */
+      1, 1, 8                      /* BlockWidth/Height,Bytes */
+   },
 };
 
 
@@ -1654,6 +1673,16 @@ _mesa_format_to_type_and_comps(gl_format format,
       *comps = 1;
       return;
 
+   case MESA_FORMAT_Z32_FLOAT:
+      *datatype = GL_FLOAT;
+      *comps = 1;
+      return;
+
+   case MESA_FORMAT_Z32_FLOAT_X24S8:
+      *datatype = GL_FLOAT_32_UNSIGNED_INT_24_8_REV;
+      *comps = 1;
+      return;
+
    case MESA_FORMAT_DUDV8:
       *datatype = GL_BYTE;
       *comps = 2;
diff --git a/mesalib/src/mesa/main/formats.h b/mesalib/src/mesa/main/formats.h
index 0640bbc4a..5b8c01781 100644
--- a/mesalib/src/mesa/main/formats.h
+++ b/mesalib/src/mesa/main/formats.h
@@ -209,6 +209,9 @@ typedef enum
    MESA_FORMAT_RGB9_E5_FLOAT,
    MESA_FORMAT_R11_G11_B10_FLOAT,
 
+   MESA_FORMAT_Z32_FLOAT,
+   MESA_FORMAT_Z32_FLOAT_X24S8,
+
    MESA_FORMAT_COUNT
 } gl_format;
 
diff --git a/mesalib/src/mesa/main/framebuffer.c b/mesalib/src/mesa/main/framebuffer.c
index 66c9bd910..6e2ce7421 100644
--- a/mesalib/src/mesa/main/framebuffer.c
+++ b/mesalib/src/mesa/main/framebuffer.c
@@ -631,8 +631,14 @@ _mesa_update_depth_buffer(struct gl_context *ctx,
           || fb->_DepthBuffer->Wrapped != depthRb
           || _mesa_get_format_base_format(fb->_DepthBuffer->Format) != GL_DEPTH_COMPONENT) {
          /* need to update wrapper */
-         struct gl_renderbuffer *wrapper
-            = _mesa_new_z24_renderbuffer_wrapper(ctx, depthRb);
+         struct gl_renderbuffer *wrapper;
+
+         if (depthRb->Format == MESA_FORMAT_Z32_FLOAT_X24S8) {
+            wrapper = _mesa_new_z32f_renderbuffer_wrapper(ctx, depthRb);
+         }
+         else {
+            wrapper = _mesa_new_z24_renderbuffer_wrapper(ctx, depthRb);
+         }
          _mesa_reference_renderbuffer(&fb->_DepthBuffer, wrapper);
          ASSERT(fb->_DepthBuffer->Wrapped == depthRb);
       }
diff --git a/mesalib/src/mesa/main/image.c b/mesalib/src/mesa/main/image.c
index a5cd5ac03..37127dcb7 100644
--- a/mesalib/src/mesa/main/image.c
+++ b/mesalib/src/mesa/main/image.c
@@ -1,1935 +1,1949 @@
-/*

- * 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 image.c

- * Image handling.

- */

-

-

-#include "glheader.h"

-#include "colormac.h"

-#include "image.h"

-#include "imports.h"

-#include "macros.h"

-#include "mfeatures.h"

-#include "mtypes.h"

-

-

-/**

- * NOTE:

- * Normally, BYTE_TO_FLOAT(0) returns 0.00392  That causes problems when

- * we later convert the float to a packed integer value (such as for

- * GL_RGB5_A1) because we'll wind up with a non-zero value.

- *

- * We redefine the macros here so zero is handled correctly.

- */

-#undef BYTE_TO_FLOAT

-#define BYTE_TO_FLOAT(B)    ((B) == 0 ? 0.0F : ((2.0F * (B) + 1.0F) * (1.0F/255.0F)))

-

-#undef SHORT_TO_FLOAT

-#define SHORT_TO_FLOAT(S)   ((S) == 0 ? 0.0F : ((2.0F * (S) + 1.0F) * (1.0F/65535.0F)))

-

-

-

-/** Compute ceiling of integer quotient of A divided by B. */

-#define CEILING( A, B )  ( (A) % (B) == 0 ? (A)/(B) : (A)/(B)+1 )

-

-

-/**

- * \return GL_TRUE if type is packed pixel type, GL_FALSE otherwise.

- */

-GLboolean

-_mesa_type_is_packed(GLenum type)

-{

-   switch (type) {

-   case GL_UNSIGNED_BYTE_3_3_2:

-   case GL_UNSIGNED_BYTE_2_3_3_REV:

-   case MESA_UNSIGNED_BYTE_4_4:

-   case GL_UNSIGNED_SHORT_5_6_5:

-   case GL_UNSIGNED_SHORT_5_6_5_REV:

-   case GL_UNSIGNED_SHORT_4_4_4_4:

-   case GL_UNSIGNED_SHORT_4_4_4_4_REV:

-   case GL_UNSIGNED_SHORT_5_5_5_1:

-   case GL_UNSIGNED_SHORT_1_5_5_5_REV:

-   case GL_UNSIGNED_INT_8_8_8_8:

-   case GL_UNSIGNED_INT_8_8_8_8_REV:

-   case GL_UNSIGNED_INT_10_10_10_2:

-   case GL_UNSIGNED_INT_2_10_10_10_REV:

-   case GL_UNSIGNED_SHORT_8_8_MESA:

-   case GL_UNSIGNED_SHORT_8_8_REV_MESA:

-   case GL_UNSIGNED_INT_24_8_EXT:

-   case GL_UNSIGNED_INT_5_9_9_9_REV:

-   case GL_UNSIGNED_INT_10F_11F_11F_REV:

-      return GL_TRUE;

-   }

-

-   return GL_FALSE;

-}

-

-

-

-/**

- * Flip the order of the 2 bytes in each word in the given array.

- *

- * \param p array.

- * \param n number of words.

- */

-void

-_mesa_swap2( GLushort *p, GLuint n )

-{

-   GLuint i;

-   for (i = 0; i < n; i++) {

-      p[i] = (p[i] >> 8) | ((p[i] << 8) & 0xff00);

-   }

-}

-

-

-

-/*

- * Flip the order of the 4 bytes in each word in the given array.

- */

-void

-_mesa_swap4( GLuint *p, GLuint n )

-{

-   GLuint i, a, b;

-   for (i = 0; i < n; i++) {

-      b = p[i];

-      a =  (b >> 24)

-	| ((b >> 8) & 0xff00)

-	| ((b << 8) & 0xff0000)

-	| ((b << 24) & 0xff000000);

-      p[i] = a;

-   }

-}

-

-

-/**

- * Get the size of a GL data type.

- *

- * \param type GL data type.

- *

- * \return the size, in bytes, of the given data type, 0 if a GL_BITMAP, or -1

- * if an invalid type enum.

- */

-GLint

-_mesa_sizeof_type( GLenum type )

-{

-   switch (type) {

-      case GL_BITMAP:

-	 return 0;

-      case GL_UNSIGNED_BYTE:

-         return sizeof(GLubyte);

-      case GL_BYTE:

-	 return sizeof(GLbyte);

-      case GL_UNSIGNED_SHORT:

-	 return sizeof(GLushort);

-      case GL_SHORT:

-	 return sizeof(GLshort);

-      case GL_UNSIGNED_INT:

-	 return sizeof(GLuint);

-      case GL_INT:

-	 return sizeof(GLint);

-      case GL_FLOAT:

-	 return sizeof(GLfloat);

-      case GL_DOUBLE:

-	 return sizeof(GLdouble);

-      case GL_HALF_FLOAT_ARB:

-	 return sizeof(GLhalfARB);

-      case GL_FIXED:

-	 return sizeof(GLfixed);

-      default:

-         return -1;

-   }

-}

-

-

-/**

- * Same as _mesa_sizeof_type() but also accepting the packed pixel

- * format data types.

- */

-GLint

-_mesa_sizeof_packed_type( GLenum type )

-{

-   switch (type) {

-      case GL_BITMAP:

-	 return 0;

-      case GL_UNSIGNED_BYTE:

-         return sizeof(GLubyte);

-      case GL_BYTE:

-	 return sizeof(GLbyte);

-      case GL_UNSIGNED_SHORT:

-	 return sizeof(GLushort);

-      case GL_SHORT:

-	 return sizeof(GLshort);

-      case GL_UNSIGNED_INT:

-	 return sizeof(GLuint);

-      case GL_INT:

-	 return sizeof(GLint);

-      case GL_HALF_FLOAT_ARB:

-	 return sizeof(GLhalfARB);

-      case GL_FLOAT:

-	 return sizeof(GLfloat);

-      case GL_UNSIGNED_BYTE_3_3_2:

-         return sizeof(GLubyte);

-      case GL_UNSIGNED_BYTE_2_3_3_REV:

-         return sizeof(GLubyte);

-      case MESA_UNSIGNED_BYTE_4_4:

-         return sizeof(GLubyte);

-      case GL_UNSIGNED_SHORT_5_6_5:

-         return sizeof(GLushort);

-      case GL_UNSIGNED_SHORT_5_6_5_REV:

-         return sizeof(GLushort);

-      case GL_UNSIGNED_SHORT_4_4_4_4:

-         return sizeof(GLushort);

-      case GL_UNSIGNED_SHORT_4_4_4_4_REV:

-         return sizeof(GLushort);

-      case GL_UNSIGNED_SHORT_5_5_5_1:

-         return sizeof(GLushort);

-      case GL_UNSIGNED_SHORT_1_5_5_5_REV:

-         return sizeof(GLushort);

-      case GL_UNSIGNED_INT_8_8_8_8:

-         return sizeof(GLuint);

-      case GL_UNSIGNED_INT_8_8_8_8_REV:

-         return sizeof(GLuint);

-      case GL_UNSIGNED_INT_10_10_10_2:

-         return sizeof(GLuint);

-      case GL_UNSIGNED_INT_2_10_10_10_REV:

-         return sizeof(GLuint);

-      case GL_UNSIGNED_SHORT_8_8_MESA:

-      case GL_UNSIGNED_SHORT_8_8_REV_MESA:

-         return sizeof(GLushort);      

-      case GL_UNSIGNED_INT_24_8_EXT:

-         return sizeof(GLuint);

-      case GL_UNSIGNED_INT_5_9_9_9_REV:

-         return sizeof(GLuint);

-      case GL_UNSIGNED_INT_10F_11F_11F_REV:

-         return sizeof(GLuint);

-      default:

-         return -1;

-   }

-}

-

-

-/**

- * Get the number of components in a pixel format.

- *

- * \param format pixel format.

- *

- * \return the number of components in the given format, or -1 if a bad format.

- */

-GLint

-_mesa_components_in_format( GLenum format )

-{

-   switch (format) {

-      case GL_COLOR_INDEX:

-      case GL_COLOR_INDEX1_EXT:

-      case GL_COLOR_INDEX2_EXT:

-      case GL_COLOR_INDEX4_EXT:

-      case GL_COLOR_INDEX8_EXT:

-      case GL_COLOR_INDEX12_EXT:

-      case GL_COLOR_INDEX16_EXT:

-      case GL_STENCIL_INDEX:

-      case GL_DEPTH_COMPONENT:

-      case GL_RED:

-      case GL_RED_INTEGER_EXT:

-      case GL_GREEN:

-      case GL_GREEN_INTEGER_EXT:

-      case GL_BLUE:

-      case GL_BLUE_INTEGER_EXT:

-      case GL_ALPHA:

-      case GL_ALPHA_INTEGER_EXT:

-      case GL_LUMINANCE:

-      case GL_LUMINANCE_INTEGER_EXT:

-      case GL_INTENSITY:

-         return 1;

-      case GL_LUMINANCE_ALPHA:

-      case GL_LUMINANCE_ALPHA_INTEGER_EXT:

-      case GL_RG:

-	 return 2;

-      case GL_RGB:

-      case GL_RGB_INTEGER_EXT:

-	 return 3;

-      case GL_RGBA:

-      case GL_RGBA_INTEGER_EXT:

-	 return 4;

-      case GL_BGR:

-	 return 3;

-      case GL_BGRA:

-	 return 4;

-      case GL_ABGR_EXT:

-         return 4;

-      case GL_YCBCR_MESA:

-         return 2;

-      case GL_DEPTH_STENCIL_EXT:

-         return 2;

-      case GL_DUDV_ATI:

-      case GL_DU8DV8_ATI:

-         return 2;

-      default:

-         return -1;

-   }

-}

-

-

-/**

- * Get the bytes per pixel of pixel format type pair.

- *

- * \param format pixel format.

- * \param type pixel type.

- *

- * \return bytes per pixel, or -1 if a bad format or type was given.

- */

-GLint

-_mesa_bytes_per_pixel( GLenum format, GLenum type )

-{

-   GLint comps = _mesa_components_in_format( format );

-   if (comps < 0)

-      return -1;

-

-   switch (type) {

-      case GL_BITMAP:

-         return 0;  /* special case */

-      case GL_BYTE:

-      case GL_UNSIGNED_BYTE:

-         return comps * sizeof(GLubyte);

-      case GL_SHORT:

-      case GL_UNSIGNED_SHORT:

-         return comps * sizeof(GLshort);

-      case GL_INT:

-      case GL_UNSIGNED_INT:

-         return comps * sizeof(GLint);

-      case GL_FLOAT:

-         return comps * sizeof(GLfloat);

-      case GL_HALF_FLOAT_ARB:

-         return comps * sizeof(GLhalfARB);

-      case GL_UNSIGNED_BYTE_3_3_2:

-      case GL_UNSIGNED_BYTE_2_3_3_REV:

-         if (format == GL_RGB || format == GL_BGR ||

-             format == GL_RGB_INTEGER_EXT || format == GL_BGR_INTEGER_EXT)

-            return sizeof(GLubyte);

-         else

-            return -1;  /* error */

-      case GL_UNSIGNED_SHORT_5_6_5:

-      case GL_UNSIGNED_SHORT_5_6_5_REV:

-         if (format == GL_RGB || format == GL_BGR ||

-             format == GL_RGB_INTEGER_EXT || format == GL_BGR_INTEGER_EXT)

-            return sizeof(GLushort);

-         else

-            return -1;  /* error */

-      case GL_UNSIGNED_SHORT_4_4_4_4:

-      case GL_UNSIGNED_SHORT_4_4_4_4_REV:

-      case GL_UNSIGNED_SHORT_5_5_5_1:

-      case GL_UNSIGNED_SHORT_1_5_5_5_REV:

-         if (format == GL_RGBA || format == GL_BGRA || format == GL_ABGR_EXT ||

-             format == GL_RGBA_INTEGER_EXT || format == GL_BGRA_INTEGER_EXT)

-            return sizeof(GLushort);

-         else

-            return -1;

-      case GL_UNSIGNED_INT_8_8_8_8:

-      case GL_UNSIGNED_INT_8_8_8_8_REV:

-      case GL_UNSIGNED_INT_10_10_10_2:

-      case GL_UNSIGNED_INT_2_10_10_10_REV:

-         if (format == GL_RGBA || format == GL_BGRA || format == GL_ABGR_EXT ||

-             format == GL_RGBA_INTEGER_EXT || format == GL_BGRA_INTEGER_EXT)

-            return sizeof(GLuint);

-         else

-            return -1;

-      case GL_UNSIGNED_SHORT_8_8_MESA:

-      case GL_UNSIGNED_SHORT_8_8_REV_MESA:

-         if (format == GL_YCBCR_MESA)

-            return sizeof(GLushort);

-         else

-            return -1;

-      case GL_UNSIGNED_INT_24_8_EXT:

-         if (format == GL_DEPTH_STENCIL_EXT)

-            return sizeof(GLuint);

-         else

-            return -1;

-      case GL_UNSIGNED_INT_5_9_9_9_REV:

-         if (format == GL_RGB)

-            return sizeof(GLuint);

-         else

-            return -1;

-      case GL_UNSIGNED_INT_10F_11F_11F_REV:

-         if (format == GL_RGB)

-            return sizeof(GLuint);

-         else

-            return -1;

-      default:

-         return -1;

-   }

-}

-

-

-/**

- * Test for a legal pixel format and type.

- *

- * \param format pixel format.

- * \param type pixel type.

- *

- * \return GL_TRUE if the given pixel format and type are legal, or GL_FALSE

- * otherwise.

- */

-GLboolean

-_mesa_is_legal_format_and_type(const struct gl_context *ctx,

-                               GLenum format, GLenum type)

-{

-   switch (format) {

-      case GL_COLOR_INDEX:

-      case GL_STENCIL_INDEX:

-         switch (type) {

-            case GL_BITMAP:

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_FLOAT:

-               return GL_TRUE;

-            case GL_HALF_FLOAT_ARB:

-               return ctx->Extensions.ARB_half_float_pixel;

-            default:

-               return GL_FALSE;

-         }

-      case GL_RED:

-      case GL_GREEN:

-      case GL_BLUE:

-      case GL_ALPHA:

-#if 0 /* not legal!  see table 3.6 of the 1.5 spec */

-      case GL_INTENSITY:

-#endif

-      case GL_LUMINANCE:

-      case GL_LUMINANCE_ALPHA:

-      case GL_DEPTH_COMPONENT:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_FLOAT:

-               return GL_TRUE;

-            case GL_HALF_FLOAT_ARB:

-               return ctx->Extensions.ARB_half_float_pixel;

-            default:

-               return GL_FALSE;

-         }

-      case GL_RG:

-	 if (!ctx->Extensions.ARB_texture_rg)

-	    return GL_FALSE;

-

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_FLOAT:

-               return GL_TRUE;

-            case GL_HALF_FLOAT_ARB:

-               return ctx->Extensions.ARB_half_float_pixel;

-            default:

-               return GL_FALSE;

-         }

-      case GL_RGB:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_FLOAT:

-            case GL_UNSIGNED_BYTE_3_3_2:

-            case GL_UNSIGNED_BYTE_2_3_3_REV:

-            case GL_UNSIGNED_SHORT_5_6_5:

-            case GL_UNSIGNED_SHORT_5_6_5_REV:

-               return GL_TRUE;

-            case GL_HALF_FLOAT_ARB:

-               return ctx->Extensions.ARB_half_float_pixel;

-            case GL_UNSIGNED_INT_5_9_9_9_REV:

-               return ctx->Extensions.EXT_texture_shared_exponent;

-            case GL_UNSIGNED_INT_10F_11F_11F_REV:

-               return ctx->Extensions.EXT_packed_float;

-            default:

-               return GL_FALSE;

-         }

-      case GL_BGR:

-         switch (type) {

-            /* NOTE: no packed types are supported with BGR.  That's

-             * intentional, according to the GL spec.

-             */

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_FLOAT:

-               return GL_TRUE;

-            case GL_HALF_FLOAT_ARB:

-               return ctx->Extensions.ARB_half_float_pixel;

-            default:

-               return GL_FALSE;

-         }

-      case GL_RGBA:

-      case GL_BGRA:

-      case GL_ABGR_EXT:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_FLOAT:

-            case GL_UNSIGNED_SHORT_4_4_4_4:

-            case GL_UNSIGNED_SHORT_4_4_4_4_REV:

-            case GL_UNSIGNED_SHORT_5_5_5_1:

-            case GL_UNSIGNED_SHORT_1_5_5_5_REV:

-            case GL_UNSIGNED_INT_8_8_8_8:

-            case GL_UNSIGNED_INT_8_8_8_8_REV:

-            case GL_UNSIGNED_INT_10_10_10_2:

-            case GL_UNSIGNED_INT_2_10_10_10_REV:

-               return GL_TRUE;

-            case GL_HALF_FLOAT_ARB:

-               return ctx->Extensions.ARB_half_float_pixel;

-            default:

-               return GL_FALSE;

-         }

-      case GL_YCBCR_MESA:

-         if (type == GL_UNSIGNED_SHORT_8_8_MESA ||

-             type == GL_UNSIGNED_SHORT_8_8_REV_MESA)

-            return GL_TRUE;

-         else

-            return GL_FALSE;

-      case GL_DEPTH_STENCIL_EXT:

-         if (ctx->Extensions.EXT_packed_depth_stencil

-             && type == GL_UNSIGNED_INT_24_8_EXT)

-            return GL_TRUE;

-         else

-            return GL_FALSE;

-      case GL_DUDV_ATI:

-      case GL_DU8DV8_ATI:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_FLOAT:

-               return GL_TRUE;

-            default:

-               return GL_FALSE;

-         }

-

-      /* integer-valued formats */

-      case GL_RED_INTEGER_EXT:

-      case GL_GREEN_INTEGER_EXT:

-      case GL_BLUE_INTEGER_EXT:

-      case GL_ALPHA_INTEGER_EXT:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-               return ctx->Extensions.EXT_texture_integer;

-            default:

-               return GL_FALSE;

-         }

-

-      case GL_RGB_INTEGER_EXT:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_UNSIGNED_BYTE_3_3_2:

-            case GL_UNSIGNED_BYTE_2_3_3_REV:

-            case GL_UNSIGNED_SHORT_5_6_5:

-            case GL_UNSIGNED_SHORT_5_6_5_REV:

-               return ctx->Extensions.EXT_texture_integer;

-            default:

-               return GL_FALSE;

-         }

-

-      case GL_BGR_INTEGER_EXT:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            /* NOTE: no packed formats w/ BGR format */

-               return ctx->Extensions.EXT_texture_integer;

-            default:

-               return GL_FALSE;

-         }

-

-      case GL_RGBA_INTEGER_EXT:

-      case GL_BGRA_INTEGER_EXT:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-            case GL_UNSIGNED_SHORT_4_4_4_4:

-            case GL_UNSIGNED_SHORT_4_4_4_4_REV:

-            case GL_UNSIGNED_SHORT_5_5_5_1:

-            case GL_UNSIGNED_SHORT_1_5_5_5_REV:

-            case GL_UNSIGNED_INT_8_8_8_8:

-            case GL_UNSIGNED_INT_8_8_8_8_REV:

-            case GL_UNSIGNED_INT_10_10_10_2:

-            case GL_UNSIGNED_INT_2_10_10_10_REV:

-               return ctx->Extensions.EXT_texture_integer;

-            default:

-               return GL_FALSE;

-         }

-

-      case GL_LUMINANCE_INTEGER_EXT:

-      case GL_LUMINANCE_ALPHA_INTEGER_EXT:

-         switch (type) {

-            case GL_BYTE:

-            case GL_UNSIGNED_BYTE:

-            case GL_SHORT:

-            case GL_UNSIGNED_SHORT:

-            case GL_INT:

-            case GL_UNSIGNED_INT:

-               return ctx->Extensions.EXT_texture_integer;

-            default:

-               return GL_FALSE;

-         }

-

-      default:

-         ; /* fall-through */

-   }

-   return GL_FALSE;

-}

-

-

-/**

- * Test if the given image format is a color/RGBA format (i.e., not color

- * index, depth, stencil, etc).

- * \param format  the image format value (may by an internal texture format)

- * \return GL_TRUE if its a color/RGBA format, GL_FALSE otherwise.

- */

-GLboolean

-_mesa_is_color_format(GLenum format)

-{

-   switch (format) {

-      case GL_RED:

-      case GL_GREEN:

-      case GL_BLUE:

-      case GL_ALPHA:

-      case GL_ALPHA4:

-      case GL_ALPHA8:

-      case GL_ALPHA12:

-      case GL_ALPHA16:

-      case 1:

-      case GL_LUMINANCE:

-      case GL_LUMINANCE4:

-      case GL_LUMINANCE8:

-      case GL_LUMINANCE12:

-      case GL_LUMINANCE16:

-      case 2:

-      case GL_LUMINANCE_ALPHA:

-      case GL_LUMINANCE4_ALPHA4:

-      case GL_LUMINANCE6_ALPHA2:

-      case GL_LUMINANCE8_ALPHA8:

-      case GL_LUMINANCE12_ALPHA4:

-      case GL_LUMINANCE12_ALPHA12:

-      case GL_LUMINANCE16_ALPHA16:

-      case GL_INTENSITY:

-      case GL_INTENSITY4:

-      case GL_INTENSITY8:

-      case GL_INTENSITY12:

-      case GL_INTENSITY16:

-      case GL_R8:

-      case GL_R16:

-      case GL_RG:

-      case GL_RG8:

-      case GL_RG16:

-      case 3:

-      case GL_RGB:

-      case GL_BGR:

-      case GL_R3_G3_B2:

-      case GL_RGB4:

-      case GL_RGB5:

-      case GL_RGB8:

-      case GL_RGB10:

-      case GL_RGB12:

-      case GL_RGB16:

-      case 4:

-      case GL_ABGR_EXT:

-      case GL_RGBA:

-      case GL_BGRA:

-      case GL_RGBA2:

-      case GL_RGBA4:

-      case GL_RGB5_A1:

-      case GL_RGBA8:

-      case GL_RGB10_A2:

-      case GL_RGBA12:

-      case GL_RGBA16:

-      /* float texture formats */

-      case GL_ALPHA16F_ARB:

-      case GL_ALPHA32F_ARB:

-      case GL_LUMINANCE16F_ARB:

-      case GL_LUMINANCE32F_ARB:

-      case GL_LUMINANCE_ALPHA16F_ARB:

-      case GL_LUMINANCE_ALPHA32F_ARB:

-      case GL_INTENSITY16F_ARB:

-      case GL_INTENSITY32F_ARB:

-      case GL_R16F:

-      case GL_R32F:

-      case GL_RG16F:

-      case GL_RG32F:

-      case GL_RGB16F_ARB:

-      case GL_RGB32F_ARB:

-      case GL_RGBA16F_ARB:

-      case GL_RGBA32F_ARB:

-      /* compressed formats */

-      case GL_COMPRESSED_ALPHA:

-      case GL_COMPRESSED_LUMINANCE:

-      case GL_COMPRESSED_LUMINANCE_ALPHA:

-      case GL_COMPRESSED_INTENSITY:

-      case GL_COMPRESSED_RED:

-      case GL_COMPRESSED_RG:

-      case GL_COMPRESSED_RGB:

-      case GL_COMPRESSED_RGBA:

-      case GL_RGB_S3TC:

-      case GL_RGB4_S3TC:

-      case GL_RGBA_S3TC:

-      case GL_RGBA4_S3TC:

-      case GL_COMPRESSED_RGB_S3TC_DXT1_EXT:

-      case GL_COMPRESSED_RGBA_S3TC_DXT1_EXT:

-      case GL_COMPRESSED_RGBA_S3TC_DXT3_EXT:

-      case GL_COMPRESSED_RGBA_S3TC_DXT5_EXT:

-      case GL_COMPRESSED_RGB_FXT1_3DFX:

-      case GL_COMPRESSED_RGBA_FXT1_3DFX:

-#if FEATURE_EXT_texture_sRGB

-      case GL_SRGB_EXT:

-      case GL_SRGB8_EXT:

-      case GL_SRGB_ALPHA_EXT:

-      case GL_SRGB8_ALPHA8_EXT:

-      case GL_SLUMINANCE_ALPHA_EXT:

-      case GL_SLUMINANCE8_ALPHA8_EXT:

-      case GL_SLUMINANCE_EXT:

-      case GL_SLUMINANCE8_EXT:

-      case GL_COMPRESSED_SRGB_EXT:

-      case GL_COMPRESSED_SRGB_S3TC_DXT1_EXT:

-      case GL_COMPRESSED_SRGB_ALPHA_EXT:

-      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT:

-      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT:

-      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT:

-      case GL_COMPRESSED_SLUMINANCE_EXT:

-      case GL_COMPRESSED_SLUMINANCE_ALPHA_EXT:

-#endif /* FEATURE_EXT_texture_sRGB */

-      case GL_COMPRESSED_RED_RGTC1:

-      case GL_COMPRESSED_SIGNED_RED_RGTC1:

-      case GL_COMPRESSED_RG_RGTC2:

-      case GL_COMPRESSED_SIGNED_RG_RGTC2:

-      case GL_COMPRESSED_LUMINANCE_LATC1_EXT:

-      case GL_COMPRESSED_SIGNED_LUMINANCE_LATC1_EXT:

-      case GL_COMPRESSED_LUMINANCE_ALPHA_LATC2_EXT:

-      case GL_COMPRESSED_SIGNED_LUMINANCE_ALPHA_LATC2_EXT:

-      case GL_COMPRESSED_LUMINANCE_ALPHA_3DC_ATI:

-      /* generic integer formats */

-      case GL_RED_INTEGER_EXT:

-      case GL_GREEN_INTEGER_EXT:

-      case GL_BLUE_INTEGER_EXT:

-      case GL_ALPHA_INTEGER_EXT:

-      case GL_RGB_INTEGER_EXT:

-      case GL_RGBA_INTEGER_EXT:

-      case GL_BGR_INTEGER_EXT:

-      case GL_BGRA_INTEGER_EXT:

-      case GL_LUMINANCE_INTEGER_EXT:

-      case GL_LUMINANCE_ALPHA_INTEGER_EXT:

-      /* sized integer formats */

-      case GL_RGBA32UI_EXT:

-      case GL_RGB32UI_EXT:

-      case GL_ALPHA32UI_EXT:

-      case GL_INTENSITY32UI_EXT:

-      case GL_LUMINANCE32UI_EXT:

-      case GL_LUMINANCE_ALPHA32UI_EXT:

-      case GL_RGBA16UI_EXT:

-      case GL_RGB16UI_EXT:

-      case GL_ALPHA16UI_EXT:

-      case GL_INTENSITY16UI_EXT:

-      case GL_LUMINANCE16UI_EXT:

-      case GL_LUMINANCE_ALPHA16UI_EXT:

-      case GL_RGBA8UI_EXT:

-      case GL_RGB8UI_EXT:

-      case GL_ALPHA8UI_EXT:

-      case GL_INTENSITY8UI_EXT:

-      case GL_LUMINANCE8UI_EXT:

-      case GL_LUMINANCE_ALPHA8UI_EXT:

-      case GL_RGBA32I_EXT:

-      case GL_RGB32I_EXT:

-      case GL_ALPHA32I_EXT:

-      case GL_INTENSITY32I_EXT:

-      case GL_LUMINANCE32I_EXT:

-      case GL_LUMINANCE_ALPHA32I_EXT:

-      case GL_RGBA16I_EXT:

-      case GL_RGB16I_EXT:

-      case GL_ALPHA16I_EXT:

-      case GL_INTENSITY16I_EXT:

-      case GL_LUMINANCE16I_EXT:

-      case GL_LUMINANCE_ALPHA16I_EXT:

-      case GL_RGBA8I_EXT:

-      case GL_RGB8I_EXT:

-      case GL_ALPHA8I_EXT:

-      case GL_INTENSITY8I_EXT:

-      case GL_LUMINANCE8I_EXT:

-      case GL_LUMINANCE_ALPHA8I_EXT:

-      /* signed, normalized texture formats */

-      case GL_RED_SNORM:

-      case GL_R8_SNORM:

-      case GL_R16_SNORM:

-      case GL_RG_SNORM:

-      case GL_RG8_SNORM:

-      case GL_RG16_SNORM:

-      case GL_RGB_SNORM:

-      case GL_RGB8_SNORM:

-      case GL_RGB16_SNORM:

-      case GL_RGBA_SNORM:

-      case GL_RGBA8_SNORM:

-      case GL_RGBA16_SNORM:

-      case GL_ALPHA_SNORM:

-      case GL_ALPHA8_SNORM:

-      case GL_ALPHA16_SNORM:

-      case GL_LUMINANCE_SNORM:

-      case GL_LUMINANCE8_SNORM:

-      case GL_LUMINANCE16_SNORM:

-      case GL_LUMINANCE_ALPHA_SNORM:

-      case GL_LUMINANCE8_ALPHA8_SNORM:

-      case GL_LUMINANCE16_ALPHA16_SNORM:

-      case GL_INTENSITY_SNORM:

-      case GL_INTENSITY8_SNORM:

-      case GL_INTENSITY16_SNORM:

-      case GL_RGB9_E5:

-      case GL_R11F_G11F_B10F:

-         return GL_TRUE;

-      case GL_YCBCR_MESA:  /* not considered to be RGB */

-         /* fall-through */

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given image format is a color index format.

- */

-GLboolean

-_mesa_is_index_format(GLenum format)

-{

-   switch (format) {

-      case GL_COLOR_INDEX:

-      case GL_COLOR_INDEX1_EXT:

-      case GL_COLOR_INDEX2_EXT:

-      case GL_COLOR_INDEX4_EXT:

-      case GL_COLOR_INDEX8_EXT:

-      case GL_COLOR_INDEX12_EXT:

-      case GL_COLOR_INDEX16_EXT:

-         return GL_TRUE;

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given image format is a depth component format.

- */

-GLboolean

-_mesa_is_depth_format(GLenum format)

-{

-   switch (format) {

-      case GL_DEPTH_COMPONENT:

-      case GL_DEPTH_COMPONENT16:

-      case GL_DEPTH_COMPONENT24:

-      case GL_DEPTH_COMPONENT32:

-         return GL_TRUE;

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given image format is a stencil format.

- */

-GLboolean

-_mesa_is_stencil_format(GLenum format)

-{

-   switch (format) {

-      case GL_STENCIL_INDEX:

-      case GL_DEPTH_STENCIL:

-         return GL_TRUE;

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given image format is a YCbCr format.

- */

-GLboolean

-_mesa_is_ycbcr_format(GLenum format)

-{

-   switch (format) {

-      case GL_YCBCR_MESA:

-         return GL_TRUE;

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given image format is a depth+stencil format.

- */

-GLboolean

-_mesa_is_depthstencil_format(GLenum format)

-{

-   switch (format) {

-      case GL_DEPTH24_STENCIL8_EXT:

-      case GL_DEPTH_STENCIL_EXT:

-         return GL_TRUE;

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given image format is a depth or stencil format.

- */

-GLboolean

-_mesa_is_depth_or_stencil_format(GLenum format)

-{

-   switch (format) {

-      case GL_DEPTH_COMPONENT:

-      case GL_DEPTH_COMPONENT16:

-      case GL_DEPTH_COMPONENT24:

-      case GL_DEPTH_COMPONENT32:

-      case GL_STENCIL_INDEX:

-      case GL_STENCIL_INDEX1_EXT:

-      case GL_STENCIL_INDEX4_EXT:

-      case GL_STENCIL_INDEX8_EXT:

-      case GL_STENCIL_INDEX16_EXT:

-      case GL_DEPTH_STENCIL_EXT:

-      case GL_DEPTH24_STENCIL8_EXT:

-         return GL_TRUE;

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given image format is a dudv format.

- */

-GLboolean

-_mesa_is_dudv_format(GLenum format)

-{

-   switch (format) {

-      case GL_DUDV_ATI:

-      case GL_DU8DV8_ATI:

-         return GL_TRUE;

-      default:

-         return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if the given format is an integer (non-normalized) format.

- */

-GLboolean

-_mesa_is_integer_format(GLenum format)

-{

-   switch (format) {

-   /* generic integer formats */

-   case GL_RED_INTEGER_EXT:

-   case GL_GREEN_INTEGER_EXT:

-   case GL_BLUE_INTEGER_EXT:

-   case GL_ALPHA_INTEGER_EXT:

-   case GL_RGB_INTEGER_EXT:

-   case GL_RGBA_INTEGER_EXT:

-   case GL_BGR_INTEGER_EXT:

-   case GL_BGRA_INTEGER_EXT:

-   case GL_LUMINANCE_INTEGER_EXT:

-   case GL_LUMINANCE_ALPHA_INTEGER_EXT:

-   /* specific integer formats */

-   case GL_RGBA32UI_EXT:

-   case GL_RGB32UI_EXT:

-   case GL_ALPHA32UI_EXT:

-   case GL_INTENSITY32UI_EXT:

-   case GL_LUMINANCE32UI_EXT:

-   case GL_LUMINANCE_ALPHA32UI_EXT:

-   case GL_RGBA16UI_EXT:

-   case GL_RGB16UI_EXT:

-   case GL_ALPHA16UI_EXT:

-   case GL_INTENSITY16UI_EXT:

-   case GL_LUMINANCE16UI_EXT:

-   case GL_LUMINANCE_ALPHA16UI_EXT:

-   case GL_RGBA8UI_EXT:

-   case GL_RGB8UI_EXT:

-   case GL_ALPHA8UI_EXT:

-   case GL_INTENSITY8UI_EXT:

-   case GL_LUMINANCE8UI_EXT:

-   case GL_LUMINANCE_ALPHA8UI_EXT:

-   case GL_RGBA32I_EXT:

-   case GL_RGB32I_EXT:

-   case GL_ALPHA32I_EXT:

-   case GL_INTENSITY32I_EXT:

-   case GL_LUMINANCE32I_EXT:

-   case GL_LUMINANCE_ALPHA32I_EXT:

-   case GL_RGBA16I_EXT:

-   case GL_RGB16I_EXT:

-   case GL_ALPHA16I_EXT:

-   case GL_INTENSITY16I_EXT:

-   case GL_LUMINANCE16I_EXT:

-   case GL_LUMINANCE_ALPHA16I_EXT:

-   case GL_RGBA8I_EXT:

-   case GL_RGB8I_EXT:

-   case GL_ALPHA8I_EXT:

-   case GL_INTENSITY8I_EXT:

-   case GL_LUMINANCE8I_EXT:

-   case GL_LUMINANCE_ALPHA8I_EXT:

-      return GL_TRUE;

-   default:

-      return GL_FALSE;

-   }

-}

-

-

-/**

- * Test if an image format is a supported compressed format.

- * \param format the internal format token provided by the user.

- * \return GL_TRUE if compressed, GL_FALSE if uncompressed

- */

-GLboolean

-_mesa_is_compressed_format(struct gl_context *ctx, GLenum format)

-{

-   switch (format) {

-   case GL_COMPRESSED_RGB_S3TC_DXT1_EXT:

-   case GL_COMPRESSED_RGBA_S3TC_DXT1_EXT:

-   case GL_COMPRESSED_RGBA_S3TC_DXT3_EXT:

-   case GL_COMPRESSED_RGBA_S3TC_DXT5_EXT:

-      return ctx->Extensions.EXT_texture_compression_s3tc;

-   case GL_RGB_S3TC:

-   case GL_RGB4_S3TC:

-   case GL_RGBA_S3TC:

-   case GL_RGBA4_S3TC:

-      return ctx->Extensions.S3_s3tc;

-   case GL_COMPRESSED_SRGB_S3TC_DXT1_EXT:

-   case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT:

-   case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT:

-   case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT:

-      return ctx->Extensions.EXT_texture_sRGB

-         && ctx->Extensions.EXT_texture_compression_s3tc;

-   case GL_COMPRESSED_RGB_FXT1_3DFX:

-   case GL_COMPRESSED_RGBA_FXT1_3DFX:

-      return ctx->Extensions.TDFX_texture_compression_FXT1;

-   case GL_COMPRESSED_RED_RGTC1:

-   case GL_COMPRESSED_SIGNED_RED_RGTC1:

-   case GL_COMPRESSED_RG_RGTC2:

-   case GL_COMPRESSED_SIGNED_RG_RGTC2:

-      return ctx->Extensions.ARB_texture_compression_rgtc;

-   case GL_COMPRESSED_LUMINANCE_LATC1_EXT:

-   case GL_COMPRESSED_SIGNED_LUMINANCE_LATC1_EXT:

-   case GL_COMPRESSED_LUMINANCE_ALPHA_LATC2_EXT:

-   case GL_COMPRESSED_SIGNED_LUMINANCE_ALPHA_LATC2_EXT:

-      return ctx->Extensions.EXT_texture_compression_latc;

-   case GL_COMPRESSED_LUMINANCE_ALPHA_3DC_ATI:

-      return ctx->Extensions.ATI_texture_compression_3dc;

-   default:

-      return GL_FALSE;

-   }

-}

-

-

-/**

- * Return the address of a specific pixel in an image (1D, 2D or 3D).

- *

- * Pixel unpacking/packing parameters are observed according to \p packing.

- *

- * \param dimensions either 1, 2 or 3 to indicate dimensionality of image

- * \param image  starting address of image data

- * \param width  the image width

- * \param height  theimage height

- * \param format  the pixel format

- * \param type  the pixel data type

- * \param packing  the pixelstore attributes

- * \param img  which image in the volume (0 for 1D or 2D images)

- * \param row  row of pixel in the image (0 for 1D images)

- * \param column column of pixel in the image

- * 

- * \return address of pixel on success, or NULL on error.

- *

- * \sa gl_pixelstore_attrib.

- */

-GLvoid *

-_mesa_image_address( GLuint dimensions,

-                     const struct gl_pixelstore_attrib *packing,

-                     const GLvoid *image,

-                     GLsizei width, GLsizei height,

-                     GLenum format, GLenum type,

-                     GLint img, GLint row, GLint column )

-{

-   GLint alignment;        /* 1, 2 or 4 */

-   GLint pixels_per_row;

-   GLint rows_per_image;

-   GLint skiprows;

-   GLint skippixels;

-   GLint skipimages;       /* for 3-D volume images */

-   GLubyte *pixel_addr;

-

-   ASSERT(dimensions >= 1 && dimensions <= 3);

-

-   alignment = packing->Alignment;

-   if (packing->RowLength > 0) {

-      pixels_per_row = packing->RowLength;

-   }

-   else {

-      pixels_per_row = width;

-   }

-   if (packing->ImageHeight > 0) {

-      rows_per_image = packing->ImageHeight;

-   }

-   else {

-      rows_per_image = height;

-   }

-

-   skippixels = packing->SkipPixels;

-   /* Note: SKIP_ROWS _is_ used for 1D images */

-   skiprows = packing->SkipRows;

-   /* Note: SKIP_IMAGES is only used for 3D images */

-   skipimages = (dimensions == 3) ? packing->SkipImages : 0;

-

-   if (type == GL_BITMAP) {

-      /* BITMAP data */

-      GLint comp_per_pixel;   /* components per pixel */

-      GLint bytes_per_comp;   /* bytes per component */

-      GLint bytes_per_row;

-      GLint bytes_per_image;

-

-      /* Compute bytes per component */

-      bytes_per_comp = _mesa_sizeof_packed_type( type );

-      if (bytes_per_comp < 0) {

-         return NULL;

-      }

-

-      /* Compute number of components per pixel */

-      comp_per_pixel = _mesa_components_in_format( format );

-      if (comp_per_pixel < 0) {

-         return NULL;

-      }

-

-      bytes_per_row = alignment

-                    * CEILING( comp_per_pixel*pixels_per_row, 8*alignment );

-

-      bytes_per_image = bytes_per_row * rows_per_image;

-

-      pixel_addr = (GLubyte *) image

-                 + (skipimages + img) * bytes_per_image

-                 + (skiprows + row) * bytes_per_row

-                 + (skippixels + column) / 8;

-   }

-   else {

-      /* Non-BITMAP data */

-      GLint bytes_per_pixel, bytes_per_row, remainder, bytes_per_image;

-      GLint topOfImage;

-

-      bytes_per_pixel = _mesa_bytes_per_pixel( format, type );

-

-      /* The pixel type and format should have been error checked earlier */

-      assert(bytes_per_pixel > 0);

-

-      bytes_per_row = pixels_per_row * bytes_per_pixel;

-      remainder = bytes_per_row % alignment;

-      if (remainder > 0)

-         bytes_per_row += (alignment - remainder);

-

-      ASSERT(bytes_per_row % alignment == 0);

-

-      bytes_per_image = bytes_per_row * rows_per_image;

-

-      if (packing->Invert) {

-         /* set pixel_addr to the last row */

-         topOfImage = bytes_per_row * (height - 1);

-         bytes_per_row = -bytes_per_row;

-      }

-      else {

-         topOfImage = 0;

-      }

-

-      /* compute final pixel address */

-      pixel_addr = (GLubyte *) image

-                 + (skipimages + img) * bytes_per_image

-                 + topOfImage

-                 + (skiprows + row) * bytes_per_row

-                 + (skippixels + column) * bytes_per_pixel;

-   }

-

-   return (GLvoid *) pixel_addr;

-}

-

-

-GLvoid *

-_mesa_image_address1d( const struct gl_pixelstore_attrib *packing,

-                       const GLvoid *image,

-                       GLsizei width,

-                       GLenum format, GLenum type,

-                       GLint column )

-{

-   return _mesa_image_address(1, packing, image, width, 1,

-                              format, type, 0, 0, column);

-}

-

-

-GLvoid *

-_mesa_image_address2d( const struct gl_pixelstore_attrib *packing,

-                       const GLvoid *image,

-                       GLsizei width, GLsizei height,

-                       GLenum format, GLenum type,

-                       GLint row, GLint column )

-{

-   return _mesa_image_address(2, packing, image, width, height,

-                              format, type, 0, row, column);

-}

-

-

-GLvoid *

-_mesa_image_address3d( const struct gl_pixelstore_attrib *packing,

-                       const GLvoid *image,

-                       GLsizei width, GLsizei height,

-                       GLenum format, GLenum type,

-                       GLint img, GLint row, GLint column )

-{

-   return _mesa_image_address(3, packing, image, width, height,

-                              format, type, img, row, column);

-}

-

-

-

-/**

- * Compute the stride (in bytes) between image rows.

- *

- * \param packing the pixelstore attributes

- * \param width image width.

- * \param format pixel format.

- * \param type pixel data type.

- * 

- * \return the stride in bytes for the given parameters, or -1 if error

- */

-GLint

-_mesa_image_row_stride( const struct gl_pixelstore_attrib *packing,

-                        GLint width, GLenum format, GLenum type )

-{

-   GLint bytesPerRow, remainder;

-

-   ASSERT(packing);

-

-   if (type == GL_BITMAP) {

-      if (packing->RowLength == 0) {

-         bytesPerRow = (width + 7) / 8;

-      }

-      else {

-         bytesPerRow = (packing->RowLength + 7) / 8;

-      }

-   }

-   else {

-      /* Non-BITMAP data */

-      const GLint bytesPerPixel = _mesa_bytes_per_pixel(format, type);

-      if (bytesPerPixel <= 0)

-         return -1;  /* error */

-      if (packing->RowLength == 0) {

-         bytesPerRow = bytesPerPixel * width;

-      }

-      else {

-         bytesPerRow = bytesPerPixel * packing->RowLength;

-      }

-   }

-

-   remainder = bytesPerRow % packing->Alignment;

-   if (remainder > 0) {

-      bytesPerRow += (packing->Alignment - remainder);

-   }

-

-   if (packing->Invert) {

-      /* negate the bytes per row (negative row stride) */

-      bytesPerRow = -bytesPerRow;

-   }

-

-   return bytesPerRow;

-}

-

-

-/*

- * Compute the stride between images in a 3D texture (in bytes) for the given

- * pixel packing parameters and image width, format and type.

- */

-GLint

-_mesa_image_image_stride( const struct gl_pixelstore_attrib *packing,

-                          GLint width, GLint height,

-                          GLenum format, GLenum type )

-{

-   GLint bytesPerRow, bytesPerImage, remainder;

-

-   ASSERT(packing);

-

-   if (type == GL_BITMAP) {

-      if (packing->RowLength == 0) {

-         bytesPerRow = (width + 7) / 8;

-      }

-      else {

-         bytesPerRow = (packing->RowLength + 7) / 8;

-      }

-   }

-   else {

-      const GLint bytesPerPixel = _mesa_bytes_per_pixel(format, type);

-

-      if (bytesPerPixel <= 0)

-         return -1;  /* error */

-      if (packing->RowLength == 0) {

-         bytesPerRow = bytesPerPixel * width;

-      }

-      else {

-         bytesPerRow = bytesPerPixel * packing->RowLength;

-      }

-   }

-

-   remainder = bytesPerRow % packing->Alignment;

-   if (remainder > 0)

-      bytesPerRow += (packing->Alignment - remainder);

-

-   if (packing->ImageHeight == 0)

-      bytesPerImage = bytesPerRow * height;

-   else

-      bytesPerImage = bytesPerRow * packing->ImageHeight;

-

-   return bytesPerImage;

-}

-

-

-

-/**

- * "Expand" a bitmap from 1-bit per pixel to 8-bits per pixel.

- * This is typically used to convert a bitmap into a GLubyte/pixel texture.

- * "On" bits will set texels to \p onValue.

- * "Off" bits will not modify texels.

- * \param width  src bitmap width in pixels

- * \param height  src bitmap height in pixels

- * \param unpack  bitmap unpacking state

- * \param bitmap  the src bitmap data

- * \param destBuffer  start of dest buffer

- * \param destStride  row stride in dest buffer

- * \param onValue  if bit is 1, set destBuffer pixel to this value

- */

-void

-_mesa_expand_bitmap(GLsizei width, GLsizei height,

-                    const struct gl_pixelstore_attrib *unpack,

-                    const GLubyte *bitmap,

-                    GLubyte *destBuffer, GLint destStride,

-                    GLubyte onValue)

-{

-   const GLubyte *srcRow = (const GLubyte *)

-      _mesa_image_address2d(unpack, bitmap, width, height,

-                            GL_COLOR_INDEX, GL_BITMAP, 0, 0);

-   const GLint srcStride = _mesa_image_row_stride(unpack, width,

-                                                  GL_COLOR_INDEX, GL_BITMAP);

-   GLint row, col;

-

-#define SET_PIXEL(COL, ROW) \

-   destBuffer[(ROW) * destStride + (COL)] = onValue;

-

-   for (row = 0; row < height; row++) {

-      const GLubyte *src = srcRow;

-

-      if (unpack->LsbFirst) {

-         /* Lsb first */

-         GLubyte mask = 1U << (unpack->SkipPixels & 0x7);

-         for (col = 0; col < width; col++) {

-

-            if (*src & mask) {

-               SET_PIXEL(col, row);

-            }

-

-            if (mask == 128U) {

-               src++;

-               mask = 1U;

-            }

-            else {

-               mask = mask << 1;

-            }

-         }

-

-         /* get ready for next row */

-         if (mask != 1)

-            src++;

-      }

-      else {

-         /* Msb first */

-         GLubyte mask = 128U >> (unpack->SkipPixels & 0x7);

-         for (col = 0; col < width; col++) {

-

-            if (*src & mask) {

-               SET_PIXEL(col, row);

-            }

-

-            if (mask == 1U) {

-               src++;

-               mask = 128U;

-            }

-            else {

-               mask = mask >> 1;

-            }

-         }

-

-         /* get ready for next row */

-         if (mask != 128)

-            src++;

-      }

-

-      srcRow += srcStride;

-   } /* row */

-

-#undef SET_PIXEL

-}

-

-

-

-

-/**

- * Convert an array of RGBA colors from one datatype to another.

- * NOTE: src may equal dst.  In that case, we use a temporary buffer.

- */

-void

-_mesa_convert_colors(GLenum srcType, const GLvoid *src,

-                     GLenum dstType, GLvoid *dst,

-                     GLuint count, const GLubyte mask[])

-{

-   GLuint tempBuffer[MAX_WIDTH][4];

-   const GLboolean useTemp = (src == dst);

-

-   ASSERT(srcType != dstType);

-

-   switch (srcType) {

-   case GL_UNSIGNED_BYTE:

-      if (dstType == GL_UNSIGNED_SHORT) {

-         const GLubyte (*src1)[4] = (const GLubyte (*)[4]) src;

-         GLushort (*dst2)[4] = (GLushort (*)[4]) (useTemp ? tempBuffer : dst);

-         GLuint i;

-         for (i = 0; i < count; i++) {

-            if (!mask || mask[i]) {

-               dst2[i][RCOMP] = UBYTE_TO_USHORT(src1[i][RCOMP]);

-               dst2[i][GCOMP] = UBYTE_TO_USHORT(src1[i][GCOMP]);

-               dst2[i][BCOMP] = UBYTE_TO_USHORT(src1[i][BCOMP]);

-               dst2[i][ACOMP] = UBYTE_TO_USHORT(src1[i][ACOMP]);

-            }

-         }

-         if (useTemp)

-            memcpy(dst, tempBuffer, count * 4 * sizeof(GLushort));

-      }

-      else {

-         const GLubyte (*src1)[4] = (const GLubyte (*)[4]) src;

-         GLfloat (*dst4)[4] = (GLfloat (*)[4]) (useTemp ? tempBuffer : dst);

-         GLuint i;

-         ASSERT(dstType == GL_FLOAT);

-         for (i = 0; i < count; i++) {

-            if (!mask || mask[i]) {

-               dst4[i][RCOMP] = UBYTE_TO_FLOAT(src1[i][RCOMP]);

-               dst4[i][GCOMP] = UBYTE_TO_FLOAT(src1[i][GCOMP]);

-               dst4[i][BCOMP] = UBYTE_TO_FLOAT(src1[i][BCOMP]);

-               dst4[i][ACOMP] = UBYTE_TO_FLOAT(src1[i][ACOMP]);

-            }

-         }

-         if (useTemp)

-            memcpy(dst, tempBuffer, count * 4 * sizeof(GLfloat));

-      }

-      break;

-   case GL_UNSIGNED_SHORT:

-      if (dstType == GL_UNSIGNED_BYTE) {

-         const GLushort (*src2)[4] = (const GLushort (*)[4]) src;

-         GLubyte (*dst1)[4] = (GLubyte (*)[4]) (useTemp ? tempBuffer : dst);

-         GLuint i;

-         for (i = 0; i < count; i++) {

-            if (!mask || mask[i]) {

-               dst1[i][RCOMP] = USHORT_TO_UBYTE(src2[i][RCOMP]);

-               dst1[i][GCOMP] = USHORT_TO_UBYTE(src2[i][GCOMP]);

-               dst1[i][BCOMP] = USHORT_TO_UBYTE(src2[i][BCOMP]);

-               dst1[i][ACOMP] = USHORT_TO_UBYTE(src2[i][ACOMP]);

-            }

-         }

-         if (useTemp)

-            memcpy(dst, tempBuffer, count * 4 * sizeof(GLubyte));

-      }

-      else {

-         const GLushort (*src2)[4] = (const GLushort (*)[4]) src;

-         GLfloat (*dst4)[4] = (GLfloat (*)[4]) (useTemp ? tempBuffer : dst);

-         GLuint i;

-         ASSERT(dstType == GL_FLOAT);

-         for (i = 0; i < count; i++) {

-            if (!mask || mask[i]) {

-               dst4[i][RCOMP] = USHORT_TO_FLOAT(src2[i][RCOMP]);

-               dst4[i][GCOMP] = USHORT_TO_FLOAT(src2[i][GCOMP]);

-               dst4[i][BCOMP] = USHORT_TO_FLOAT(src2[i][BCOMP]);

-               dst4[i][ACOMP] = USHORT_TO_FLOAT(src2[i][ACOMP]);

-            }

-         }

-         if (useTemp)

-            memcpy(dst, tempBuffer, count * 4 * sizeof(GLfloat));

-      }

-      break;

-   case GL_FLOAT:

-      if (dstType == GL_UNSIGNED_BYTE) {

-         const GLfloat (*src4)[4] = (const GLfloat (*)[4]) src;

-         GLubyte (*dst1)[4] = (GLubyte (*)[4]) (useTemp ? tempBuffer : dst);

-         GLuint i;

-         for (i = 0; i < count; i++) {

-            if (!mask || mask[i]) {

-               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][RCOMP], src4[i][RCOMP]);

-               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][GCOMP], src4[i][GCOMP]);

-               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][BCOMP], src4[i][BCOMP]);

-               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][ACOMP], src4[i][ACOMP]);

-            }

-         }

-         if (useTemp)

-            memcpy(dst, tempBuffer, count * 4 * sizeof(GLubyte));

-      }

-      else {

-         const GLfloat (*src4)[4] = (const GLfloat (*)[4]) src;

-         GLushort (*dst2)[4] = (GLushort (*)[4]) (useTemp ? tempBuffer : dst);

-         GLuint i;

-         ASSERT(dstType == GL_UNSIGNED_SHORT);

-         for (i = 0; i < count; i++) {

-            if (!mask || mask[i]) {

-               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][RCOMP], src4[i][RCOMP]);

-               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][GCOMP], src4[i][GCOMP]);

-               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][BCOMP], src4[i][BCOMP]);

-               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][ACOMP], src4[i][ACOMP]);

-            }

-         }

-         if (useTemp)

-            memcpy(dst, tempBuffer, count * 4 * sizeof(GLushort));

-      }

-      break;

-   default:

-      _mesa_problem(NULL, "Invalid datatype in _mesa_convert_colors");

-   }

-}

-

-

-

-

-/**

- * Perform basic clipping for glDrawPixels.  The image's position and size

- * and the unpack SkipPixels and SkipRows are adjusted so that the image

- * region is entirely within the window and scissor bounds.

- * NOTE: this will only work when glPixelZoom is (1, 1) or (1, -1).

- * If Pixel.ZoomY is -1, *destY will be changed to be the first row which

- * we'll actually write.  Beforehand, *destY-1 is the first drawing row.

- *

- * \return  GL_TRUE if image is ready for drawing or

- *          GL_FALSE if image was completely clipped away (draw nothing)

- */

-GLboolean

-_mesa_clip_drawpixels(const struct gl_context *ctx,

-                      GLint *destX, GLint *destY,

-                      GLsizei *width, GLsizei *height,

-                      struct gl_pixelstore_attrib *unpack)

-{

-   const struct gl_framebuffer *buffer = ctx->DrawBuffer;

-

-   if (unpack->RowLength == 0) {

-      unpack->RowLength = *width;

-   }

-

-   ASSERT(ctx->Pixel.ZoomX == 1.0F);

-   ASSERT(ctx->Pixel.ZoomY == 1.0F || ctx->Pixel.ZoomY == -1.0F);

-

-   /* left clipping */

-   if (*destX < buffer->_Xmin) {

-      unpack->SkipPixels += (buffer->_Xmin - *destX);

-      *width -= (buffer->_Xmin - *destX);

-      *destX = buffer->_Xmin;

-   }

-   /* right clipping */

-   if (*destX + *width > buffer->_Xmax)

-      *width -= (*destX + *width - buffer->_Xmax);

-

-   if (*width <= 0)

-      return GL_FALSE;

-

-   if (ctx->Pixel.ZoomY == 1.0F) {

-      /* bottom clipping */

-      if (*destY < buffer->_Ymin) {

-         unpack->SkipRows += (buffer->_Ymin - *destY);

-         *height -= (buffer->_Ymin - *destY);

-         *destY = buffer->_Ymin;

-      }

-      /* top clipping */

-      if (*destY + *height > buffer->_Ymax)

-         *height -= (*destY + *height - buffer->_Ymax);

-   }

-   else { /* upside down */

-      /* top clipping */

-      if (*destY > buffer->_Ymax) {

-         unpack->SkipRows += (*destY - buffer->_Ymax);

-         *height -= (*destY - buffer->_Ymax);

-         *destY = buffer->_Ymax;

-      }

-      /* bottom clipping */

-      if (*destY - *height < buffer->_Ymin)

-         *height -= (buffer->_Ymin - (*destY - *height));

-      /* adjust destY so it's the first row to write to */

-      (*destY)--;

-   }

-

-   if (*height <= 0)

-      return GL_FALSE;

-

-   return GL_TRUE;

-}

-

-

-/**

- * Perform clipping for glReadPixels.  The image's window position

- * and size, and the pack skipPixels, skipRows and rowLength are adjusted

- * so that the image region is entirely within the window bounds.

- * Note: this is different from _mesa_clip_drawpixels() in that the

- * scissor box is ignored, and we use the bounds of the current readbuffer

- * surface.

- *

- * \return  GL_TRUE if region to read is in bounds

- *          GL_FALSE if region is completely out of bounds (nothing to read)

- */

-GLboolean

-_mesa_clip_readpixels(const struct gl_context *ctx,

-                      GLint *srcX, GLint *srcY,

-                      GLsizei *width, GLsizei *height,

-                      struct gl_pixelstore_attrib *pack)

-{

-   const struct gl_framebuffer *buffer = ctx->ReadBuffer;

-

-   if (pack->RowLength == 0) {

-      pack->RowLength = *width;

-   }

-

-   /* left clipping */

-   if (*srcX < 0) {

-      pack->SkipPixels += (0 - *srcX);

-      *width -= (0 - *srcX);

-      *srcX = 0;

-   }

-   /* right clipping */

-   if (*srcX + *width > (GLsizei) buffer->Width)

-      *width -= (*srcX + *width - buffer->Width);

-

-   if (*width <= 0)

-      return GL_FALSE;

-

-   /* bottom clipping */

-   if (*srcY < 0) {

-      pack->SkipRows += (0 - *srcY);

-      *height -= (0 - *srcY);

-      *srcY = 0;

-   }

-   /* top clipping */

-   if (*srcY + *height > (GLsizei) buffer->Height)

-      *height -= (*srcY + *height - buffer->Height);

-

-   if (*height <= 0)

-      return GL_FALSE;

-

-   return GL_TRUE;

-}

-

-

-/**

- * Do clipping for a glCopyTexSubImage call.

- * The framebuffer source region might extend outside the framebuffer

- * bounds.  Clip the source region against the framebuffer bounds and

- * adjust the texture/dest position and size accordingly.

- *

- * \return GL_FALSE if region is totally clipped, GL_TRUE otherwise.

- */

-GLboolean

-_mesa_clip_copytexsubimage(const struct gl_context *ctx,

-                           GLint *destX, GLint *destY,

-                           GLint *srcX, GLint *srcY,

-                           GLsizei *width, GLsizei *height)

-{

-   const struct gl_framebuffer *fb = ctx->ReadBuffer;

-   const GLint srcX0 = *srcX, srcY0 = *srcY;

-

-   if (_mesa_clip_to_region(0, 0, fb->Width, fb->Height,

-                            srcX, srcY, width, height)) {

-      *destX = *destX + *srcX - srcX0;

-      *destY = *destY + *srcY - srcY0;

-

-      return GL_TRUE;

-   }

-   else {

-      return GL_FALSE;

-   }

-}

-

-

-

-/**

- * Clip the rectangle defined by (x, y, width, height) against the bounds

- * specified by [xmin, xmax) and [ymin, ymax).

- * \return GL_FALSE if rect is totally clipped, GL_TRUE otherwise.

- */

-GLboolean

-_mesa_clip_to_region(GLint xmin, GLint ymin,

-                     GLint xmax, GLint ymax,

-                     GLint *x, GLint *y,

-                     GLsizei *width, GLsizei *height )

-{

-   /* left clipping */

-   if (*x < xmin) {

-      *width -= (xmin - *x);

-      *x = xmin;

-   }

-

-   /* right clipping */

-   if (*x + *width > xmax)

-      *width -= (*x + *width - xmax);

-

-   if (*width <= 0)

-      return GL_FALSE;

-

-   /* bottom (or top) clipping */

-   if (*y < ymin) {

-      *height -= (ymin - *y);

-      *y = ymin;

-   }

-

-   /* top (or bottom) clipping */

-   if (*y + *height > ymax)

-      *height -= (*y + *height - ymax);

-

-   if (*height <= 0)

-      return GL_FALSE;

-

-   return GL_TRUE;

-}

-

-

-/**

- * Clip dst coords against Xmax (or Ymax).

- */

-static INLINE void

-clip_right_or_top(GLint *srcX0, GLint *srcX1,

-                  GLint *dstX0, GLint *dstX1,

-                  GLint maxValue)

-{

-   GLfloat t, bias;

-

-   if (*dstX1 > maxValue) {

-      /* X1 outside right edge */

-      ASSERT(*dstX0 < maxValue); /* X0 should be inside right edge */

-      t = (GLfloat) (maxValue - *dstX0) / (GLfloat) (*dstX1 - *dstX0);

-      /* chop off [t, 1] part */

-      ASSERT(t >= 0.0 && t <= 1.0);

-      *dstX1 = maxValue;

-      bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F;

-      *srcX1 = *srcX0 + (GLint) (t * (*srcX1 - *srcX0) + bias);

-   }

-   else if (*dstX0 > maxValue) {

-      /* X0 outside right edge */

-      ASSERT(*dstX1 < maxValue); /* X1 should be inside right edge */

-      t = (GLfloat) (maxValue - *dstX1) / (GLfloat) (*dstX0 - *dstX1);

-      /* chop off [t, 1] part */

-      ASSERT(t >= 0.0 && t <= 1.0);

-      *dstX0 = maxValue;

-      bias = (*srcX0 < *srcX1) ? -0.5F : 0.5F;

-      *srcX0 = *srcX1 + (GLint) (t * (*srcX0 - *srcX1) + bias);

-   }

-}

-

-

-/**

- * Clip dst coords against Xmin (or Ymin).

- */

-static INLINE void

-clip_left_or_bottom(GLint *srcX0, GLint *srcX1,

-                    GLint *dstX0, GLint *dstX1,

-                    GLint minValue)

-{

-   GLfloat t, bias;

-

-   if (*dstX0 < minValue) {

-      /* X0 outside left edge */

-      ASSERT(*dstX1 > minValue); /* X1 should be inside left edge */

-      t = (GLfloat) (minValue - *dstX0) / (GLfloat) (*dstX1 - *dstX0);

-      /* chop off [0, t] part */

-      ASSERT(t >= 0.0 && t <= 1.0);

-      *dstX0 = minValue;

-      bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F; /* flipped??? */

-      *srcX0 = *srcX0 + (GLint) (t * (*srcX1 - *srcX0) + bias);

-   }

-   else if (*dstX1 < minValue) {

-      /* X1 outside left edge */

-      ASSERT(*dstX0 > minValue); /* X0 should be inside left edge */

-      t = (GLfloat) (minValue - *dstX1) / (GLfloat) (*dstX0 - *dstX1);

-      /* chop off [0, t] part */

-      ASSERT(t >= 0.0 && t <= 1.0);

-      *dstX1 = minValue;

-      bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F;

-      *srcX1 = *srcX1 + (GLint) (t * (*srcX0 - *srcX1) + bias);

-   }

-}

-

-

-/**

- * Do clipping of blit src/dest rectangles.

- * The dest rect is clipped against both the buffer bounds and scissor bounds.

- * The src rect is just clipped against the buffer bounds.

- *

- * When either the src or dest rect is clipped, the other is also clipped

- * proportionately!

- *

- * Note that X0 need not be less than X1 (same for Y) for either the source

- * and dest rects.  That makes the clipping a little trickier.

- *

- * \return GL_TRUE if anything is left to draw, GL_FALSE if totally clipped

- */

-GLboolean

-_mesa_clip_blit(struct gl_context *ctx,

-                GLint *srcX0, GLint *srcY0, GLint *srcX1, GLint *srcY1,

-                GLint *dstX0, GLint *dstY0, GLint *dstX1, GLint *dstY1)

-{

-   const GLint srcXmin = 0;

-   const GLint srcXmax = ctx->ReadBuffer->Width;

-   const GLint srcYmin = 0;

-   const GLint srcYmax = ctx->ReadBuffer->Height;

-

-   /* these include scissor bounds */

-   const GLint dstXmin = ctx->DrawBuffer->_Xmin;

-   const GLint dstXmax = ctx->DrawBuffer->_Xmax;

-   const GLint dstYmin = ctx->DrawBuffer->_Ymin;

-   const GLint dstYmax = ctx->DrawBuffer->_Ymax;

-

-   /*

-   printf("PreClipX:  src: %d .. %d  dst: %d .. %d\n",

-          *srcX0, *srcX1, *dstX0, *dstX1);

-   printf("PreClipY:  src: %d .. %d  dst: %d .. %d\n",

-          *srcY0, *srcY1, *dstY0, *dstY1);

-   */

-

-   /* trivial rejection tests */

-   if (*dstX0 == *dstX1)

-      return GL_FALSE; /* no width */

-   if (*dstX0 <= dstXmin && *dstX1 <= dstXmin)

-      return GL_FALSE; /* totally out (left) of bounds */

-   if (*dstX0 >= dstXmax && *dstX1 >= dstXmax)

-      return GL_FALSE; /* totally out (right) of bounds */

-

-   if (*dstY0 == *dstY1)

-      return GL_FALSE;

-   if (*dstY0 <= dstYmin && *dstY1 <= dstYmin)

-      return GL_FALSE;

-   if (*dstY0 >= dstYmax && *dstY1 >= dstYmax)

-      return GL_FALSE;

-

-   if (*srcX0 == *srcX1)

-      return GL_FALSE;

-   if (*srcX0 <= srcXmin && *srcX1 <= srcXmin)

-      return GL_FALSE;

-   if (*srcX0 >= srcXmax && *srcX1 >= srcXmax)

-      return GL_FALSE;

-

-   if (*srcY0 == *srcY1)

-      return GL_FALSE;

-   if (*srcY0 <= srcYmin && *srcY1 <= srcYmin)

-      return GL_FALSE;

-   if (*srcY0 >= srcYmax && *srcY1 >= srcYmax)

-      return GL_FALSE;

-

-   /*

-    * dest clip

-    */

-   clip_right_or_top(srcX0, srcX1, dstX0, dstX1, dstXmax);

-   clip_right_or_top(srcY0, srcY1, dstY0, dstY1, dstYmax);

-   clip_left_or_bottom(srcX0, srcX1, dstX0, dstX1, dstXmin);

-   clip_left_or_bottom(srcY0, srcY1, dstY0, dstY1, dstYmin);

-

-   /*

-    * src clip (just swap src/dst values from above)

-    */

-   clip_right_or_top(dstX0, dstX1, srcX0, srcX1, srcXmax);

-   clip_right_or_top(dstY0, dstY1, srcY0, srcY1, srcYmax);

-   clip_left_or_bottom(dstX0, dstX1, srcX0, srcX1, srcXmin);

-   clip_left_or_bottom(dstY0, dstY1, srcY0, srcY1, srcYmin);

-

-   /*

-   printf("PostClipX: src: %d .. %d  dst: %d .. %d\n",

-          *srcX0, *srcX1, *dstX0, *dstX1);

-   printf("PostClipY: src: %d .. %d  dst: %d .. %d\n",

-          *srcY0, *srcY1, *dstY0, *dstY1);

-   */

-

-   ASSERT(*dstX0 >= dstXmin);

-   ASSERT(*dstX0 <= dstXmax);

-   ASSERT(*dstX1 >= dstXmin);

-   ASSERT(*dstX1 <= dstXmax);

-

-   ASSERT(*dstY0 >= dstYmin);

-   ASSERT(*dstY0 <= dstYmax);

-   ASSERT(*dstY1 >= dstYmin);

-   ASSERT(*dstY1 <= dstYmax);

-

-   ASSERT(*srcX0 >= srcXmin);

-   ASSERT(*srcX0 <= srcXmax);

-   ASSERT(*srcX1 >= srcXmin);

-   ASSERT(*srcX1 <= srcXmax);

-

-   ASSERT(*srcY0 >= srcYmin);

-   ASSERT(*srcY0 <= srcYmax);

-   ASSERT(*srcY1 >= srcYmin);

-   ASSERT(*srcY1 <= srcYmax);

-

-   return GL_TRUE;

-}

+/*
+ * 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 image.c
+ * Image handling.
+ */
+
+
+#include "glheader.h"
+#include "colormac.h"
+#include "image.h"
+#include "imports.h"
+#include "macros.h"
+#include "mfeatures.h"
+#include "mtypes.h"
+
+
+/**
+ * NOTE:
+ * Normally, BYTE_TO_FLOAT(0) returns 0.00392  That causes problems when
+ * we later convert the float to a packed integer value (such as for
+ * GL_RGB5_A1) because we'll wind up with a non-zero value.
+ *
+ * We redefine the macros here so zero is handled correctly.
+ */
+#undef BYTE_TO_FLOAT
+#define BYTE_TO_FLOAT(B)    ((B) == 0 ? 0.0F : ((2.0F * (B) + 1.0F) * (1.0F/255.0F)))
+
+#undef SHORT_TO_FLOAT
+#define SHORT_TO_FLOAT(S)   ((S) == 0 ? 0.0F : ((2.0F * (S) + 1.0F) * (1.0F/65535.0F)))
+
+
+
+/** Compute ceiling of integer quotient of A divided by B. */
+#define CEILING( A, B )  ( (A) % (B) == 0 ? (A)/(B) : (A)/(B)+1 )
+
+
+/**
+ * \return GL_TRUE if type is packed pixel type, GL_FALSE otherwise.
+ */
+GLboolean
+_mesa_type_is_packed(GLenum type)
+{
+   switch (type) {
+   case GL_UNSIGNED_BYTE_3_3_2:
+   case GL_UNSIGNED_BYTE_2_3_3_REV:
+   case MESA_UNSIGNED_BYTE_4_4:
+   case GL_UNSIGNED_SHORT_5_6_5:
+   case GL_UNSIGNED_SHORT_5_6_5_REV:
+   case GL_UNSIGNED_SHORT_4_4_4_4:
+   case GL_UNSIGNED_SHORT_4_4_4_4_REV:
+   case GL_UNSIGNED_SHORT_5_5_5_1:
+   case GL_UNSIGNED_SHORT_1_5_5_5_REV:
+   case GL_UNSIGNED_INT_8_8_8_8:
+   case GL_UNSIGNED_INT_8_8_8_8_REV:
+   case GL_UNSIGNED_INT_10_10_10_2:
+   case GL_UNSIGNED_INT_2_10_10_10_REV:
+   case GL_UNSIGNED_SHORT_8_8_MESA:
+   case GL_UNSIGNED_SHORT_8_8_REV_MESA:
+   case GL_UNSIGNED_INT_24_8_EXT:
+   case GL_UNSIGNED_INT_5_9_9_9_REV:
+   case GL_UNSIGNED_INT_10F_11F_11F_REV:
+   case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+      return GL_TRUE;
+   }
+
+   return GL_FALSE;
+}
+
+
+
+/**
+ * Flip the order of the 2 bytes in each word in the given array.
+ *
+ * \param p array.
+ * \param n number of words.
+ */
+void
+_mesa_swap2( GLushort *p, GLuint n )
+{
+   GLuint i;
+   for (i = 0; i < n; i++) {
+      p[i] = (p[i] >> 8) | ((p[i] << 8) & 0xff00);
+   }
+}
+
+
+
+/*
+ * Flip the order of the 4 bytes in each word in the given array.
+ */
+void
+_mesa_swap4( GLuint *p, GLuint n )
+{
+   GLuint i, a, b;
+   for (i = 0; i < n; i++) {
+      b = p[i];
+      a =  (b >> 24)
+	| ((b >> 8) & 0xff00)
+	| ((b << 8) & 0xff0000)
+	| ((b << 24) & 0xff000000);
+      p[i] = a;
+   }
+}
+
+
+/**
+ * Get the size of a GL data type.
+ *
+ * \param type GL data type.
+ *
+ * \return the size, in bytes, of the given data type, 0 if a GL_BITMAP, or -1
+ * if an invalid type enum.
+ */
+GLint
+_mesa_sizeof_type( GLenum type )
+{
+   switch (type) {
+      case GL_BITMAP:
+	 return 0;
+      case GL_UNSIGNED_BYTE:
+         return sizeof(GLubyte);
+      case GL_BYTE:
+	 return sizeof(GLbyte);
+      case GL_UNSIGNED_SHORT:
+	 return sizeof(GLushort);
+      case GL_SHORT:
+	 return sizeof(GLshort);
+      case GL_UNSIGNED_INT:
+	 return sizeof(GLuint);
+      case GL_INT:
+	 return sizeof(GLint);
+      case GL_FLOAT:
+	 return sizeof(GLfloat);
+      case GL_DOUBLE:
+	 return sizeof(GLdouble);
+      case GL_HALF_FLOAT_ARB:
+	 return sizeof(GLhalfARB);
+      case GL_FIXED:
+	 return sizeof(GLfixed);
+      default:
+         return -1;
+   }
+}
+
+
+/**
+ * Same as _mesa_sizeof_type() but also accepting the packed pixel
+ * format data types.
+ */
+GLint
+_mesa_sizeof_packed_type( GLenum type )
+{
+   switch (type) {
+      case GL_BITMAP:
+	 return 0;
+      case GL_UNSIGNED_BYTE:
+         return sizeof(GLubyte);
+      case GL_BYTE:
+	 return sizeof(GLbyte);
+      case GL_UNSIGNED_SHORT:
+	 return sizeof(GLushort);
+      case GL_SHORT:
+	 return sizeof(GLshort);
+      case GL_UNSIGNED_INT:
+	 return sizeof(GLuint);
+      case GL_INT:
+	 return sizeof(GLint);
+      case GL_HALF_FLOAT_ARB:
+	 return sizeof(GLhalfARB);
+      case GL_FLOAT:
+	 return sizeof(GLfloat);
+      case GL_UNSIGNED_BYTE_3_3_2:
+         return sizeof(GLubyte);
+      case GL_UNSIGNED_BYTE_2_3_3_REV:
+         return sizeof(GLubyte);
+      case MESA_UNSIGNED_BYTE_4_4:
+         return sizeof(GLubyte);
+      case GL_UNSIGNED_SHORT_5_6_5:
+         return sizeof(GLushort);
+      case GL_UNSIGNED_SHORT_5_6_5_REV:
+         return sizeof(GLushort);
+      case GL_UNSIGNED_SHORT_4_4_4_4:
+         return sizeof(GLushort);
+      case GL_UNSIGNED_SHORT_4_4_4_4_REV:
+         return sizeof(GLushort);
+      case GL_UNSIGNED_SHORT_5_5_5_1:
+         return sizeof(GLushort);
+      case GL_UNSIGNED_SHORT_1_5_5_5_REV:
+         return sizeof(GLushort);
+      case GL_UNSIGNED_INT_8_8_8_8:
+         return sizeof(GLuint);
+      case GL_UNSIGNED_INT_8_8_8_8_REV:
+         return sizeof(GLuint);
+      case GL_UNSIGNED_INT_10_10_10_2:
+         return sizeof(GLuint);
+      case GL_UNSIGNED_INT_2_10_10_10_REV:
+         return sizeof(GLuint);
+      case GL_UNSIGNED_SHORT_8_8_MESA:
+      case GL_UNSIGNED_SHORT_8_8_REV_MESA:
+         return sizeof(GLushort);      
+      case GL_UNSIGNED_INT_24_8_EXT:
+         return sizeof(GLuint);
+      case GL_UNSIGNED_INT_5_9_9_9_REV:
+         return sizeof(GLuint);
+      case GL_UNSIGNED_INT_10F_11F_11F_REV:
+         return sizeof(GLuint);
+      case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+         return 8;
+      default:
+         return -1;
+   }
+}
+
+
+/**
+ * Get the number of components in a pixel format.
+ *
+ * \param format pixel format.
+ *
+ * \return the number of components in the given format, or -1 if a bad format.
+ */
+GLint
+_mesa_components_in_format( GLenum format )
+{
+   switch (format) {
+      case GL_COLOR_INDEX:
+      case GL_COLOR_INDEX1_EXT:
+      case GL_COLOR_INDEX2_EXT:
+      case GL_COLOR_INDEX4_EXT:
+      case GL_COLOR_INDEX8_EXT:
+      case GL_COLOR_INDEX12_EXT:
+      case GL_COLOR_INDEX16_EXT:
+      case GL_STENCIL_INDEX:
+      case GL_DEPTH_COMPONENT:
+      case GL_RED:
+      case GL_RED_INTEGER_EXT:
+      case GL_GREEN:
+      case GL_GREEN_INTEGER_EXT:
+      case GL_BLUE:
+      case GL_BLUE_INTEGER_EXT:
+      case GL_ALPHA:
+      case GL_ALPHA_INTEGER_EXT:
+      case GL_LUMINANCE:
+      case GL_LUMINANCE_INTEGER_EXT:
+      case GL_INTENSITY:
+         return 1;
+      case GL_LUMINANCE_ALPHA:
+      case GL_LUMINANCE_ALPHA_INTEGER_EXT:
+      case GL_RG:
+	 return 2;
+      case GL_RGB:
+      case GL_RGB_INTEGER_EXT:
+	 return 3;
+      case GL_RGBA:
+      case GL_RGBA_INTEGER_EXT:
+	 return 4;
+      case GL_BGR:
+	 return 3;
+      case GL_BGRA:
+	 return 4;
+      case GL_ABGR_EXT:
+         return 4;
+      case GL_YCBCR_MESA:
+         return 2;
+      case GL_DEPTH_STENCIL_EXT:
+         return 2;
+      case GL_DUDV_ATI:
+      case GL_DU8DV8_ATI:
+         return 2;
+      default:
+         return -1;
+   }
+}
+
+
+/**
+ * Get the bytes per pixel of pixel format type pair.
+ *
+ * \param format pixel format.
+ * \param type pixel type.
+ *
+ * \return bytes per pixel, or -1 if a bad format or type was given.
+ */
+GLint
+_mesa_bytes_per_pixel( GLenum format, GLenum type )
+{
+   GLint comps = _mesa_components_in_format( format );
+   if (comps < 0)
+      return -1;
+
+   switch (type) {
+      case GL_BITMAP:
+         return 0;  /* special case */
+      case GL_BYTE:
+      case GL_UNSIGNED_BYTE:
+         return comps * sizeof(GLubyte);
+      case GL_SHORT:
+      case GL_UNSIGNED_SHORT:
+         return comps * sizeof(GLshort);
+      case GL_INT:
+      case GL_UNSIGNED_INT:
+         return comps * sizeof(GLint);
+      case GL_FLOAT:
+         return comps * sizeof(GLfloat);
+      case GL_HALF_FLOAT_ARB:
+         return comps * sizeof(GLhalfARB);
+      case GL_UNSIGNED_BYTE_3_3_2:
+      case GL_UNSIGNED_BYTE_2_3_3_REV:
+         if (format == GL_RGB || format == GL_BGR ||
+             format == GL_RGB_INTEGER_EXT || format == GL_BGR_INTEGER_EXT)
+            return sizeof(GLubyte);
+         else
+            return -1;  /* error */
+      case GL_UNSIGNED_SHORT_5_6_5:
+      case GL_UNSIGNED_SHORT_5_6_5_REV:
+         if (format == GL_RGB || format == GL_BGR ||
+             format == GL_RGB_INTEGER_EXT || format == GL_BGR_INTEGER_EXT)
+            return sizeof(GLushort);
+         else
+            return -1;  /* error */
+      case GL_UNSIGNED_SHORT_4_4_4_4:
+      case GL_UNSIGNED_SHORT_4_4_4_4_REV:
+      case GL_UNSIGNED_SHORT_5_5_5_1:
+      case GL_UNSIGNED_SHORT_1_5_5_5_REV:
+         if (format == GL_RGBA || format == GL_BGRA || format == GL_ABGR_EXT ||
+             format == GL_RGBA_INTEGER_EXT || format == GL_BGRA_INTEGER_EXT)
+            return sizeof(GLushort);
+         else
+            return -1;
+      case GL_UNSIGNED_INT_8_8_8_8:
+      case GL_UNSIGNED_INT_8_8_8_8_REV:
+      case GL_UNSIGNED_INT_10_10_10_2:
+      case GL_UNSIGNED_INT_2_10_10_10_REV:
+         if (format == GL_RGBA || format == GL_BGRA || format == GL_ABGR_EXT ||
+             format == GL_RGBA_INTEGER_EXT || format == GL_BGRA_INTEGER_EXT)
+            return sizeof(GLuint);
+         else
+            return -1;
+      case GL_UNSIGNED_SHORT_8_8_MESA:
+      case GL_UNSIGNED_SHORT_8_8_REV_MESA:
+         if (format == GL_YCBCR_MESA)
+            return sizeof(GLushort);
+         else
+            return -1;
+      case GL_UNSIGNED_INT_24_8_EXT:
+         if (format == GL_DEPTH_STENCIL_EXT)
+            return sizeof(GLuint);
+         else
+            return -1;
+      case GL_UNSIGNED_INT_5_9_9_9_REV:
+         if (format == GL_RGB)
+            return sizeof(GLuint);
+         else
+            return -1;
+      case GL_UNSIGNED_INT_10F_11F_11F_REV:
+         if (format == GL_RGB)
+            return sizeof(GLuint);
+         else
+            return -1;
+      case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+         if (format == GL_DEPTH_STENCIL)
+            return 8;
+         else
+            return -1;
+      default:
+         return -1;
+   }
+}
+
+
+/**
+ * Test for a legal pixel format and type.
+ *
+ * \param format pixel format.
+ * \param type pixel type.
+ *
+ * \return GL_TRUE if the given pixel format and type are legal, or GL_FALSE
+ * otherwise.
+ */
+GLboolean
+_mesa_is_legal_format_and_type(const struct gl_context *ctx,
+                               GLenum format, GLenum type)
+{
+   switch (format) {
+      case GL_COLOR_INDEX:
+      case GL_STENCIL_INDEX:
+         switch (type) {
+            case GL_BITMAP:
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+               return GL_TRUE;
+            case GL_HALF_FLOAT_ARB:
+               return ctx->Extensions.ARB_half_float_pixel;
+            default:
+               return GL_FALSE;
+         }
+      case GL_RED:
+      case GL_GREEN:
+      case GL_BLUE:
+      case GL_ALPHA:
+#if 0 /* not legal!  see table 3.6 of the 1.5 spec */
+      case GL_INTENSITY:
+#endif
+      case GL_LUMINANCE:
+      case GL_LUMINANCE_ALPHA:
+      case GL_DEPTH_COMPONENT:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+               return GL_TRUE;
+            case GL_HALF_FLOAT_ARB:
+               return ctx->Extensions.ARB_half_float_pixel;
+            default:
+               return GL_FALSE;
+         }
+      case GL_RG:
+	 if (!ctx->Extensions.ARB_texture_rg)
+	    return GL_FALSE;
+
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+               return GL_TRUE;
+            case GL_HALF_FLOAT_ARB:
+               return ctx->Extensions.ARB_half_float_pixel;
+            default:
+               return GL_FALSE;
+         }
+      case GL_RGB:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+            case GL_UNSIGNED_BYTE_3_3_2:
+            case GL_UNSIGNED_BYTE_2_3_3_REV:
+            case GL_UNSIGNED_SHORT_5_6_5:
+            case GL_UNSIGNED_SHORT_5_6_5_REV:
+               return GL_TRUE;
+            case GL_HALF_FLOAT_ARB:
+               return ctx->Extensions.ARB_half_float_pixel;
+            case GL_UNSIGNED_INT_5_9_9_9_REV:
+               return ctx->Extensions.EXT_texture_shared_exponent;
+            case GL_UNSIGNED_INT_10F_11F_11F_REV:
+               return ctx->Extensions.EXT_packed_float;
+            default:
+               return GL_FALSE;
+         }
+      case GL_BGR:
+         switch (type) {
+            /* NOTE: no packed types are supported with BGR.  That's
+             * intentional, according to the GL spec.
+             */
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+               return GL_TRUE;
+            case GL_HALF_FLOAT_ARB:
+               return ctx->Extensions.ARB_half_float_pixel;
+            default:
+               return GL_FALSE;
+         }
+      case GL_RGBA:
+      case GL_BGRA:
+      case GL_ABGR_EXT:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+            case GL_UNSIGNED_SHORT_4_4_4_4:
+            case GL_UNSIGNED_SHORT_4_4_4_4_REV:
+            case GL_UNSIGNED_SHORT_5_5_5_1:
+            case GL_UNSIGNED_SHORT_1_5_5_5_REV:
+            case GL_UNSIGNED_INT_8_8_8_8:
+            case GL_UNSIGNED_INT_8_8_8_8_REV:
+            case GL_UNSIGNED_INT_10_10_10_2:
+            case GL_UNSIGNED_INT_2_10_10_10_REV:
+               return GL_TRUE;
+            case GL_HALF_FLOAT_ARB:
+               return ctx->Extensions.ARB_half_float_pixel;
+            default:
+               return GL_FALSE;
+         }
+      case GL_YCBCR_MESA:
+         if (type == GL_UNSIGNED_SHORT_8_8_MESA ||
+             type == GL_UNSIGNED_SHORT_8_8_REV_MESA)
+            return GL_TRUE;
+         else
+            return GL_FALSE;
+      case GL_DEPTH_STENCIL_EXT:
+         if ((ctx->Extensions.EXT_packed_depth_stencil &&
+              type == GL_UNSIGNED_INT_24_8_EXT) ||
+             (ctx->Extensions.ARB_depth_buffer_float &&
+              type == GL_FLOAT_32_UNSIGNED_INT_24_8_REV))
+            return GL_TRUE;
+         else
+            return GL_FALSE;
+      case GL_DUDV_ATI:
+      case GL_DU8DV8_ATI:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_FLOAT:
+               return GL_TRUE;
+            default:
+               return GL_FALSE;
+         }
+
+      /* integer-valued formats */
+      case GL_RED_INTEGER_EXT:
+      case GL_GREEN_INTEGER_EXT:
+      case GL_BLUE_INTEGER_EXT:
+      case GL_ALPHA_INTEGER_EXT:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+               return ctx->Extensions.EXT_texture_integer;
+            default:
+               return GL_FALSE;
+         }
+
+      case GL_RGB_INTEGER_EXT:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_UNSIGNED_BYTE_3_3_2:
+            case GL_UNSIGNED_BYTE_2_3_3_REV:
+            case GL_UNSIGNED_SHORT_5_6_5:
+            case GL_UNSIGNED_SHORT_5_6_5_REV:
+               return ctx->Extensions.EXT_texture_integer;
+            default:
+               return GL_FALSE;
+         }
+
+      case GL_BGR_INTEGER_EXT:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            /* NOTE: no packed formats w/ BGR format */
+               return ctx->Extensions.EXT_texture_integer;
+            default:
+               return GL_FALSE;
+         }
+
+      case GL_RGBA_INTEGER_EXT:
+      case GL_BGRA_INTEGER_EXT:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+            case GL_UNSIGNED_SHORT_4_4_4_4:
+            case GL_UNSIGNED_SHORT_4_4_4_4_REV:
+            case GL_UNSIGNED_SHORT_5_5_5_1:
+            case GL_UNSIGNED_SHORT_1_5_5_5_REV:
+            case GL_UNSIGNED_INT_8_8_8_8:
+            case GL_UNSIGNED_INT_8_8_8_8_REV:
+            case GL_UNSIGNED_INT_10_10_10_2:
+            case GL_UNSIGNED_INT_2_10_10_10_REV:
+               return ctx->Extensions.EXT_texture_integer;
+            default:
+               return GL_FALSE;
+         }
+
+      case GL_LUMINANCE_INTEGER_EXT:
+      case GL_LUMINANCE_ALPHA_INTEGER_EXT:
+         switch (type) {
+            case GL_BYTE:
+            case GL_UNSIGNED_BYTE:
+            case GL_SHORT:
+            case GL_UNSIGNED_SHORT:
+            case GL_INT:
+            case GL_UNSIGNED_INT:
+               return ctx->Extensions.EXT_texture_integer;
+            default:
+               return GL_FALSE;
+         }
+
+      default:
+         ; /* fall-through */
+   }
+   return GL_FALSE;
+}
+
+
+/**
+ * Test if the given image format is a color/RGBA format (i.e., not color
+ * index, depth, stencil, etc).
+ * \param format  the image format value (may by an internal texture format)
+ * \return GL_TRUE if its a color/RGBA format, GL_FALSE otherwise.
+ */
+GLboolean
+_mesa_is_color_format(GLenum format)
+{
+   switch (format) {
+      case GL_RED:
+      case GL_GREEN:
+      case GL_BLUE:
+      case GL_ALPHA:
+      case GL_ALPHA4:
+      case GL_ALPHA8:
+      case GL_ALPHA12:
+      case GL_ALPHA16:
+      case 1:
+      case GL_LUMINANCE:
+      case GL_LUMINANCE4:
+      case GL_LUMINANCE8:
+      case GL_LUMINANCE12:
+      case GL_LUMINANCE16:
+      case 2:
+      case GL_LUMINANCE_ALPHA:
+      case GL_LUMINANCE4_ALPHA4:
+      case GL_LUMINANCE6_ALPHA2:
+      case GL_LUMINANCE8_ALPHA8:
+      case GL_LUMINANCE12_ALPHA4:
+      case GL_LUMINANCE12_ALPHA12:
+      case GL_LUMINANCE16_ALPHA16:
+      case GL_INTENSITY:
+      case GL_INTENSITY4:
+      case GL_INTENSITY8:
+      case GL_INTENSITY12:
+      case GL_INTENSITY16:
+      case GL_R8:
+      case GL_R16:
+      case GL_RG:
+      case GL_RG8:
+      case GL_RG16:
+      case 3:
+      case GL_RGB:
+      case GL_BGR:
+      case GL_R3_G3_B2:
+      case GL_RGB4:
+      case GL_RGB5:
+      case GL_RGB8:
+      case GL_RGB10:
+      case GL_RGB12:
+      case GL_RGB16:
+      case 4:
+      case GL_ABGR_EXT:
+      case GL_RGBA:
+      case GL_BGRA:
+      case GL_RGBA2:
+      case GL_RGBA4:
+      case GL_RGB5_A1:
+      case GL_RGBA8:
+      case GL_RGB10_A2:
+      case GL_RGBA12:
+      case GL_RGBA16:
+      /* float texture formats */
+      case GL_ALPHA16F_ARB:
+      case GL_ALPHA32F_ARB:
+      case GL_LUMINANCE16F_ARB:
+      case GL_LUMINANCE32F_ARB:
+      case GL_LUMINANCE_ALPHA16F_ARB:
+      case GL_LUMINANCE_ALPHA32F_ARB:
+      case GL_INTENSITY16F_ARB:
+      case GL_INTENSITY32F_ARB:
+      case GL_R16F:
+      case GL_R32F:
+      case GL_RG16F:
+      case GL_RG32F:
+      case GL_RGB16F_ARB:
+      case GL_RGB32F_ARB:
+      case GL_RGBA16F_ARB:
+      case GL_RGBA32F_ARB:
+      /* compressed formats */
+      case GL_COMPRESSED_ALPHA:
+      case GL_COMPRESSED_LUMINANCE:
+      case GL_COMPRESSED_LUMINANCE_ALPHA:
+      case GL_COMPRESSED_INTENSITY:
+      case GL_COMPRESSED_RED:
+      case GL_COMPRESSED_RG:
+      case GL_COMPRESSED_RGB:
+      case GL_COMPRESSED_RGBA:
+      case GL_RGB_S3TC:
+      case GL_RGB4_S3TC:
+      case GL_RGBA_S3TC:
+      case GL_RGBA4_S3TC:
+      case GL_COMPRESSED_RGB_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_RGBA_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_RGBA_S3TC_DXT3_EXT:
+      case GL_COMPRESSED_RGBA_S3TC_DXT5_EXT:
+      case GL_COMPRESSED_RGB_FXT1_3DFX:
+      case GL_COMPRESSED_RGBA_FXT1_3DFX:
+#if FEATURE_EXT_texture_sRGB
+      case GL_SRGB_EXT:
+      case GL_SRGB8_EXT:
+      case GL_SRGB_ALPHA_EXT:
+      case GL_SRGB8_ALPHA8_EXT:
+      case GL_SLUMINANCE_ALPHA_EXT:
+      case GL_SLUMINANCE8_ALPHA8_EXT:
+      case GL_SLUMINANCE_EXT:
+      case GL_SLUMINANCE8_EXT:
+      case GL_COMPRESSED_SRGB_EXT:
+      case GL_COMPRESSED_SRGB_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_SRGB_ALPHA_EXT:
+      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT:
+      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT:
+      case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT:
+      case GL_COMPRESSED_SLUMINANCE_EXT:
+      case GL_COMPRESSED_SLUMINANCE_ALPHA_EXT:
+#endif /* FEATURE_EXT_texture_sRGB */
+      case GL_COMPRESSED_RED_RGTC1:
+      case GL_COMPRESSED_SIGNED_RED_RGTC1:
+      case GL_COMPRESSED_RG_RGTC2:
+      case GL_COMPRESSED_SIGNED_RG_RGTC2:
+      case GL_COMPRESSED_LUMINANCE_LATC1_EXT:
+      case GL_COMPRESSED_SIGNED_LUMINANCE_LATC1_EXT:
+      case GL_COMPRESSED_LUMINANCE_ALPHA_LATC2_EXT:
+      case GL_COMPRESSED_SIGNED_LUMINANCE_ALPHA_LATC2_EXT:
+      case GL_COMPRESSED_LUMINANCE_ALPHA_3DC_ATI:
+      /* generic integer formats */
+      case GL_RED_INTEGER_EXT:
+      case GL_GREEN_INTEGER_EXT:
+      case GL_BLUE_INTEGER_EXT:
+      case GL_ALPHA_INTEGER_EXT:
+      case GL_RGB_INTEGER_EXT:
+      case GL_RGBA_INTEGER_EXT:
+      case GL_BGR_INTEGER_EXT:
+      case GL_BGRA_INTEGER_EXT:
+      case GL_LUMINANCE_INTEGER_EXT:
+      case GL_LUMINANCE_ALPHA_INTEGER_EXT:
+      /* sized integer formats */
+      case GL_RGBA32UI_EXT:
+      case GL_RGB32UI_EXT:
+      case GL_ALPHA32UI_EXT:
+      case GL_INTENSITY32UI_EXT:
+      case GL_LUMINANCE32UI_EXT:
+      case GL_LUMINANCE_ALPHA32UI_EXT:
+      case GL_RGBA16UI_EXT:
+      case GL_RGB16UI_EXT:
+      case GL_ALPHA16UI_EXT:
+      case GL_INTENSITY16UI_EXT:
+      case GL_LUMINANCE16UI_EXT:
+      case GL_LUMINANCE_ALPHA16UI_EXT:
+      case GL_RGBA8UI_EXT:
+      case GL_RGB8UI_EXT:
+      case GL_ALPHA8UI_EXT:
+      case GL_INTENSITY8UI_EXT:
+      case GL_LUMINANCE8UI_EXT:
+      case GL_LUMINANCE_ALPHA8UI_EXT:
+      case GL_RGBA32I_EXT:
+      case GL_RGB32I_EXT:
+      case GL_ALPHA32I_EXT:
+      case GL_INTENSITY32I_EXT:
+      case GL_LUMINANCE32I_EXT:
+      case GL_LUMINANCE_ALPHA32I_EXT:
+      case GL_RGBA16I_EXT:
+      case GL_RGB16I_EXT:
+      case GL_ALPHA16I_EXT:
+      case GL_INTENSITY16I_EXT:
+      case GL_LUMINANCE16I_EXT:
+      case GL_LUMINANCE_ALPHA16I_EXT:
+      case GL_RGBA8I_EXT:
+      case GL_RGB8I_EXT:
+      case GL_ALPHA8I_EXT:
+      case GL_INTENSITY8I_EXT:
+      case GL_LUMINANCE8I_EXT:
+      case GL_LUMINANCE_ALPHA8I_EXT:
+      /* signed, normalized texture formats */
+      case GL_RED_SNORM:
+      case GL_R8_SNORM:
+      case GL_R16_SNORM:
+      case GL_RG_SNORM:
+      case GL_RG8_SNORM:
+      case GL_RG16_SNORM:
+      case GL_RGB_SNORM:
+      case GL_RGB8_SNORM:
+      case GL_RGB16_SNORM:
+      case GL_RGBA_SNORM:
+      case GL_RGBA8_SNORM:
+      case GL_RGBA16_SNORM:
+      case GL_ALPHA_SNORM:
+      case GL_ALPHA8_SNORM:
+      case GL_ALPHA16_SNORM:
+      case GL_LUMINANCE_SNORM:
+      case GL_LUMINANCE8_SNORM:
+      case GL_LUMINANCE16_SNORM:
+      case GL_LUMINANCE_ALPHA_SNORM:
+      case GL_LUMINANCE8_ALPHA8_SNORM:
+      case GL_LUMINANCE16_ALPHA16_SNORM:
+      case GL_INTENSITY_SNORM:
+      case GL_INTENSITY8_SNORM:
+      case GL_INTENSITY16_SNORM:
+      case GL_RGB9_E5:
+      case GL_R11F_G11F_B10F:
+         return GL_TRUE;
+      case GL_YCBCR_MESA:  /* not considered to be RGB */
+         /* fall-through */
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given image format is a color index format.
+ */
+GLboolean
+_mesa_is_index_format(GLenum format)
+{
+   switch (format) {
+      case GL_COLOR_INDEX:
+      case GL_COLOR_INDEX1_EXT:
+      case GL_COLOR_INDEX2_EXT:
+      case GL_COLOR_INDEX4_EXT:
+      case GL_COLOR_INDEX8_EXT:
+      case GL_COLOR_INDEX12_EXT:
+      case GL_COLOR_INDEX16_EXT:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given image format is a depth component format.
+ */
+GLboolean
+_mesa_is_depth_format(GLenum format)
+{
+   switch (format) {
+      case GL_DEPTH_COMPONENT:
+      case GL_DEPTH_COMPONENT16:
+      case GL_DEPTH_COMPONENT24:
+      case GL_DEPTH_COMPONENT32:
+      case GL_DEPTH_COMPONENT32F:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given image format is a stencil format.
+ */
+GLboolean
+_mesa_is_stencil_format(GLenum format)
+{
+   switch (format) {
+      case GL_STENCIL_INDEX:
+      case GL_DEPTH_STENCIL:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given image format is a YCbCr format.
+ */
+GLboolean
+_mesa_is_ycbcr_format(GLenum format)
+{
+   switch (format) {
+      case GL_YCBCR_MESA:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given image format is a depth+stencil format.
+ */
+GLboolean
+_mesa_is_depthstencil_format(GLenum format)
+{
+   switch (format) {
+      case GL_DEPTH24_STENCIL8_EXT:
+      case GL_DEPTH_STENCIL_EXT:
+      case GL_DEPTH32F_STENCIL8:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given image format is a depth or stencil format.
+ */
+GLboolean
+_mesa_is_depth_or_stencil_format(GLenum format)
+{
+   switch (format) {
+      case GL_DEPTH_COMPONENT:
+      case GL_DEPTH_COMPONENT16:
+      case GL_DEPTH_COMPONENT24:
+      case GL_DEPTH_COMPONENT32:
+      case GL_STENCIL_INDEX:
+      case GL_STENCIL_INDEX1_EXT:
+      case GL_STENCIL_INDEX4_EXT:
+      case GL_STENCIL_INDEX8_EXT:
+      case GL_STENCIL_INDEX16_EXT:
+      case GL_DEPTH_STENCIL_EXT:
+      case GL_DEPTH24_STENCIL8_EXT:
+      case GL_DEPTH_COMPONENT32F:
+      case GL_DEPTH32F_STENCIL8:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given image format is a dudv format.
+ */
+GLboolean
+_mesa_is_dudv_format(GLenum format)
+{
+   switch (format) {
+      case GL_DUDV_ATI:
+      case GL_DU8DV8_ATI:
+         return GL_TRUE;
+      default:
+         return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if the given format is an integer (non-normalized) format.
+ */
+GLboolean
+_mesa_is_integer_format(GLenum format)
+{
+   switch (format) {
+   /* generic integer formats */
+   case GL_RED_INTEGER_EXT:
+   case GL_GREEN_INTEGER_EXT:
+   case GL_BLUE_INTEGER_EXT:
+   case GL_ALPHA_INTEGER_EXT:
+   case GL_RGB_INTEGER_EXT:
+   case GL_RGBA_INTEGER_EXT:
+   case GL_BGR_INTEGER_EXT:
+   case GL_BGRA_INTEGER_EXT:
+   case GL_LUMINANCE_INTEGER_EXT:
+   case GL_LUMINANCE_ALPHA_INTEGER_EXT:
+   /* specific integer formats */
+   case GL_RGBA32UI_EXT:
+   case GL_RGB32UI_EXT:
+   case GL_ALPHA32UI_EXT:
+   case GL_INTENSITY32UI_EXT:
+   case GL_LUMINANCE32UI_EXT:
+   case GL_LUMINANCE_ALPHA32UI_EXT:
+   case GL_RGBA16UI_EXT:
+   case GL_RGB16UI_EXT:
+   case GL_ALPHA16UI_EXT:
+   case GL_INTENSITY16UI_EXT:
+   case GL_LUMINANCE16UI_EXT:
+   case GL_LUMINANCE_ALPHA16UI_EXT:
+   case GL_RGBA8UI_EXT:
+   case GL_RGB8UI_EXT:
+   case GL_ALPHA8UI_EXT:
+   case GL_INTENSITY8UI_EXT:
+   case GL_LUMINANCE8UI_EXT:
+   case GL_LUMINANCE_ALPHA8UI_EXT:
+   case GL_RGBA32I_EXT:
+   case GL_RGB32I_EXT:
+   case GL_ALPHA32I_EXT:
+   case GL_INTENSITY32I_EXT:
+   case GL_LUMINANCE32I_EXT:
+   case GL_LUMINANCE_ALPHA32I_EXT:
+   case GL_RGBA16I_EXT:
+   case GL_RGB16I_EXT:
+   case GL_ALPHA16I_EXT:
+   case GL_INTENSITY16I_EXT:
+   case GL_LUMINANCE16I_EXT:
+   case GL_LUMINANCE_ALPHA16I_EXT:
+   case GL_RGBA8I_EXT:
+   case GL_RGB8I_EXT:
+   case GL_ALPHA8I_EXT:
+   case GL_INTENSITY8I_EXT:
+   case GL_LUMINANCE8I_EXT:
+   case GL_LUMINANCE_ALPHA8I_EXT:
+      return GL_TRUE;
+   default:
+      return GL_FALSE;
+   }
+}
+
+
+/**
+ * Test if an image format is a supported compressed format.
+ * \param format the internal format token provided by the user.
+ * \return GL_TRUE if compressed, GL_FALSE if uncompressed
+ */
+GLboolean
+_mesa_is_compressed_format(struct gl_context *ctx, GLenum format)
+{
+   switch (format) {
+   case GL_COMPRESSED_RGB_S3TC_DXT1_EXT:
+   case GL_COMPRESSED_RGBA_S3TC_DXT1_EXT:
+   case GL_COMPRESSED_RGBA_S3TC_DXT3_EXT:
+   case GL_COMPRESSED_RGBA_S3TC_DXT5_EXT:
+      return ctx->Extensions.EXT_texture_compression_s3tc;
+   case GL_RGB_S3TC:
+   case GL_RGB4_S3TC:
+   case GL_RGBA_S3TC:
+   case GL_RGBA4_S3TC:
+      return ctx->Extensions.S3_s3tc;
+   case GL_COMPRESSED_SRGB_S3TC_DXT1_EXT:
+   case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT1_EXT:
+   case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT3_EXT:
+   case GL_COMPRESSED_SRGB_ALPHA_S3TC_DXT5_EXT:
+      return ctx->Extensions.EXT_texture_sRGB
+         && ctx->Extensions.EXT_texture_compression_s3tc;
+   case GL_COMPRESSED_RGB_FXT1_3DFX:
+   case GL_COMPRESSED_RGBA_FXT1_3DFX:
+      return ctx->Extensions.TDFX_texture_compression_FXT1;
+   case GL_COMPRESSED_RED_RGTC1:
+   case GL_COMPRESSED_SIGNED_RED_RGTC1:
+   case GL_COMPRESSED_RG_RGTC2:
+   case GL_COMPRESSED_SIGNED_RG_RGTC2:
+      return ctx->Extensions.ARB_texture_compression_rgtc;
+   case GL_COMPRESSED_LUMINANCE_LATC1_EXT:
+   case GL_COMPRESSED_SIGNED_LUMINANCE_LATC1_EXT:
+   case GL_COMPRESSED_LUMINANCE_ALPHA_LATC2_EXT:
+   case GL_COMPRESSED_SIGNED_LUMINANCE_ALPHA_LATC2_EXT:
+      return ctx->Extensions.EXT_texture_compression_latc;
+   case GL_COMPRESSED_LUMINANCE_ALPHA_3DC_ATI:
+      return ctx->Extensions.ATI_texture_compression_3dc;
+   default:
+      return GL_FALSE;
+   }
+}
+
+
+/**
+ * Return the address of a specific pixel in an image (1D, 2D or 3D).
+ *
+ * Pixel unpacking/packing parameters are observed according to \p packing.
+ *
+ * \param dimensions either 1, 2 or 3 to indicate dimensionality of image
+ * \param image  starting address of image data
+ * \param width  the image width
+ * \param height  theimage height
+ * \param format  the pixel format
+ * \param type  the pixel data type
+ * \param packing  the pixelstore attributes
+ * \param img  which image in the volume (0 for 1D or 2D images)
+ * \param row  row of pixel in the image (0 for 1D images)
+ * \param column column of pixel in the image
+ * 
+ * \return address of pixel on success, or NULL on error.
+ *
+ * \sa gl_pixelstore_attrib.
+ */
+GLvoid *
+_mesa_image_address( GLuint dimensions,
+                     const struct gl_pixelstore_attrib *packing,
+                     const GLvoid *image,
+                     GLsizei width, GLsizei height,
+                     GLenum format, GLenum type,
+                     GLint img, GLint row, GLint column )
+{
+   GLint alignment;        /* 1, 2 or 4 */
+   GLint pixels_per_row;
+   GLint rows_per_image;
+   GLint skiprows;
+   GLint skippixels;
+   GLint skipimages;       /* for 3-D volume images */
+   GLubyte *pixel_addr;
+
+   ASSERT(dimensions >= 1 && dimensions <= 3);
+
+   alignment = packing->Alignment;
+   if (packing->RowLength > 0) {
+      pixels_per_row = packing->RowLength;
+   }
+   else {
+      pixels_per_row = width;
+   }
+   if (packing->ImageHeight > 0) {
+      rows_per_image = packing->ImageHeight;
+   }
+   else {
+      rows_per_image = height;
+   }
+
+   skippixels = packing->SkipPixels;
+   /* Note: SKIP_ROWS _is_ used for 1D images */
+   skiprows = packing->SkipRows;
+   /* Note: SKIP_IMAGES is only used for 3D images */
+   skipimages = (dimensions == 3) ? packing->SkipImages : 0;
+
+   if (type == GL_BITMAP) {
+      /* BITMAP data */
+      GLint comp_per_pixel;   /* components per pixel */
+      GLint bytes_per_comp;   /* bytes per component */
+      GLint bytes_per_row;
+      GLint bytes_per_image;
+
+      /* Compute bytes per component */
+      bytes_per_comp = _mesa_sizeof_packed_type( type );
+      if (bytes_per_comp < 0) {
+         return NULL;
+      }
+
+      /* Compute number of components per pixel */
+      comp_per_pixel = _mesa_components_in_format( format );
+      if (comp_per_pixel < 0) {
+         return NULL;
+      }
+
+      bytes_per_row = alignment
+                    * CEILING( comp_per_pixel*pixels_per_row, 8*alignment );
+
+      bytes_per_image = bytes_per_row * rows_per_image;
+
+      pixel_addr = (GLubyte *) image
+                 + (skipimages + img) * bytes_per_image
+                 + (skiprows + row) * bytes_per_row
+                 + (skippixels + column) / 8;
+   }
+   else {
+      /* Non-BITMAP data */
+      GLint bytes_per_pixel, bytes_per_row, remainder, bytes_per_image;
+      GLint topOfImage;
+
+      bytes_per_pixel = _mesa_bytes_per_pixel( format, type );
+
+      /* The pixel type and format should have been error checked earlier */
+      assert(bytes_per_pixel > 0);
+
+      bytes_per_row = pixels_per_row * bytes_per_pixel;
+      remainder = bytes_per_row % alignment;
+      if (remainder > 0)
+         bytes_per_row += (alignment - remainder);
+
+      ASSERT(bytes_per_row % alignment == 0);
+
+      bytes_per_image = bytes_per_row * rows_per_image;
+
+      if (packing->Invert) {
+         /* set pixel_addr to the last row */
+         topOfImage = bytes_per_row * (height - 1);
+         bytes_per_row = -bytes_per_row;
+      }
+      else {
+         topOfImage = 0;
+      }
+
+      /* compute final pixel address */
+      pixel_addr = (GLubyte *) image
+                 + (skipimages + img) * bytes_per_image
+                 + topOfImage
+                 + (skiprows + row) * bytes_per_row
+                 + (skippixels + column) * bytes_per_pixel;
+   }
+
+   return (GLvoid *) pixel_addr;
+}
+
+
+GLvoid *
+_mesa_image_address1d( const struct gl_pixelstore_attrib *packing,
+                       const GLvoid *image,
+                       GLsizei width,
+                       GLenum format, GLenum type,
+                       GLint column )
+{
+   return _mesa_image_address(1, packing, image, width, 1,
+                              format, type, 0, 0, column);
+}
+
+
+GLvoid *
+_mesa_image_address2d( const struct gl_pixelstore_attrib *packing,
+                       const GLvoid *image,
+                       GLsizei width, GLsizei height,
+                       GLenum format, GLenum type,
+                       GLint row, GLint column )
+{
+   return _mesa_image_address(2, packing, image, width, height,
+                              format, type, 0, row, column);
+}
+
+
+GLvoid *
+_mesa_image_address3d( const struct gl_pixelstore_attrib *packing,
+                       const GLvoid *image,
+                       GLsizei width, GLsizei height,
+                       GLenum format, GLenum type,
+                       GLint img, GLint row, GLint column )
+{
+   return _mesa_image_address(3, packing, image, width, height,
+                              format, type, img, row, column);
+}
+
+
+
+/**
+ * Compute the stride (in bytes) between image rows.
+ *
+ * \param packing the pixelstore attributes
+ * \param width image width.
+ * \param format pixel format.
+ * \param type pixel data type.
+ * 
+ * \return the stride in bytes for the given parameters, or -1 if error
+ */
+GLint
+_mesa_image_row_stride( const struct gl_pixelstore_attrib *packing,
+                        GLint width, GLenum format, GLenum type )
+{
+   GLint bytesPerRow, remainder;
+
+   ASSERT(packing);
+
+   if (type == GL_BITMAP) {
+      if (packing->RowLength == 0) {
+         bytesPerRow = (width + 7) / 8;
+      }
+      else {
+         bytesPerRow = (packing->RowLength + 7) / 8;
+      }
+   }
+   else {
+      /* Non-BITMAP data */
+      const GLint bytesPerPixel = _mesa_bytes_per_pixel(format, type);
+      if (bytesPerPixel <= 0)
+         return -1;  /* error */
+      if (packing->RowLength == 0) {
+         bytesPerRow = bytesPerPixel * width;
+      }
+      else {
+         bytesPerRow = bytesPerPixel * packing->RowLength;
+      }
+   }
+
+   remainder = bytesPerRow % packing->Alignment;
+   if (remainder > 0) {
+      bytesPerRow += (packing->Alignment - remainder);
+   }
+
+   if (packing->Invert) {
+      /* negate the bytes per row (negative row stride) */
+      bytesPerRow = -bytesPerRow;
+   }
+
+   return bytesPerRow;
+}
+
+
+/*
+ * Compute the stride between images in a 3D texture (in bytes) for the given
+ * pixel packing parameters and image width, format and type.
+ */
+GLint
+_mesa_image_image_stride( const struct gl_pixelstore_attrib *packing,
+                          GLint width, GLint height,
+                          GLenum format, GLenum type )
+{
+   GLint bytesPerRow, bytesPerImage, remainder;
+
+   ASSERT(packing);
+
+   if (type == GL_BITMAP) {
+      if (packing->RowLength == 0) {
+         bytesPerRow = (width + 7) / 8;
+      }
+      else {
+         bytesPerRow = (packing->RowLength + 7) / 8;
+      }
+   }
+   else {
+      const GLint bytesPerPixel = _mesa_bytes_per_pixel(format, type);
+
+      if (bytesPerPixel <= 0)
+         return -1;  /* error */
+      if (packing->RowLength == 0) {
+         bytesPerRow = bytesPerPixel * width;
+      }
+      else {
+         bytesPerRow = bytesPerPixel * packing->RowLength;
+      }
+   }
+
+   remainder = bytesPerRow % packing->Alignment;
+   if (remainder > 0)
+      bytesPerRow += (packing->Alignment - remainder);
+
+   if (packing->ImageHeight == 0)
+      bytesPerImage = bytesPerRow * height;
+   else
+      bytesPerImage = bytesPerRow * packing->ImageHeight;
+
+   return bytesPerImage;
+}
+
+
+
+/**
+ * "Expand" a bitmap from 1-bit per pixel to 8-bits per pixel.
+ * This is typically used to convert a bitmap into a GLubyte/pixel texture.
+ * "On" bits will set texels to \p onValue.
+ * "Off" bits will not modify texels.
+ * \param width  src bitmap width in pixels
+ * \param height  src bitmap height in pixels
+ * \param unpack  bitmap unpacking state
+ * \param bitmap  the src bitmap data
+ * \param destBuffer  start of dest buffer
+ * \param destStride  row stride in dest buffer
+ * \param onValue  if bit is 1, set destBuffer pixel to this value
+ */
+void
+_mesa_expand_bitmap(GLsizei width, GLsizei height,
+                    const struct gl_pixelstore_attrib *unpack,
+                    const GLubyte *bitmap,
+                    GLubyte *destBuffer, GLint destStride,
+                    GLubyte onValue)
+{
+   const GLubyte *srcRow = (const GLubyte *)
+      _mesa_image_address2d(unpack, bitmap, width, height,
+                            GL_COLOR_INDEX, GL_BITMAP, 0, 0);
+   const GLint srcStride = _mesa_image_row_stride(unpack, width,
+                                                  GL_COLOR_INDEX, GL_BITMAP);
+   GLint row, col;
+
+#define SET_PIXEL(COL, ROW) \
+   destBuffer[(ROW) * destStride + (COL)] = onValue;
+
+   for (row = 0; row < height; row++) {
+      const GLubyte *src = srcRow;
+
+      if (unpack->LsbFirst) {
+         /* Lsb first */
+         GLubyte mask = 1U << (unpack->SkipPixels & 0x7);
+         for (col = 0; col < width; col++) {
+
+            if (*src & mask) {
+               SET_PIXEL(col, row);
+            }
+
+            if (mask == 128U) {
+               src++;
+               mask = 1U;
+            }
+            else {
+               mask = mask << 1;
+            }
+         }
+
+         /* get ready for next row */
+         if (mask != 1)
+            src++;
+      }
+      else {
+         /* Msb first */
+         GLubyte mask = 128U >> (unpack->SkipPixels & 0x7);
+         for (col = 0; col < width; col++) {
+
+            if (*src & mask) {
+               SET_PIXEL(col, row);
+            }
+
+            if (mask == 1U) {
+               src++;
+               mask = 128U;
+            }
+            else {
+               mask = mask >> 1;
+            }
+         }
+
+         /* get ready for next row */
+         if (mask != 128)
+            src++;
+      }
+
+      srcRow += srcStride;
+   } /* row */
+
+#undef SET_PIXEL
+}
+
+
+
+
+/**
+ * Convert an array of RGBA colors from one datatype to another.
+ * NOTE: src may equal dst.  In that case, we use a temporary buffer.
+ */
+void
+_mesa_convert_colors(GLenum srcType, const GLvoid *src,
+                     GLenum dstType, GLvoid *dst,
+                     GLuint count, const GLubyte mask[])
+{
+   GLuint tempBuffer[MAX_WIDTH][4];
+   const GLboolean useTemp = (src == dst);
+
+   ASSERT(srcType != dstType);
+
+   switch (srcType) {
+   case GL_UNSIGNED_BYTE:
+      if (dstType == GL_UNSIGNED_SHORT) {
+         const GLubyte (*src1)[4] = (const GLubyte (*)[4]) src;
+         GLushort (*dst2)[4] = (GLushort (*)[4]) (useTemp ? tempBuffer : dst);
+         GLuint i;
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               dst2[i][RCOMP] = UBYTE_TO_USHORT(src1[i][RCOMP]);
+               dst2[i][GCOMP] = UBYTE_TO_USHORT(src1[i][GCOMP]);
+               dst2[i][BCOMP] = UBYTE_TO_USHORT(src1[i][BCOMP]);
+               dst2[i][ACOMP] = UBYTE_TO_USHORT(src1[i][ACOMP]);
+            }
+         }
+         if (useTemp)
+            memcpy(dst, tempBuffer, count * 4 * sizeof(GLushort));
+      }
+      else {
+         const GLubyte (*src1)[4] = (const GLubyte (*)[4]) src;
+         GLfloat (*dst4)[4] = (GLfloat (*)[4]) (useTemp ? tempBuffer : dst);
+         GLuint i;
+         ASSERT(dstType == GL_FLOAT);
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               dst4[i][RCOMP] = UBYTE_TO_FLOAT(src1[i][RCOMP]);
+               dst4[i][GCOMP] = UBYTE_TO_FLOAT(src1[i][GCOMP]);
+               dst4[i][BCOMP] = UBYTE_TO_FLOAT(src1[i][BCOMP]);
+               dst4[i][ACOMP] = UBYTE_TO_FLOAT(src1[i][ACOMP]);
+            }
+         }
+         if (useTemp)
+            memcpy(dst, tempBuffer, count * 4 * sizeof(GLfloat));
+      }
+      break;
+   case GL_UNSIGNED_SHORT:
+      if (dstType == GL_UNSIGNED_BYTE) {
+         const GLushort (*src2)[4] = (const GLushort (*)[4]) src;
+         GLubyte (*dst1)[4] = (GLubyte (*)[4]) (useTemp ? tempBuffer : dst);
+         GLuint i;
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               dst1[i][RCOMP] = USHORT_TO_UBYTE(src2[i][RCOMP]);
+               dst1[i][GCOMP] = USHORT_TO_UBYTE(src2[i][GCOMP]);
+               dst1[i][BCOMP] = USHORT_TO_UBYTE(src2[i][BCOMP]);
+               dst1[i][ACOMP] = USHORT_TO_UBYTE(src2[i][ACOMP]);
+            }
+         }
+         if (useTemp)
+            memcpy(dst, tempBuffer, count * 4 * sizeof(GLubyte));
+      }
+      else {
+         const GLushort (*src2)[4] = (const GLushort (*)[4]) src;
+         GLfloat (*dst4)[4] = (GLfloat (*)[4]) (useTemp ? tempBuffer : dst);
+         GLuint i;
+         ASSERT(dstType == GL_FLOAT);
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               dst4[i][RCOMP] = USHORT_TO_FLOAT(src2[i][RCOMP]);
+               dst4[i][GCOMP] = USHORT_TO_FLOAT(src2[i][GCOMP]);
+               dst4[i][BCOMP] = USHORT_TO_FLOAT(src2[i][BCOMP]);
+               dst4[i][ACOMP] = USHORT_TO_FLOAT(src2[i][ACOMP]);
+            }
+         }
+         if (useTemp)
+            memcpy(dst, tempBuffer, count * 4 * sizeof(GLfloat));
+      }
+      break;
+   case GL_FLOAT:
+      if (dstType == GL_UNSIGNED_BYTE) {
+         const GLfloat (*src4)[4] = (const GLfloat (*)[4]) src;
+         GLubyte (*dst1)[4] = (GLubyte (*)[4]) (useTemp ? tempBuffer : dst);
+         GLuint i;
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][RCOMP], src4[i][RCOMP]);
+               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][GCOMP], src4[i][GCOMP]);
+               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][BCOMP], src4[i][BCOMP]);
+               UNCLAMPED_FLOAT_TO_UBYTE(dst1[i][ACOMP], src4[i][ACOMP]);
+            }
+         }
+         if (useTemp)
+            memcpy(dst, tempBuffer, count * 4 * sizeof(GLubyte));
+      }
+      else {
+         const GLfloat (*src4)[4] = (const GLfloat (*)[4]) src;
+         GLushort (*dst2)[4] = (GLushort (*)[4]) (useTemp ? tempBuffer : dst);
+         GLuint i;
+         ASSERT(dstType == GL_UNSIGNED_SHORT);
+         for (i = 0; i < count; i++) {
+            if (!mask || mask[i]) {
+               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][RCOMP], src4[i][RCOMP]);
+               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][GCOMP], src4[i][GCOMP]);
+               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][BCOMP], src4[i][BCOMP]);
+               UNCLAMPED_FLOAT_TO_USHORT(dst2[i][ACOMP], src4[i][ACOMP]);
+            }
+         }
+         if (useTemp)
+            memcpy(dst, tempBuffer, count * 4 * sizeof(GLushort));
+      }
+      break;
+   default:
+      _mesa_problem(NULL, "Invalid datatype in _mesa_convert_colors");
+   }
+}
+
+
+
+
+/**
+ * Perform basic clipping for glDrawPixels.  The image's position and size
+ * and the unpack SkipPixels and SkipRows are adjusted so that the image
+ * region is entirely within the window and scissor bounds.
+ * NOTE: this will only work when glPixelZoom is (1, 1) or (1, -1).
+ * If Pixel.ZoomY is -1, *destY will be changed to be the first row which
+ * we'll actually write.  Beforehand, *destY-1 is the first drawing row.
+ *
+ * \return  GL_TRUE if image is ready for drawing or
+ *          GL_FALSE if image was completely clipped away (draw nothing)
+ */
+GLboolean
+_mesa_clip_drawpixels(const struct gl_context *ctx,
+                      GLint *destX, GLint *destY,
+                      GLsizei *width, GLsizei *height,
+                      struct gl_pixelstore_attrib *unpack)
+{
+   const struct gl_framebuffer *buffer = ctx->DrawBuffer;
+
+   if (unpack->RowLength == 0) {
+      unpack->RowLength = *width;
+   }
+
+   ASSERT(ctx->Pixel.ZoomX == 1.0F);
+   ASSERT(ctx->Pixel.ZoomY == 1.0F || ctx->Pixel.ZoomY == -1.0F);
+
+   /* left clipping */
+   if (*destX < buffer->_Xmin) {
+      unpack->SkipPixels += (buffer->_Xmin - *destX);
+      *width -= (buffer->_Xmin - *destX);
+      *destX = buffer->_Xmin;
+   }
+   /* right clipping */
+   if (*destX + *width > buffer->_Xmax)
+      *width -= (*destX + *width - buffer->_Xmax);
+
+   if (*width <= 0)
+      return GL_FALSE;
+
+   if (ctx->Pixel.ZoomY == 1.0F) {
+      /* bottom clipping */
+      if (*destY < buffer->_Ymin) {
+         unpack->SkipRows += (buffer->_Ymin - *destY);
+         *height -= (buffer->_Ymin - *destY);
+         *destY = buffer->_Ymin;
+      }
+      /* top clipping */
+      if (*destY + *height > buffer->_Ymax)
+         *height -= (*destY + *height - buffer->_Ymax);
+   }
+   else { /* upside down */
+      /* top clipping */
+      if (*destY > buffer->_Ymax) {
+         unpack->SkipRows += (*destY - buffer->_Ymax);
+         *height -= (*destY - buffer->_Ymax);
+         *destY = buffer->_Ymax;
+      }
+      /* bottom clipping */
+      if (*destY - *height < buffer->_Ymin)
+         *height -= (buffer->_Ymin - (*destY - *height));
+      /* adjust destY so it's the first row to write to */
+      (*destY)--;
+   }
+
+   if (*height <= 0)
+      return GL_FALSE;
+
+   return GL_TRUE;
+}
+
+
+/**
+ * Perform clipping for glReadPixels.  The image's window position
+ * and size, and the pack skipPixels, skipRows and rowLength are adjusted
+ * so that the image region is entirely within the window bounds.
+ * Note: this is different from _mesa_clip_drawpixels() in that the
+ * scissor box is ignored, and we use the bounds of the current readbuffer
+ * surface.
+ *
+ * \return  GL_TRUE if region to read is in bounds
+ *          GL_FALSE if region is completely out of bounds (nothing to read)
+ */
+GLboolean
+_mesa_clip_readpixels(const struct gl_context *ctx,
+                      GLint *srcX, GLint *srcY,
+                      GLsizei *width, GLsizei *height,
+                      struct gl_pixelstore_attrib *pack)
+{
+   const struct gl_framebuffer *buffer = ctx->ReadBuffer;
+
+   if (pack->RowLength == 0) {
+      pack->RowLength = *width;
+   }
+
+   /* left clipping */
+   if (*srcX < 0) {
+      pack->SkipPixels += (0 - *srcX);
+      *width -= (0 - *srcX);
+      *srcX = 0;
+   }
+   /* right clipping */
+   if (*srcX + *width > (GLsizei) buffer->Width)
+      *width -= (*srcX + *width - buffer->Width);
+
+   if (*width <= 0)
+      return GL_FALSE;
+
+   /* bottom clipping */
+   if (*srcY < 0) {
+      pack->SkipRows += (0 - *srcY);
+      *height -= (0 - *srcY);
+      *srcY = 0;
+   }
+   /* top clipping */
+   if (*srcY + *height > (GLsizei) buffer->Height)
+      *height -= (*srcY + *height - buffer->Height);
+
+   if (*height <= 0)
+      return GL_FALSE;
+
+   return GL_TRUE;
+}
+
+
+/**
+ * Do clipping for a glCopyTexSubImage call.
+ * The framebuffer source region might extend outside the framebuffer
+ * bounds.  Clip the source region against the framebuffer bounds and
+ * adjust the texture/dest position and size accordingly.
+ *
+ * \return GL_FALSE if region is totally clipped, GL_TRUE otherwise.
+ */
+GLboolean
+_mesa_clip_copytexsubimage(const struct gl_context *ctx,
+                           GLint *destX, GLint *destY,
+                           GLint *srcX, GLint *srcY,
+                           GLsizei *width, GLsizei *height)
+{
+   const struct gl_framebuffer *fb = ctx->ReadBuffer;
+   const GLint srcX0 = *srcX, srcY0 = *srcY;
+
+   if (_mesa_clip_to_region(0, 0, fb->Width, fb->Height,
+                            srcX, srcY, width, height)) {
+      *destX = *destX + *srcX - srcX0;
+      *destY = *destY + *srcY - srcY0;
+
+      return GL_TRUE;
+   }
+   else {
+      return GL_FALSE;
+   }
+}
+
+
+
+/**
+ * Clip the rectangle defined by (x, y, width, height) against the bounds
+ * specified by [xmin, xmax) and [ymin, ymax).
+ * \return GL_FALSE if rect is totally clipped, GL_TRUE otherwise.
+ */
+GLboolean
+_mesa_clip_to_region(GLint xmin, GLint ymin,
+                     GLint xmax, GLint ymax,
+                     GLint *x, GLint *y,
+                     GLsizei *width, GLsizei *height )
+{
+   /* left clipping */
+   if (*x < xmin) {
+      *width -= (xmin - *x);
+      *x = xmin;
+   }
+
+   /* right clipping */
+   if (*x + *width > xmax)
+      *width -= (*x + *width - xmax);
+
+   if (*width <= 0)
+      return GL_FALSE;
+
+   /* bottom (or top) clipping */
+   if (*y < ymin) {
+      *height -= (ymin - *y);
+      *y = ymin;
+   }
+
+   /* top (or bottom) clipping */
+   if (*y + *height > ymax)
+      *height -= (*y + *height - ymax);
+
+   if (*height <= 0)
+      return GL_FALSE;
+
+   return GL_TRUE;
+}
+
+
+/**
+ * Clip dst coords against Xmax (or Ymax).
+ */
+static INLINE void
+clip_right_or_top(GLint *srcX0, GLint *srcX1,
+                  GLint *dstX0, GLint *dstX1,
+                  GLint maxValue)
+{
+   GLfloat t, bias;
+
+   if (*dstX1 > maxValue) {
+      /* X1 outside right edge */
+      ASSERT(*dstX0 < maxValue); /* X0 should be inside right edge */
+      t = (GLfloat) (maxValue - *dstX0) / (GLfloat) (*dstX1 - *dstX0);
+      /* chop off [t, 1] part */
+      ASSERT(t >= 0.0 && t <= 1.0);
+      *dstX1 = maxValue;
+      bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F;
+      *srcX1 = *srcX0 + (GLint) (t * (*srcX1 - *srcX0) + bias);
+   }
+   else if (*dstX0 > maxValue) {
+      /* X0 outside right edge */
+      ASSERT(*dstX1 < maxValue); /* X1 should be inside right edge */
+      t = (GLfloat) (maxValue - *dstX1) / (GLfloat) (*dstX0 - *dstX1);
+      /* chop off [t, 1] part */
+      ASSERT(t >= 0.0 && t <= 1.0);
+      *dstX0 = maxValue;
+      bias = (*srcX0 < *srcX1) ? -0.5F : 0.5F;
+      *srcX0 = *srcX1 + (GLint) (t * (*srcX0 - *srcX1) + bias);
+   }
+}
+
+
+/**
+ * Clip dst coords against Xmin (or Ymin).
+ */
+static INLINE void
+clip_left_or_bottom(GLint *srcX0, GLint *srcX1,
+                    GLint *dstX0, GLint *dstX1,
+                    GLint minValue)
+{
+   GLfloat t, bias;
+
+   if (*dstX0 < minValue) {
+      /* X0 outside left edge */
+      ASSERT(*dstX1 > minValue); /* X1 should be inside left edge */
+      t = (GLfloat) (minValue - *dstX0) / (GLfloat) (*dstX1 - *dstX0);
+      /* chop off [0, t] part */
+      ASSERT(t >= 0.0 && t <= 1.0);
+      *dstX0 = minValue;
+      bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F; /* flipped??? */
+      *srcX0 = *srcX0 + (GLint) (t * (*srcX1 - *srcX0) + bias);
+   }
+   else if (*dstX1 < minValue) {
+      /* X1 outside left edge */
+      ASSERT(*dstX0 > minValue); /* X0 should be inside left edge */
+      t = (GLfloat) (minValue - *dstX1) / (GLfloat) (*dstX0 - *dstX1);
+      /* chop off [0, t] part */
+      ASSERT(t >= 0.0 && t <= 1.0);
+      *dstX1 = minValue;
+      bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F;
+      *srcX1 = *srcX1 + (GLint) (t * (*srcX0 - *srcX1) + bias);
+   }
+}
+
+
+/**
+ * Do clipping of blit src/dest rectangles.
+ * The dest rect is clipped against both the buffer bounds and scissor bounds.
+ * The src rect is just clipped against the buffer bounds.
+ *
+ * When either the src or dest rect is clipped, the other is also clipped
+ * proportionately!
+ *
+ * Note that X0 need not be less than X1 (same for Y) for either the source
+ * and dest rects.  That makes the clipping a little trickier.
+ *
+ * \return GL_TRUE if anything is left to draw, GL_FALSE if totally clipped
+ */
+GLboolean
+_mesa_clip_blit(struct gl_context *ctx,
+                GLint *srcX0, GLint *srcY0, GLint *srcX1, GLint *srcY1,
+                GLint *dstX0, GLint *dstY0, GLint *dstX1, GLint *dstY1)
+{
+   const GLint srcXmin = 0;
+   const GLint srcXmax = ctx->ReadBuffer->Width;
+   const GLint srcYmin = 0;
+   const GLint srcYmax = ctx->ReadBuffer->Height;
+
+   /* these include scissor bounds */
+   const GLint dstXmin = ctx->DrawBuffer->_Xmin;
+   const GLint dstXmax = ctx->DrawBuffer->_Xmax;
+   const GLint dstYmin = ctx->DrawBuffer->_Ymin;
+   const GLint dstYmax = ctx->DrawBuffer->_Ymax;
+
+   /*
+   printf("PreClipX:  src: %d .. %d  dst: %d .. %d\n",
+          *srcX0, *srcX1, *dstX0, *dstX1);
+   printf("PreClipY:  src: %d .. %d  dst: %d .. %d\n",
+          *srcY0, *srcY1, *dstY0, *dstY1);
+   */
+
+   /* trivial rejection tests */
+   if (*dstX0 == *dstX1)
+      return GL_FALSE; /* no width */
+   if (*dstX0 <= dstXmin && *dstX1 <= dstXmin)
+      return GL_FALSE; /* totally out (left) of bounds */
+   if (*dstX0 >= dstXmax && *dstX1 >= dstXmax)
+      return GL_FALSE; /* totally out (right) of bounds */
+
+   if (*dstY0 == *dstY1)
+      return GL_FALSE;
+   if (*dstY0 <= dstYmin && *dstY1 <= dstYmin)
+      return GL_FALSE;
+   if (*dstY0 >= dstYmax && *dstY1 >= dstYmax)
+      return GL_FALSE;
+
+   if (*srcX0 == *srcX1)
+      return GL_FALSE;
+   if (*srcX0 <= srcXmin && *srcX1 <= srcXmin)
+      return GL_FALSE;
+   if (*srcX0 >= srcXmax && *srcX1 >= srcXmax)
+      return GL_FALSE;
+
+   if (*srcY0 == *srcY1)
+      return GL_FALSE;
+   if (*srcY0 <= srcYmin && *srcY1 <= srcYmin)
+      return GL_FALSE;
+   if (*srcY0 >= srcYmax && *srcY1 >= srcYmax)
+      return GL_FALSE;
+
+   /*
+    * dest clip
+    */
+   clip_right_or_top(srcX0, srcX1, dstX0, dstX1, dstXmax);
+   clip_right_or_top(srcY0, srcY1, dstY0, dstY1, dstYmax);
+   clip_left_or_bottom(srcX0, srcX1, dstX0, dstX1, dstXmin);
+   clip_left_or_bottom(srcY0, srcY1, dstY0, dstY1, dstYmin);
+
+   /*
+    * src clip (just swap src/dst values from above)
+    */
+   clip_right_or_top(dstX0, dstX1, srcX0, srcX1, srcXmax);
+   clip_right_or_top(dstY0, dstY1, srcY0, srcY1, srcYmax);
+   clip_left_or_bottom(dstX0, dstX1, srcX0, srcX1, srcXmin);
+   clip_left_or_bottom(dstY0, dstY1, srcY0, srcY1, srcYmin);
+
+   /*
+   printf("PostClipX: src: %d .. %d  dst: %d .. %d\n",
+          *srcX0, *srcX1, *dstX0, *dstX1);
+   printf("PostClipY: src: %d .. %d  dst: %d .. %d\n",
+          *srcY0, *srcY1, *dstY0, *dstY1);
+   */
+
+   ASSERT(*dstX0 >= dstXmin);
+   ASSERT(*dstX0 <= dstXmax);
+   ASSERT(*dstX1 >= dstXmin);
+   ASSERT(*dstX1 <= dstXmax);
+
+   ASSERT(*dstY0 >= dstYmin);
+   ASSERT(*dstY0 <= dstYmax);
+   ASSERT(*dstY1 >= dstYmin);
+   ASSERT(*dstY1 <= dstYmax);
+
+   ASSERT(*srcX0 >= srcXmin);
+   ASSERT(*srcX0 <= srcXmax);
+   ASSERT(*srcX1 >= srcXmin);
+   ASSERT(*srcX1 <= srcXmax);
+
+   ASSERT(*srcY0 >= srcYmin);
+   ASSERT(*srcY0 <= srcYmax);
+   ASSERT(*srcY1 >= srcYmin);
+   ASSERT(*srcY1 <= srcYmax);
+
+   return GL_TRUE;
+}
diff --git a/mesalib/src/mesa/main/mipmap.c b/mesalib/src/mesa/main/mipmap.c
index f2724dbca..8a811cb72 100644
--- a/mesalib/src/mesa/main/mipmap.c
+++ b/mesalib/src/mesa/main/mipmap.c
@@ -706,6 +706,17 @@ do_row(GLenum datatype, GLuint comps, GLint srcWidth,
       }
    }
 
+   else if (datatype == GL_FLOAT_32_UNSIGNED_INT_24_8_REV && comps == 1) {
+      GLuint i, j, k;
+      const GLfloat *rowA = (const GLfloat *) srcRowA;
+      const GLfloat *rowB = (const GLfloat *) srcRowB;
+      GLfloat *dst = (GLfloat *) dstRow;
+      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
+           i++, j += colStride, k += colStride) {
+         dst[i*2] = (rowA[j*2] + rowA[k*2] + rowB[j*2] + rowB[k*2]) * 0.25F;
+      }
+   }
+
    else {
       _mesa_problem(NULL, "bad format in do_row()");
    }
@@ -1341,6 +1352,15 @@ do_row_3D(GLenum datatype, GLuint comps, GLint srcWidth,
       }
    }
 
+   else if (datatype == GL_FLOAT_32_UNSIGNED_INT_24_8_REV && comps == 1) {
+      DECLARE_ROW_POINTERS(GLfloat, 2);
+
+      for (i = j = 0, k = k0; i < (GLuint) dstWidth;
+           i++, j += colStride, k += colStride) {
+         FILTER_F_3D(0);
+      }
+   }
+
    else {
       _mesa_problem(NULL, "bad format in do_row()");
    }
diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h
index f018c75cc..b88118366 100644
--- a/mesalib/src/mesa/main/mtypes.h
+++ b/mesalib/src/mesa/main/mtypes.h
@@ -1856,7 +1856,7 @@ struct gl_program
    GLbitfield SystemValuesRead;   /**< Bitmask of SYSTEM_VALUE_x inputs used */
    GLbitfield InputFlags[MAX_PROGRAM_INPUTS];   /**< PROG_PARAM_BIT_x flags */
    GLbitfield OutputFlags[MAX_PROGRAM_OUTPUTS]; /**< PROG_PARAM_BIT_x flags */
-   GLbitfield TexturesUsed[MAX_TEXTURE_UNITS];  /**< TEXTURE_x_BIT bitmask */
+   GLbitfield TexturesUsed[MAX_COMBINED_TEXTURE_IMAGE_UNITS];  /**< TEXTURE_x_BIT bitmask */
    GLbitfield SamplersUsed;   /**< Bitfield of which samplers are used */
    GLbitfield ShadowSamplers; /**< Texture units used for shadow sampling. */
 
diff --git a/mesalib/src/mesa/main/pack.c b/mesalib/src/mesa/main/pack.c
index a232a51c3..7de1d05b9 100644
--- a/mesalib/src/mesa/main/pack.c
+++ b/mesalib/src/mesa/main/pack.c
@@ -1971,7 +1971,8 @@ extract_uint_indexes(GLuint n, GLuint indexes[],
           srcType == GL_INT ||
           srcType == GL_UNSIGNED_INT_24_8_EXT ||
           srcType == GL_HALF_FLOAT_ARB ||
-          srcType == GL_FLOAT);
+          srcType == GL_FLOAT ||
+          srcType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
    switch (srcType) {
       case GL_BITMAP:
@@ -2142,6 +2143,23 @@ extract_uint_indexes(GLuint n, GLuint indexes[],
             }
          }
          break;
+      case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+         {
+            GLuint i;
+            const GLuint *s = (const GLuint *) src;
+            if (unpack->SwapBytes) {
+               for (i = 0; i < n; i++) {
+                  GLuint value = s[i*2+1];
+                  SWAP4BYTE(value);
+                  indexes[i] = value & 0xff;  /* lower 8 bits */
+               }
+            }
+            else {
+               for (i = 0; i < n; i++)
+                  indexes[i] = s[i*2+1] & 0xff;  /* lower 8 bits */
+            }
+         }
+         break;
 
       default:
          _mesa_problem(NULL, "bad srcType in extract_uint_indexes");
@@ -4412,11 +4430,13 @@ _mesa_unpack_stencil_span( struct gl_context *ctx, GLuint n,
           srcType == GL_INT ||
           srcType == GL_UNSIGNED_INT_24_8_EXT ||
           srcType == GL_HALF_FLOAT_ARB ||
-          srcType == GL_FLOAT);
+          srcType == GL_FLOAT ||
+          srcType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
    ASSERT(dstType == GL_UNSIGNED_BYTE ||
           dstType == GL_UNSIGNED_SHORT ||
-          dstType == GL_UNSIGNED_INT);
+          dstType == GL_UNSIGNED_INT ||
+          dstType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
    /* only shift and offset apply to stencil */
    transferOps &= IMAGE_SHIFT_OFFSET_BIT;
@@ -4488,6 +4508,15 @@ _mesa_unpack_stencil_span( struct gl_context *ctx, GLuint n,
          case GL_UNSIGNED_INT:
             memcpy(dest, indexes, n * sizeof(GLuint));
             break;
+         case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+            {
+               GLuint *dst = (GLuint *) dest;
+               GLuint i;
+               for (i = 0; i < n; i++) {
+                  dst[i*2+1] = indexes[i] & 0xff; /* lower 8 bits */
+               }
+            }
+            break;
          default:
             _mesa_problem(ctx, "bad dstType in _mesa_unpack_stencil_span");
       }
@@ -4798,6 +4827,20 @@ _mesa_unpack_depth_span( struct gl_context *ctx, GLuint n,
             }
          }
          break;
+      case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+         {
+            GLuint i;
+            const GLfloat *src = (const GLfloat *)source;
+            for (i = 0; i < n; i++) {
+               GLfloat value = src[i * 2];
+               if (srcPacking->SwapBytes) {
+                  SWAP4BYTE(value);
+               }
+               depthValues[i] = value;
+            }
+            needClamp = GL_TRUE;
+         }
+         break;
       case GL_FLOAT:
          DEPTH_VALUES(GLfloat, 1*);
          needClamp = GL_TRUE;
@@ -4874,9 +4917,18 @@ _mesa_unpack_depth_span( struct gl_context *ctx, GLuint n,
          zValues[i] = (GLushort) (depthValues[i] * (GLfloat) depthMax);
       }
    }
+   else if (dstType == GL_FLOAT) {
+      /* Nothing to do. depthValues is pointing to dest. */
+   }
+   else if (dstType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV) {
+      GLfloat *zValues = (GLfloat*) dest;
+      GLuint i;
+      for (i = 0; i < n; i++) {
+         zValues[i*2] = depthValues[i];
+      }
+   }
    else {
-      ASSERT(dstType == GL_FLOAT);
-      /*ASSERT(depthMax == 1.0F);*/
+      ASSERT(0);
    }
 
    free(depthTemp);
@@ -5004,10 +5056,11 @@ _mesa_pack_depth_span( struct gl_context *ctx, GLuint n, GLvoid *dest,
 
 
 /**
- * Pack depth and stencil values as GL_DEPTH_STENCIL/GL_UNSIGNED_INT_24_8.
+ * Pack depth and stencil values as GL_DEPTH_STENCIL (GL_UNSIGNED_INT_24_8 etc)
  */
 void
-_mesa_pack_depth_stencil_span(struct gl_context *ctx, GLuint n, GLuint *dest,
+_mesa_pack_depth_stencil_span(struct gl_context *ctx,GLuint n,
+                              GLenum dstType, GLuint *dest,
                               const GLfloat *depthVals,
                               const GLstencil *stencilVals,
                               const struct gl_pixelstore_attrib *dstPacking)
@@ -5037,9 +5090,19 @@ _mesa_pack_depth_stencil_span(struct gl_context *ctx, GLuint n, GLuint *dest,
       stencilVals = stencilCopy;
    }
 
-   for (i = 0; i < n; i++) {
-      GLuint z = (GLuint) (depthVals[i] * 0xffffff);
-      dest[i] = (z << 8) | (stencilVals[i] & 0xff);
+   switch (dstType) {
+   case GL_UNSIGNED_INT_24_8:
+      for (i = 0; i < n; i++) {
+         GLuint z = (GLuint) (depthVals[i] * 0xffffff);
+         dest[i] = (z << 8) | (stencilVals[i] & 0xff);
+      }
+      break;
+   case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+      for (i = 0; i < n; i++) {
+         ((GLfloat*)dest)[i*2] = depthVals[i];
+         dest[i*2+1] = stencilVals[i] & 0xff;
+      }
+      break;
    }
 
    if (dstPacking->SwapBytes) {
diff --git a/mesalib/src/mesa/main/pack.h b/mesalib/src/mesa/main/pack.h
index 1f19a5a7c..00aab409e 100644
--- a/mesalib/src/mesa/main/pack.h
+++ b/mesalib/src/mesa/main/pack.h
@@ -1,147 +1,147 @@
-/*

- * Mesa 3-D graphics library

- *

- * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.

- * Copyright (C) 1999-2010  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, 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 BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN

- * AN ACTION 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 PACK_H

-#define PACK_H

-

-

-#include "mtypes.h"

-

-

-extern void

-_mesa_unpack_polygon_stipple(const GLubyte *pattern, GLuint dest[32],

-                             const struct gl_pixelstore_attrib *unpacking);

-

-

-extern void

-_mesa_pack_polygon_stipple(const GLuint pattern[32], GLubyte *dest,

-                           const struct gl_pixelstore_attrib *packing);

-

-

-extern GLvoid *

-_mesa_unpack_bitmap(GLint width, GLint height, const GLubyte *pixels,

-                    const struct gl_pixelstore_attrib *packing);

-

-extern void

-_mesa_pack_bitmap(GLint width, GLint height, const GLubyte *source,

-                  GLubyte *dest, const struct gl_pixelstore_attrib *packing);

-

-

-extern void

-_mesa_pack_rgba_span_float(struct gl_context *ctx, GLuint n,

-                           GLfloat rgba[][4],

-                           GLenum dstFormat, GLenum dstType, GLvoid *dstAddr,

-                           const struct gl_pixelstore_attrib *dstPacking,

-                           GLbitfield transferOps);

-

-

-extern void

-_mesa_unpack_color_span_chan(struct gl_context *ctx,

-                             GLuint n, GLenum dstFormat, GLchan dest[],

-                             GLenum srcFormat, GLenum srcType,

-                             const GLvoid *source,

-                             const struct gl_pixelstore_attrib *srcPacking,

-                             GLbitfield transferOps);

-

-

-extern void

-_mesa_unpack_color_span_float(struct gl_context *ctx,

-                              GLuint n, GLenum dstFormat, GLfloat dest[],

-                              GLenum srcFormat, GLenum srcType,

-                              const GLvoid *source,

-                              const struct gl_pixelstore_attrib *srcPacking,

-                              GLbitfield transferOps);

-

-extern void

-_mesa_unpack_color_span_uint(struct gl_context *ctx,

-                             GLuint n, GLenum dstFormat, GLuint *dest,

-                             GLenum srcFormat, GLenum srcType,

-                             const GLvoid *source,

-                             const struct gl_pixelstore_attrib *srcPacking);

-

-extern void

-_mesa_unpack_dudv_span_byte(struct gl_context *ctx,

-                            GLuint n, GLenum dstFormat, GLbyte dest[],

-                            GLenum srcFormat, GLenum srcType,

-                            const GLvoid *source,

-                            const struct gl_pixelstore_attrib *srcPacking,

-                            GLbitfield transferOps);

-

-extern void

-_mesa_unpack_index_span(struct gl_context *ctx, GLuint n,

-                        GLenum dstType, GLvoid *dest,

-                        GLenum srcType, const GLvoid *source,

-                        const struct gl_pixelstore_attrib *srcPacking,

-                        GLbitfield transferOps);

-

-

-extern void

-_mesa_pack_index_span(struct gl_context *ctx, GLuint n,

-                      GLenum dstType, GLvoid *dest, const GLuint *source,

-                      const struct gl_pixelstore_attrib *dstPacking,

-                      GLbitfield transferOps);

-

-

-extern void

-_mesa_unpack_stencil_span(struct gl_context *ctx, GLuint n,

-                          GLenum dstType, GLvoid *dest,

-                          GLenum srcType, const GLvoid *source,

-                          const struct gl_pixelstore_attrib *srcPacking,

-                          GLbitfield transferOps);

-

-extern void

-_mesa_pack_stencil_span(struct gl_context *ctx, GLuint n,

-                        GLenum dstType, GLvoid *dest, const GLstencil *source,

-                        const struct gl_pixelstore_attrib *dstPacking);

-

-

-extern void

-_mesa_unpack_depth_span(struct gl_context *ctx, GLuint n,

-                        GLenum dstType, GLvoid *dest, GLuint depthMax,

-                        GLenum srcType, const GLvoid *source,

-                        const struct gl_pixelstore_attrib *srcPacking);

-

-extern void

-_mesa_pack_depth_span(struct gl_context *ctx, GLuint n, GLvoid *dest,

-                      GLenum dstType, const GLfloat *depthSpan,

-                      const struct gl_pixelstore_attrib *dstPacking);

-

-

-extern void

-_mesa_pack_depth_stencil_span(struct gl_context *ctx,

-                              GLuint n, GLuint *dest,

-                              const GLfloat *depthVals,

-                              const GLstencil *stencilVals,

-                              const struct gl_pixelstore_attrib *dstPacking);

-

-

-extern void *

-_mesa_unpack_image(GLuint dimensions,

-                   GLsizei width, GLsizei height, GLsizei depth,

-                   GLenum format, GLenum type, const GLvoid *pixels,

-                   const struct gl_pixelstore_attrib *unpack);

-

-

-#endif

+/*
+ * Mesa 3-D graphics library
+ *
+ * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.
+ * Copyright (C) 1999-2010  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, 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 BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
+ * AN ACTION 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 PACK_H
+#define PACK_H
+
+
+#include "mtypes.h"
+
+
+extern void
+_mesa_unpack_polygon_stipple(const GLubyte *pattern, GLuint dest[32],
+                             const struct gl_pixelstore_attrib *unpacking);
+
+
+extern void
+_mesa_pack_polygon_stipple(const GLuint pattern[32], GLubyte *dest,
+                           const struct gl_pixelstore_attrib *packing);
+
+
+extern GLvoid *
+_mesa_unpack_bitmap(GLint width, GLint height, const GLubyte *pixels,
+                    const struct gl_pixelstore_attrib *packing);
+
+extern void
+_mesa_pack_bitmap(GLint width, GLint height, const GLubyte *source,
+                  GLubyte *dest, const struct gl_pixelstore_attrib *packing);
+
+
+extern void
+_mesa_pack_rgba_span_float(struct gl_context *ctx, GLuint n,
+                           GLfloat rgba[][4],
+                           GLenum dstFormat, GLenum dstType, GLvoid *dstAddr,
+                           const struct gl_pixelstore_attrib *dstPacking,
+                           GLbitfield transferOps);
+
+
+extern void
+_mesa_unpack_color_span_chan(struct gl_context *ctx,
+                             GLuint n, GLenum dstFormat, GLchan dest[],
+                             GLenum srcFormat, GLenum srcType,
+                             const GLvoid *source,
+                             const struct gl_pixelstore_attrib *srcPacking,
+                             GLbitfield transferOps);
+
+
+extern void
+_mesa_unpack_color_span_float(struct gl_context *ctx,
+                              GLuint n, GLenum dstFormat, GLfloat dest[],
+                              GLenum srcFormat, GLenum srcType,
+                              const GLvoid *source,
+                              const struct gl_pixelstore_attrib *srcPacking,
+                              GLbitfield transferOps);
+
+extern void
+_mesa_unpack_color_span_uint(struct gl_context *ctx,
+                             GLuint n, GLenum dstFormat, GLuint *dest,
+                             GLenum srcFormat, GLenum srcType,
+                             const GLvoid *source,
+                             const struct gl_pixelstore_attrib *srcPacking);
+
+extern void
+_mesa_unpack_dudv_span_byte(struct gl_context *ctx,
+                            GLuint n, GLenum dstFormat, GLbyte dest[],
+                            GLenum srcFormat, GLenum srcType,
+                            const GLvoid *source,
+                            const struct gl_pixelstore_attrib *srcPacking,
+                            GLbitfield transferOps);
+
+extern void
+_mesa_unpack_index_span(struct gl_context *ctx, GLuint n,
+                        GLenum dstType, GLvoid *dest,
+                        GLenum srcType, const GLvoid *source,
+                        const struct gl_pixelstore_attrib *srcPacking,
+                        GLbitfield transferOps);
+
+
+extern void
+_mesa_pack_index_span(struct gl_context *ctx, GLuint n,
+                      GLenum dstType, GLvoid *dest, const GLuint *source,
+                      const struct gl_pixelstore_attrib *dstPacking,
+                      GLbitfield transferOps);
+
+
+extern void
+_mesa_unpack_stencil_span(struct gl_context *ctx, GLuint n,
+                          GLenum dstType, GLvoid *dest,
+                          GLenum srcType, const GLvoid *source,
+                          const struct gl_pixelstore_attrib *srcPacking,
+                          GLbitfield transferOps);
+
+extern void
+_mesa_pack_stencil_span(struct gl_context *ctx, GLuint n,
+                        GLenum dstType, GLvoid *dest, const GLstencil *source,
+                        const struct gl_pixelstore_attrib *dstPacking);
+
+
+extern void
+_mesa_unpack_depth_span(struct gl_context *ctx, GLuint n,
+                        GLenum dstType, GLvoid *dest, GLuint depthMax,
+                        GLenum srcType, const GLvoid *source,
+                        const struct gl_pixelstore_attrib *srcPacking);
+
+extern void
+_mesa_pack_depth_span(struct gl_context *ctx, GLuint n, GLvoid *dest,
+                      GLenum dstType, const GLfloat *depthSpan,
+                      const struct gl_pixelstore_attrib *dstPacking);
+
+
+extern void
+_mesa_pack_depth_stencil_span(struct gl_context *ctx,GLuint n,
+                              GLenum dstType, GLuint *dest,
+                              const GLfloat *depthVals,
+                              const GLstencil *stencilVals,
+                              const struct gl_pixelstore_attrib *dstPacking);
+
+
+extern void *
+_mesa_unpack_image(GLuint dimensions,
+                   GLsizei width, GLsizei height, GLsizei depth,
+                   GLenum format, GLenum type, const GLvoid *pixels,
+                   const struct gl_pixelstore_attrib *unpack);
+
+
+#endif
diff --git a/mesalib/src/mesa/main/readpix.c b/mesalib/src/mesa/main/readpix.c
index 0331a8ca2..84c5b2228 100644
--- a/mesalib/src/mesa/main/readpix.c
+++ b/mesalib/src/mesa/main/readpix.c
@@ -61,6 +61,14 @@ _mesa_error_check_format_type(struct gl_context *ctx, GLenum format,
       return GL_TRUE;
    }
 
+   if (ctx->Extensions.ARB_depth_buffer_float
+       && type == GL_FLOAT_32_UNSIGNED_INT_24_8_REV
+       && format != GL_DEPTH_STENCIL_EXT) {
+      _mesa_error(ctx, GL_INVALID_OPERATION,
+                  "gl%sPixels(format is not GL_DEPTH_STENCIL_EXT)", readDraw);
+      return GL_TRUE;
+   }
+
    /* basic combinations test */
    if (!_mesa_is_legal_format_and_type(ctx, format, type)) {
       _mesa_error(ctx, GL_INVALID_ENUM,
@@ -142,10 +150,23 @@ _mesa_error_check_format_type(struct gl_context *ctx, GLenum format,
       }
       break;
    case GL_DEPTH_STENCIL_EXT:
-      if (!ctx->Extensions.EXT_packed_depth_stencil ||
-          type != GL_UNSIGNED_INT_24_8_EXT) {
-         _mesa_error(ctx, GL_INVALID_ENUM, "gl%sPixels(type)", readDraw);
-         return GL_TRUE;
+      /* Check validity of the type first. */
+      switch (type) {
+         case GL_UNSIGNED_INT_24_8_EXT:
+            if (!ctx->Extensions.EXT_packed_depth_stencil) {
+               _mesa_error(ctx, GL_INVALID_ENUM, "gl%sPixels(type)", readDraw);
+               return GL_TRUE;
+            }
+            break;
+         case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+            if (!ctx->Extensions.ARB_depth_buffer_float) {
+               _mesa_error(ctx, GL_INVALID_ENUM, "gl%sPixels(type)", readDraw);
+               return GL_TRUE;
+            }
+            break;
+         default:
+            _mesa_error(ctx, GL_INVALID_ENUM, "gl%sPixels(type)", readDraw);
+            return GL_TRUE;
       }
       if ((drawing && !_mesa_dest_buffer_exists(ctx, format)) ||
           (reading && !_mesa_source_buffer_exists(ctx, format))) {
diff --git a/mesalib/src/mesa/main/renderbuffer.c b/mesalib/src/mesa/main/renderbuffer.c
index c36175c60..f5b20020d 100644
--- a/mesalib/src/mesa/main/renderbuffer.c
+++ b/mesalib/src/mesa/main/renderbuffer.c
@@ -66,6 +66,9 @@ get_datatype_bytes(struct gl_renderbuffer *rb)
    int component_size;
 
    switch (rb->DataType) {
+   case GL_FLOAT_32_UNSIGNED_INT_24_8_REV:
+      component_size = 8;
+      break;
    case GL_FLOAT:
    case GL_UNSIGNED_INT:
    case GL_UNSIGNED_INT_24_8_EXT:
diff --git a/mesalib/src/mesa/main/samplerobj.c b/mesalib/src/mesa/main/samplerobj.c
index f7774fdd7..8f8d87b90 100644
--- a/mesalib/src/mesa/main/samplerobj.c
+++ b/mesalib/src/mesa/main/samplerobj.c
@@ -251,7 +251,7 @@ _mesa_BindSampler(GLuint unit, GLuint sampler)
    struct gl_sampler_object *sampObj;
    GET_CURRENT_CONTEXT(ctx);
 
-   if (unit >= ctx->Const.MaxTextureImageUnits) {
+   if (unit >= ctx->Const.MaxCombinedTextureImageUnits) {
       _mesa_error(ctx, GL_INVALID_VALUE, "glBindSampler(unit %u)", unit);
       return;
    }
diff --git a/mesalib/src/mesa/main/shaderapi.c b/mesalib/src/mesa/main/shaderapi.c
index b58e30de9..8df25c3f9 100644
--- a/mesalib/src/mesa/main/shaderapi.c
+++ b/mesalib/src/mesa/main/shaderapi.c
@@ -1032,7 +1032,7 @@ validate_samplers(const struct gl_program *prog, char *errMsg)
       "TEXTURE_2D",
       "TEXTURE_1D",
    };
-   GLint targetUsed[MAX_TEXTURE_IMAGE_UNITS];
+   GLint targetUsed[MAX_COMBINED_TEXTURE_IMAGE_UNITS];
    GLbitfield samplersUsed = prog->SamplersUsed;
    GLuint i;
 
@@ -1050,7 +1050,7 @@ validate_samplers(const struct gl_program *prog, char *errMsg)
       gl_texture_index target;
       GLint sampler = _mesa_ffs(samplersUsed) - 1;
       assert(sampler >= 0);
-      assert(sampler < MAX_TEXTURE_IMAGE_UNITS);
+      assert(sampler < Elements(prog->SamplerUnits));
       unit = prog->SamplerUnits[sampler];
       target = prog->SamplerTargets[sampler];
       if (targetUsed[unit] != -1 && targetUsed[unit] != (int) target) {
diff --git a/mesalib/src/mesa/main/texenv.c b/mesalib/src/mesa/main/texenv.c
index 9228e354a..c0d0f3779 100644
--- a/mesalib/src/mesa/main/texenv.c
+++ b/mesalib/src/mesa/main/texenv.c
@@ -419,7 +419,7 @@ _mesa_TexEnvfv( GLenum target, GLenum pname, const GLfloat *param )
    ASSERT_OUTSIDE_BEGIN_END(ctx);
 
    maxUnit = (target == GL_POINT_SPRITE_NV && pname == GL_COORD_REPLACE_NV)
-      ? ctx->Const.MaxTextureCoordUnits : ctx->Const.MaxTextureImageUnits;
+      ? ctx->Const.MaxTextureCoordUnits : ctx->Const.MaxCombinedTextureImageUnits;
    if (ctx->Texture.CurrentUnit >= maxUnit) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "glTexEnvfv(current unit)");
       return;
@@ -748,7 +748,7 @@ _mesa_GetTexEnvfv( GLenum target, GLenum pname, GLfloat *params )
    ASSERT_OUTSIDE_BEGIN_END(ctx);
 
    maxUnit = (target == GL_POINT_SPRITE_NV && pname == GL_COORD_REPLACE_NV)
-      ? ctx->Const.MaxTextureCoordUnits : ctx->Const.MaxTextureImageUnits;
+      ? ctx->Const.MaxTextureCoordUnits : ctx->Const.MaxCombinedTextureImageUnits;
    if (ctx->Texture.CurrentUnit >= maxUnit) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexEnvfv(current unit)");
       return;
@@ -817,7 +817,7 @@ _mesa_GetTexEnviv( GLenum target, GLenum pname, GLint *params )
    ASSERT_OUTSIDE_BEGIN_END(ctx);
 
    maxUnit = (target == GL_POINT_SPRITE_NV && pname == GL_COORD_REPLACE_NV)
-      ? ctx->Const.MaxTextureCoordUnits : ctx->Const.MaxTextureImageUnits;
+      ? ctx->Const.MaxTextureCoordUnits : ctx->Const.MaxCombinedTextureImageUnits;
    if (ctx->Texture.CurrentUnit >= maxUnit) {
       _mesa_error(ctx, GL_INVALID_OPERATION, "glGetTexEnviv(current unit)");
       return;
diff --git a/mesalib/src/mesa/main/texfetch.c b/mesalib/src/mesa/main/texfetch.c
index ab3abf657..72283eb68 100644
--- a/mesalib/src/mesa/main/texfetch.c
+++ b/mesalib/src/mesa/main/texfetch.c
@@ -1,1050 +1,1064 @@
-/*

- * Mesa 3-D graphics library

- * Version:  7.7

- *

- * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.

- * Copyright (c) 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.

- */

-

-

-/**

- * \file texfetch.c

- *

- * Texel fetch/store functions

- *

- * \author Gareth Hughes

- */

-

-

-#include "colormac.h"

-#include "macros.h"

-#include "texcompress.h"

-#include "texcompress_fxt1.h"

-#include "texcompress_s3tc.h"

-#include "texcompress_rgtc.h"

-#include "texfetch.h"

-#include "teximage.h"

-#include "../../gallium/auxiliary/util/u_format_rgb9e5.h"

-#include "../../gallium/auxiliary/util/u_format_r11g11b10f.h"

-

-

-/**

- * Convert an 8-bit sRGB value from non-linear space to a

- * linear RGB value in [0, 1].

- * Implemented with a 256-entry lookup table.

- */

-static INLINE GLfloat

-nonlinear_to_linear(GLubyte cs8)

-{

-   static GLfloat table[256];

-   static GLboolean tableReady = GL_FALSE;

-   if (!tableReady) {

-      /* compute lookup table now */

-      GLuint i;

-      for (i = 0; i < 256; i++) {

-         const GLfloat cs = UBYTE_TO_FLOAT(i);

-         if (cs <= 0.04045) {

-            table[i] = cs / 12.92f;

-         }

-         else {

-            table[i] = (GLfloat) pow((cs + 0.055) / 1.055, 2.4);

-         }

-      }

-      tableReady = GL_TRUE;

-   }

-   return table[cs8];

-}

-

-

-

-/* Texel fetch routines for all supported formats

- */

-#define DIM 1

-#include "texfetch_tmp.h"

-

-#define DIM 2

-#include "texfetch_tmp.h"

-

-#define DIM 3

-#include "texfetch_tmp.h"

-

-/**

- * Null texel fetch function.

- *

- * Have to have this so the FetchTexel function pointer is never NULL.

- */

-static void fetch_null_texelf( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   (void) texImage; (void) i; (void) j; (void) k;

-   texel[RCOMP] = 0.0;

-   texel[GCOMP] = 0.0;

-   texel[BCOMP] = 0.0;

-   texel[ACOMP] = 0.0;

-   _mesa_warning(NULL, "fetch_null_texelf() called!");

-}

-

-static void store_null_texel(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   (void) texImage;

-   (void) i;

-   (void) j;

-   (void) k;

-   (void) texel;

-   /* no-op */

-}

-

-

-

-/**

- * Table to map MESA_FORMAT_ to texel fetch/store funcs.

- * XXX this is somewhat temporary.

- */

-static struct {

-   gl_format Name;

-   FetchTexelFuncF Fetch1D;

-   FetchTexelFuncF Fetch2D;

-   FetchTexelFuncF Fetch3D;

-   StoreTexelFunc StoreTexel;

-}

-texfetch_funcs[MESA_FORMAT_COUNT] =

-{

-   {

-      MESA_FORMAT_NONE,

-      fetch_null_texelf,

-      fetch_null_texelf,

-      fetch_null_texelf,

-      store_null_texel

-   },

-

-   {

-      MESA_FORMAT_RGBA8888,

-      fetch_texel_1d_f_rgba8888,

-      fetch_texel_2d_f_rgba8888,

-      fetch_texel_3d_f_rgba8888,

-      store_texel_rgba8888

-   },

-   {

-      MESA_FORMAT_RGBA8888_REV,

-      fetch_texel_1d_f_rgba8888_rev,

-      fetch_texel_2d_f_rgba8888_rev,

-      fetch_texel_3d_f_rgba8888_rev,

-      store_texel_rgba8888_rev

-   },

-   {

-      MESA_FORMAT_ARGB8888,

-      fetch_texel_1d_f_argb8888,

-      fetch_texel_2d_f_argb8888,

-      fetch_texel_3d_f_argb8888,

-      store_texel_argb8888

-   },

-   {

-      MESA_FORMAT_ARGB8888_REV,

-      fetch_texel_1d_f_argb8888_rev,

-      fetch_texel_2d_f_argb8888_rev,

-      fetch_texel_3d_f_argb8888_rev,

-      store_texel_argb8888_rev

-   },

-   {

-      MESA_FORMAT_XRGB8888,

-      fetch_texel_1d_f_xrgb8888,

-      fetch_texel_2d_f_xrgb8888,

-      fetch_texel_3d_f_xrgb8888,

-      store_texel_xrgb8888

-   },

-   {

-      MESA_FORMAT_XRGB8888_REV,

-      fetch_texel_1d_f_xrgb8888_rev,

-      fetch_texel_2d_f_xrgb8888_rev,

-      fetch_texel_3d_f_xrgb8888_rev,

-      store_texel_xrgb8888_rev,

-   },

-   {

-      MESA_FORMAT_RGB888,

-      fetch_texel_1d_f_rgb888,

-      fetch_texel_2d_f_rgb888,

-      fetch_texel_3d_f_rgb888,

-      store_texel_rgb888

-   },

-   {

-      MESA_FORMAT_BGR888,

-      fetch_texel_1d_f_bgr888,

-      fetch_texel_2d_f_bgr888,

-      fetch_texel_3d_f_bgr888,

-      store_texel_bgr888

-   },

-   {

-      MESA_FORMAT_RGB565,

-      fetch_texel_1d_f_rgb565,

-      fetch_texel_2d_f_rgb565,

-      fetch_texel_3d_f_rgb565,

-      store_texel_rgb565

-   },

-   {

-      MESA_FORMAT_RGB565_REV,

-      fetch_texel_1d_f_rgb565_rev,

-      fetch_texel_2d_f_rgb565_rev,

-      fetch_texel_3d_f_rgb565_rev,

-      store_texel_rgb565_rev

-   },

-   {

-      MESA_FORMAT_ARGB4444,

-      fetch_texel_1d_f_argb4444,

-      fetch_texel_2d_f_argb4444,

-      fetch_texel_3d_f_argb4444,

-      store_texel_argb4444

-   },

-   {

-      MESA_FORMAT_ARGB4444_REV,

-      fetch_texel_1d_f_argb4444_rev,

-      fetch_texel_2d_f_argb4444_rev,

-      fetch_texel_3d_f_argb4444_rev,

-      store_texel_argb4444_rev

-   },

-   {

-      MESA_FORMAT_RGBA5551,

-      fetch_texel_1d_f_rgba5551,

-      fetch_texel_2d_f_rgba5551,

-      fetch_texel_3d_f_rgba5551,

-      store_texel_rgba5551

-   },

-   {

-      MESA_FORMAT_ARGB1555,

-      fetch_texel_1d_f_argb1555,

-      fetch_texel_2d_f_argb1555,

-      fetch_texel_3d_f_argb1555,

-      store_texel_argb1555

-   },

-   {

-      MESA_FORMAT_ARGB1555_REV,

-      fetch_texel_1d_f_argb1555_rev,

-      fetch_texel_2d_f_argb1555_rev,

-      fetch_texel_3d_f_argb1555_rev,

-      store_texel_argb1555_rev

-   },

-   {

-      MESA_FORMAT_AL44,

-      fetch_texel_1d_f_al44,

-      fetch_texel_2d_f_al44,

-      fetch_texel_3d_f_al44,

-      store_texel_al44

-   },

-   {

-      MESA_FORMAT_AL88,

-      fetch_texel_1d_f_al88,

-      fetch_texel_2d_f_al88,

-      fetch_texel_3d_f_al88,

-      store_texel_al88

-   },

-   {

-      MESA_FORMAT_AL88_REV,

-      fetch_texel_1d_f_al88_rev,

-      fetch_texel_2d_f_al88_rev,

-      fetch_texel_3d_f_al88_rev,

-      store_texel_al88_rev

-   },

-   {

-      MESA_FORMAT_AL1616,

-      fetch_texel_1d_f_al1616,

-      fetch_texel_2d_f_al1616,

-      fetch_texel_3d_f_al1616,

-      store_texel_al1616

-   },

-   {

-      MESA_FORMAT_AL1616_REV,

-      fetch_texel_1d_f_al1616_rev,

-      fetch_texel_2d_f_al1616_rev,

-      fetch_texel_3d_f_al1616_rev,

-      store_texel_al1616_rev

-   },

-   {

-      MESA_FORMAT_RGB332,

-      fetch_texel_1d_f_rgb332,

-      fetch_texel_2d_f_rgb332,

-      fetch_texel_3d_f_rgb332,

-      store_texel_rgb332

-   },

-   {

-      MESA_FORMAT_A8,

-      fetch_texel_1d_f_a8,

-      fetch_texel_2d_f_a8,

-      fetch_texel_3d_f_a8,

-      store_texel_a8

-   },

-   {

-      MESA_FORMAT_A16,

-      fetch_texel_1d_f_a16,

-      fetch_texel_2d_f_a16,

-      fetch_texel_3d_f_a16,

-      store_texel_a16

-   },

-   {

-      MESA_FORMAT_L8,

-      fetch_texel_1d_f_l8,

-      fetch_texel_2d_f_l8,

-      fetch_texel_3d_f_l8,

-      store_texel_l8

-   },

-   {

-      MESA_FORMAT_L16,

-      fetch_texel_1d_f_l16,

-      fetch_texel_2d_f_l16,

-      fetch_texel_3d_f_l16,

-      store_texel_l16

-   },

-   {

-      MESA_FORMAT_I8,

-      fetch_texel_1d_f_i8,

-      fetch_texel_2d_f_i8,

-      fetch_texel_3d_f_i8,

-      store_texel_i8

-   },

-   {

-      MESA_FORMAT_I16,

-      fetch_texel_1d_f_i16,

-      fetch_texel_2d_f_i16,

-      fetch_texel_3d_f_i16,

-      store_texel_i16

-   },

-   {

-      MESA_FORMAT_CI8,

-      fetch_texel_1d_f_ci8,

-      fetch_texel_2d_f_ci8,

-      fetch_texel_3d_f_ci8,

-      store_texel_ci8

-   },

-   {

-      MESA_FORMAT_YCBCR,

-      fetch_texel_1d_f_ycbcr,

-      fetch_texel_2d_f_ycbcr,

-      fetch_texel_3d_f_ycbcr,

-      store_texel_ycbcr

-   },

-   {

-      MESA_FORMAT_YCBCR_REV,

-      fetch_texel_1d_f_ycbcr_rev,

-      fetch_texel_2d_f_ycbcr_rev,

-      fetch_texel_3d_f_ycbcr_rev,

-      store_texel_ycbcr_rev

-   },

-   {

-      MESA_FORMAT_R8,

-      fetch_texel_1d_f_r8,

-      fetch_texel_2d_f_r8,

-      fetch_texel_3d_f_r8,

-      store_texel_r8,

-   },

-   {

-      MESA_FORMAT_RG88,

-      fetch_texel_1d_f_rg88,

-      fetch_texel_2d_f_rg88,

-      fetch_texel_3d_f_rg88,

-      store_texel_rg88,

-   },

-   {

-      MESA_FORMAT_RG88_REV,

-      fetch_texel_1d_f_rg88_rev,

-      fetch_texel_2d_f_rg88_rev,

-      fetch_texel_3d_f_rg88_rev,

-      store_texel_rg88_rev,

-   },

-   {

-      MESA_FORMAT_R16,

-      fetch_texel_1d_f_r16,

-      fetch_texel_2d_f_r16,

-      fetch_texel_3d_f_r16,

-      store_texel_r16,

-   },

-   {

-      MESA_FORMAT_RG1616,

-      fetch_texel_1d_f_rg1616,

-      fetch_texel_2d_f_rg1616,

-      fetch_texel_3d_f_rg1616,

-      store_texel_rg1616,

-   },

-   {

-      MESA_FORMAT_RG1616_REV,

-      fetch_texel_1d_f_rg1616_rev,

-      fetch_texel_2d_f_rg1616_rev,

-      fetch_texel_3d_f_rg1616_rev,

-      store_texel_rg1616_rev,

-   },

-   {

-      MESA_FORMAT_ARGB2101010,

-      fetch_texel_1d_f_argb2101010,

-      fetch_texel_2d_f_argb2101010,

-      fetch_texel_3d_f_argb2101010,

-      store_texel_argb2101010

-   },

-   {

-      MESA_FORMAT_Z24_S8,

-      fetch_texel_1d_f_z24_s8,

-      fetch_texel_2d_f_z24_s8,

-      fetch_texel_3d_f_z24_s8,

-      store_texel_z24_s8

-   },

-   {

-      MESA_FORMAT_S8_Z24,

-      fetch_texel_1d_f_s8_z24,

-      fetch_texel_2d_f_s8_z24,

-      fetch_texel_3d_f_s8_z24,

-      store_texel_s8_z24

-   },

-   {

-      MESA_FORMAT_Z16,

-      fetch_texel_1d_f_z16,

-      fetch_texel_2d_f_z16,

-      fetch_texel_3d_f_z16,

-      store_texel_z16

-   },

-   {

-      MESA_FORMAT_X8_Z24,

-      fetch_texel_1d_f_s8_z24,

-      fetch_texel_2d_f_s8_z24,

-      fetch_texel_3d_f_s8_z24,

-      store_texel_s8_z24

-   },

-   {

-      MESA_FORMAT_Z24_X8,

-      fetch_texel_1d_f_z24_s8,

-      fetch_texel_2d_f_z24_s8,

-      fetch_texel_3d_f_z24_s8,

-      store_texel_z24_s8

-   },

-   {

-      MESA_FORMAT_Z32,

-      fetch_texel_1d_f_z32,

-      fetch_texel_2d_f_z32,

-      fetch_texel_3d_f_z32,

-      store_texel_z32

-   },

-   {

-      MESA_FORMAT_S8,

-      NULL,

-      NULL,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SRGB8,

-      fetch_texel_1d_srgb8,

-      fetch_texel_2d_srgb8,

-      fetch_texel_3d_srgb8,

-      store_texel_srgb8

-   },

-   {

-      MESA_FORMAT_SRGBA8,

-      fetch_texel_1d_srgba8,

-      fetch_texel_2d_srgba8,

-      fetch_texel_3d_srgba8,

-      store_texel_srgba8

-   },

-   {

-      MESA_FORMAT_SARGB8,

-      fetch_texel_1d_sargb8,

-      fetch_texel_2d_sargb8,

-      fetch_texel_3d_sargb8,

-      store_texel_sargb8

-   },

-   {

-      MESA_FORMAT_SL8,

-      fetch_texel_1d_sl8,

-      fetch_texel_2d_sl8,

-      fetch_texel_3d_sl8,

-      store_texel_sl8

-   },

-   {

-      MESA_FORMAT_SLA8,

-      fetch_texel_1d_sla8,

-      fetch_texel_2d_sla8,

-      fetch_texel_3d_sla8,

-      store_texel_sla8

-   },

-   {

-      MESA_FORMAT_SRGB_DXT1,

-      NULL,

-      _mesa_fetch_texel_2d_f_srgb_dxt1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SRGBA_DXT1,

-      NULL,

-      _mesa_fetch_texel_2d_f_srgba_dxt1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SRGBA_DXT3,

-      NULL,

-      _mesa_fetch_texel_2d_f_srgba_dxt3,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SRGBA_DXT5,

-      NULL,

-      _mesa_fetch_texel_2d_f_srgba_dxt5,

-      NULL,

-      NULL

-   },

-

-   {

-      MESA_FORMAT_RGB_FXT1,

-      NULL,

-      _mesa_fetch_texel_2d_f_rgb_fxt1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_RGBA_FXT1,

-      NULL,

-      _mesa_fetch_texel_2d_f_rgba_fxt1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_RGB_DXT1,

-      NULL,

-      _mesa_fetch_texel_2d_f_rgb_dxt1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_RGBA_DXT1,

-      NULL,

-      _mesa_fetch_texel_2d_f_rgba_dxt1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_RGBA_DXT3,

-      NULL,

-      _mesa_fetch_texel_2d_f_rgba_dxt3,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_RGBA_DXT5,

-      NULL,

-      _mesa_fetch_texel_2d_f_rgba_dxt5,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_RGBA_FLOAT32,

-      fetch_texel_1d_f_rgba_f32,

-      fetch_texel_2d_f_rgba_f32,

-      fetch_texel_3d_f_rgba_f32,

-      store_texel_rgba_f32

-   },

-   {

-      MESA_FORMAT_RGBA_FLOAT16,

-      fetch_texel_1d_f_rgba_f16,

-      fetch_texel_2d_f_rgba_f16,

-      fetch_texel_3d_f_rgba_f16,

-      store_texel_rgba_f16

-   },

-   {

-      MESA_FORMAT_RGB_FLOAT32,

-      fetch_texel_1d_f_rgb_f32,

-      fetch_texel_2d_f_rgb_f32,

-      fetch_texel_3d_f_rgb_f32,

-      store_texel_rgb_f32

-   },

-   {

-      MESA_FORMAT_RGB_FLOAT16,

-      fetch_texel_1d_f_rgb_f16,

-      fetch_texel_2d_f_rgb_f16,

-      fetch_texel_3d_f_rgb_f16,

-      store_texel_rgb_f16

-   },

-   {

-      MESA_FORMAT_ALPHA_FLOAT32,

-      fetch_texel_1d_f_alpha_f32,

-      fetch_texel_2d_f_alpha_f32,

-      fetch_texel_3d_f_alpha_f32,

-      store_texel_alpha_f32

-   },

-   {

-      MESA_FORMAT_ALPHA_FLOAT16,

-      fetch_texel_1d_f_alpha_f16,

-      fetch_texel_2d_f_alpha_f16,

-      fetch_texel_3d_f_alpha_f16,

-      store_texel_alpha_f16

-   },

-   {

-      MESA_FORMAT_LUMINANCE_FLOAT32,

-      fetch_texel_1d_f_luminance_f32,

-      fetch_texel_2d_f_luminance_f32,

-      fetch_texel_3d_f_luminance_f32,

-      store_texel_luminance_f32

-   },

-   {

-      MESA_FORMAT_LUMINANCE_FLOAT16,

-      fetch_texel_1d_f_luminance_f16,

-      fetch_texel_2d_f_luminance_f16,

-      fetch_texel_3d_f_luminance_f16,

-      store_texel_luminance_f16

-   },

-   {

-      MESA_FORMAT_LUMINANCE_ALPHA_FLOAT32,

-      fetch_texel_1d_f_luminance_alpha_f32,

-      fetch_texel_2d_f_luminance_alpha_f32,

-      fetch_texel_3d_f_luminance_alpha_f32,

-      store_texel_luminance_alpha_f32

-   },

-   {

-      MESA_FORMAT_LUMINANCE_ALPHA_FLOAT16,

-      fetch_texel_1d_f_luminance_alpha_f16,

-      fetch_texel_2d_f_luminance_alpha_f16,

-      fetch_texel_3d_f_luminance_alpha_f16,

-      store_texel_luminance_alpha_f16

-   },

-   {

-      MESA_FORMAT_INTENSITY_FLOAT32,

-      fetch_texel_1d_f_intensity_f32,

-      fetch_texel_2d_f_intensity_f32,

-      fetch_texel_3d_f_intensity_f32,

-      store_texel_intensity_f32

-   },

-   {

-      MESA_FORMAT_INTENSITY_FLOAT16,

-      fetch_texel_1d_f_intensity_f16,

-      fetch_texel_2d_f_intensity_f16,

-      fetch_texel_3d_f_intensity_f16,

-      store_texel_intensity_f16

-   },

-   {

-      MESA_FORMAT_R_FLOAT32,

-      fetch_texel_1d_f_r_f32,

-      fetch_texel_2d_f_r_f32,

-      fetch_texel_3d_f_r_f32,

-      store_texel_r_f32

-   },

-   {

-      MESA_FORMAT_R_FLOAT16,

-      fetch_texel_1d_f_r_f16,

-      fetch_texel_2d_f_r_f16,

-      fetch_texel_3d_f_r_f16,

-      store_texel_r_f16

-   },

-   {

-      MESA_FORMAT_RG_FLOAT32,

-      fetch_texel_1d_f_rg_f32,

-      fetch_texel_2d_f_rg_f32,

-      fetch_texel_3d_f_rg_f32,

-      store_texel_rg_f32

-   },

-   {

-      MESA_FORMAT_RG_FLOAT16,

-      fetch_texel_1d_f_rg_f16,

-      fetch_texel_2d_f_rg_f16,

-      fetch_texel_3d_f_rg_f16,

-      store_texel_rg_f16

-   },

-

-   /* non-normalized, signed int */

-   {

-      MESA_FORMAT_RGBA_INT8,

-      fetch_texel_1d_rgba_int8,

-      fetch_texel_2d_rgba_int8,

-      fetch_texel_3d_rgba_int8,

-      store_texel_rgba_int8

-   },

-   {

-      MESA_FORMAT_RGBA_INT16,

-      fetch_texel_1d_rgba_int16,

-      fetch_texel_2d_rgba_int16,

-      fetch_texel_3d_rgba_int16,

-      store_texel_rgba_int16

-   },

-   {

-      MESA_FORMAT_RGBA_INT32,

-      fetch_texel_1d_rgba_int32,

-      fetch_texel_2d_rgba_int32,

-      fetch_texel_3d_rgba_int32,

-      store_texel_rgba_int32

-   },

-

-   /* non-normalized, unsigned int */

-   {

-      MESA_FORMAT_RGBA_UINT8,

-      fetch_texel_1d_rgba_uint8,

-      fetch_texel_2d_rgba_uint8,

-      fetch_texel_3d_rgba_uint8,

-      store_texel_rgba_uint8

-   },

-   {

-      MESA_FORMAT_RGBA_UINT16,

-      fetch_texel_1d_rgba_uint16,

-      fetch_texel_2d_rgba_uint16,

-      fetch_texel_3d_rgba_uint16,

-      store_texel_rgba_uint16

-   },

-   {

-      MESA_FORMAT_RGBA_UINT32,

-      fetch_texel_1d_rgba_uint32,

-      fetch_texel_2d_rgba_uint32,

-      fetch_texel_3d_rgba_uint32,

-      store_texel_rgba_uint32

-   },

-

-   /* dudv */

-   {

-      MESA_FORMAT_DUDV8,

-      fetch_texel_1d_dudv8,

-      fetch_texel_2d_dudv8,

-      fetch_texel_3d_dudv8,

-      NULL

-   },

-

-   /* signed, normalized */

-   {

-      MESA_FORMAT_SIGNED_R8,

-      fetch_texel_1d_signed_r8,

-      fetch_texel_2d_signed_r8,

-      fetch_texel_3d_signed_r8,

-      store_texel_signed_r8

-   },

-   {

-      MESA_FORMAT_SIGNED_RG88_REV,

-      fetch_texel_1d_signed_rg88_rev,

-      fetch_texel_2d_signed_rg88_rev,

-      fetch_texel_3d_signed_rg88_rev,

-      store_texel_signed_rg88_rev

-   },

-   {

-      MESA_FORMAT_SIGNED_RGBX8888,

-      fetch_texel_1d_signed_rgbx8888,

-      fetch_texel_2d_signed_rgbx8888,

-      fetch_texel_3d_signed_rgbx8888,

-      store_texel_signed_rgbx8888

-   },

-   {

-      MESA_FORMAT_SIGNED_RGBA8888,

-      fetch_texel_1d_signed_rgba8888,

-      fetch_texel_2d_signed_rgba8888,

-      fetch_texel_3d_signed_rgba8888,

-      store_texel_signed_rgba8888

-   },

-   {

-      MESA_FORMAT_SIGNED_RGBA8888_REV,

-      fetch_texel_1d_signed_rgba8888_rev,

-      fetch_texel_2d_signed_rgba8888_rev,

-      fetch_texel_3d_signed_rgba8888_rev,

-      store_texel_signed_rgba8888_rev

-   },

-   {

-      MESA_FORMAT_SIGNED_R16,

-      fetch_texel_1d_signed_r16,

-      fetch_texel_2d_signed_r16,

-      fetch_texel_3d_signed_r16,

-      store_texel_signed_r16

-   },

-   {

-      MESA_FORMAT_SIGNED_GR1616,

-      fetch_texel_1d_signed_rg1616,

-      fetch_texel_2d_signed_rg1616,

-      fetch_texel_3d_signed_rg1616,

-      store_texel_signed_rg1616

-   },

-   {

-      MESA_FORMAT_SIGNED_RGB_16,

-      fetch_texel_1d_signed_rgb_16,

-      fetch_texel_2d_signed_rgb_16,

-      fetch_texel_3d_signed_rgb_16,

-      store_texel_signed_rgb_16

-   },

-   {

-      MESA_FORMAT_SIGNED_RGBA_16,

-      fetch_texel_1d_signed_rgba_16,

-      fetch_texel_2d_signed_rgba_16,

-      fetch_texel_3d_signed_rgba_16,

-      store_texel_signed_rgba_16

-   },

-   {

-      MESA_FORMAT_RGBA_16,

-      fetch_texel_1d_rgba_16,

-      fetch_texel_2d_rgba_16,

-      fetch_texel_3d_rgba_16,

-      store_texel_rgba_16

-   },

-   {

-      MESA_FORMAT_RED_RGTC1,

-      NULL,

-      _mesa_fetch_texel_2d_f_red_rgtc1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SIGNED_RED_RGTC1,

-      NULL,

-      _mesa_fetch_texel_2d_f_signed_red_rgtc1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_RG_RGTC2,

-      NULL,

-      _mesa_fetch_texel_2d_f_rg_rgtc2,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SIGNED_RG_RGTC2,

-      NULL,

-      _mesa_fetch_texel_2d_f_signed_rg_rgtc2,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_L_LATC1,

-      NULL,

-      _mesa_fetch_texel_2d_f_l_latc1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SIGNED_L_LATC1,

-      NULL,

-      _mesa_fetch_texel_2d_f_signed_l_latc1,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_LA_LATC2,

-      NULL,

-      _mesa_fetch_texel_2d_f_la_latc2,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SIGNED_LA_LATC2,

-      NULL,

-      _mesa_fetch_texel_2d_f_signed_la_latc2,

-      NULL,

-      NULL

-   },

-   {

-      MESA_FORMAT_SIGNED_A8,

-      fetch_texel_1d_signed_a8,

-      fetch_texel_2d_signed_a8,

-      fetch_texel_3d_signed_a8,

-      store_texel_signed_a8

-   },

-   {

-      MESA_FORMAT_SIGNED_L8,

-      fetch_texel_1d_signed_l8,

-      fetch_texel_2d_signed_l8,

-      fetch_texel_3d_signed_l8,

-      store_texel_signed_l8

-   },

-   {

-      MESA_FORMAT_SIGNED_AL88,

-      fetch_texel_1d_signed_al88,

-      fetch_texel_2d_signed_al88,

-      fetch_texel_3d_signed_al88,

-      store_texel_signed_al88

-   },

-   {

-      MESA_FORMAT_SIGNED_I8,

-      fetch_texel_1d_signed_i8,

-      fetch_texel_2d_signed_i8,

-      fetch_texel_3d_signed_i8,

-      store_texel_signed_i8

-   },

-   {

-      MESA_FORMAT_SIGNED_A16,

-      fetch_texel_1d_signed_a16,

-      fetch_texel_2d_signed_a16,

-      fetch_texel_3d_signed_a16,

-      store_texel_signed_a16

-   },

-   {

-      MESA_FORMAT_SIGNED_L16,

-      fetch_texel_1d_signed_l16,

-      fetch_texel_2d_signed_l16,

-      fetch_texel_3d_signed_l16,

-      store_texel_signed_l16

-   },

-   {

-      MESA_FORMAT_SIGNED_AL1616,

-      fetch_texel_1d_signed_al1616,

-      fetch_texel_2d_signed_al1616,

-      fetch_texel_3d_signed_al1616,

-      store_texel_signed_al1616

-   },

-   {

-      MESA_FORMAT_SIGNED_I16,

-      fetch_texel_1d_signed_i16,

-      fetch_texel_2d_signed_i16,

-      fetch_texel_3d_signed_i16,

-      store_texel_signed_i16

-   },

-   {

-      MESA_FORMAT_RGB9_E5_FLOAT,

-      fetch_texel_1d_rgb9_e5,

-      fetch_texel_2d_rgb9_e5,

-      fetch_texel_3d_rgb9_e5,

-      store_texel_rgb9_e5

-   },

-   {

-      MESA_FORMAT_R11_G11_B10_FLOAT,

-      fetch_texel_1d_r11_g11_b10f,

-      fetch_texel_2d_r11_g11_b10f,

-      fetch_texel_3d_r11_g11_b10f,

-      store_texel_r11_g11_b10f

-   }

-};

-

-

-FetchTexelFuncF

-_mesa_get_texel_fetch_func(gl_format format, GLuint dims)

-{

-#ifdef DEBUG

-   /* check that the table entries are sorted by format name */

-   gl_format fmt;

-   for (fmt = 0; fmt < MESA_FORMAT_COUNT; fmt++) {

-      assert(texfetch_funcs[fmt].Name == fmt);

-   }

-#endif

-

-   assert(Elements(texfetch_funcs) == MESA_FORMAT_COUNT);

-   assert(format < MESA_FORMAT_COUNT);

-

-   switch (dims) {

-   case 1:

-      return texfetch_funcs[format].Fetch1D;

-   case 2:

-      return texfetch_funcs[format].Fetch2D;

-   case 3:

-      return texfetch_funcs[format].Fetch3D;

-   default:

-      assert(0 && "bad dims in _mesa_get_texel_fetch_func");

-      return NULL;

-   }

-}

-

-

-StoreTexelFunc

-_mesa_get_texel_store_func(gl_format format)

-{

-   assert(format < MESA_FORMAT_COUNT);

-   return texfetch_funcs[format].StoreTexel;

-}

-

-

-/**

- * Adaptor for fetching a GLchan texel from a float-valued texture.

- */

-static void

-fetch_texel_float_to_chan(const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLchan *texelOut)

-{

-   GLfloat temp[4];

-   GLenum baseFormat = _mesa_get_format_base_format(texImage->TexFormat);

-

-   ASSERT(texImage->FetchTexelf);

-   texImage->FetchTexelf(texImage, i, j, k, temp);

-   if (baseFormat == GL_DEPTH_COMPONENT ||

-       baseFormat == GL_DEPTH_STENCIL_EXT) {

-      /* just one channel */

-      UNCLAMPED_FLOAT_TO_CHAN(texelOut[0], temp[0]);

-   }

-   else {

-      /* four channels */

-      UNCLAMPED_FLOAT_TO_CHAN(texelOut[0], temp[0]);

-      UNCLAMPED_FLOAT_TO_CHAN(texelOut[1], temp[1]);

-      UNCLAMPED_FLOAT_TO_CHAN(texelOut[2], temp[2]);

-      UNCLAMPED_FLOAT_TO_CHAN(texelOut[3], temp[3]);

-   }

-}

-

-

-#if 0

-/**

- * Adaptor for fetching a float texel from a GLchan-valued texture.

- */

-static void

-fetch_texel_chan_to_float(const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texelOut)

-{

-   GLchan temp[4];

-   GLenum baseFormat = _mesa_get_format_base_format(texImage->TexFormat);

-

-   ASSERT(texImage->FetchTexelc);

-   texImage->FetchTexelc(texImage, i, j, k, temp);

-   if (baseFormat == GL_DEPTH_COMPONENT ||

-       baseFormat == GL_DEPTH_STENCIL_EXT) {

-      /* just one channel */

-      texelOut[0] = CHAN_TO_FLOAT(temp[0]);

-   }

-   else {

-      /* four channels */

-      texelOut[0] = CHAN_TO_FLOAT(temp[0]);

-      texelOut[1] = CHAN_TO_FLOAT(temp[1]);

-      texelOut[2] = CHAN_TO_FLOAT(temp[2]);

-      texelOut[3] = CHAN_TO_FLOAT(temp[3]);

-   }

-}

-#endif

-

-

-/**

- * Initialize the texture image's FetchTexelc and FetchTexelf methods.

- */

-void

-_mesa_set_fetch_functions(struct gl_texture_image *texImage, GLuint dims)

-{

-   gl_format format = texImage->TexFormat;

-

-   ASSERT(dims == 1 || dims == 2 || dims == 3);

-

-   if (texImage->TexObject->Sampler.sRGBDecode == GL_SKIP_DECODE_EXT &&

-       _mesa_get_format_color_encoding(format) == GL_SRGB) {

-      format = _mesa_get_srgb_format_linear(format);

-   }

-

-   texImage->FetchTexelf = _mesa_get_texel_fetch_func(format, dims);

-

-   texImage->FetchTexelc = fetch_texel_float_to_chan;

-

-   ASSERT(texImage->FetchTexelc);

-   ASSERT(texImage->FetchTexelf);

-}

-

-void

-_mesa_update_fetch_functions(struct gl_texture_object *texObj)

-{

-   GLuint face, i;

-   GLuint dims;

-

-   dims = _mesa_get_texture_dimensions(texObj->Target);

-

-   for (face = 0; face < 6; face++) {

-      for (i = 0; i < MAX_TEXTURE_LEVELS; i++) {

-         if (texObj->Image[face][i]) {

-	    _mesa_set_fetch_functions(texObj->Image[face][i], dims);

-         }

-      }

-   }

-}

+/*
+ * Mesa 3-D graphics library
+ * Version:  7.7
+ *
+ * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.
+ * Copyright (c) 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.
+ */
+
+
+/**
+ * \file texfetch.c
+ *
+ * Texel fetch/store functions
+ *
+ * \author Gareth Hughes
+ */
+
+
+#include "colormac.h"
+#include "macros.h"
+#include "texcompress.h"
+#include "texcompress_fxt1.h"
+#include "texcompress_s3tc.h"
+#include "texcompress_rgtc.h"
+#include "texfetch.h"
+#include "teximage.h"
+#include "../../gallium/auxiliary/util/u_format_rgb9e5.h"
+#include "../../gallium/auxiliary/util/u_format_r11g11b10f.h"
+
+
+/**
+ * Convert an 8-bit sRGB value from non-linear space to a
+ * linear RGB value in [0, 1].
+ * Implemented with a 256-entry lookup table.
+ */
+static INLINE GLfloat
+nonlinear_to_linear(GLubyte cs8)
+{
+   static GLfloat table[256];
+   static GLboolean tableReady = GL_FALSE;
+   if (!tableReady) {
+      /* compute lookup table now */
+      GLuint i;
+      for (i = 0; i < 256; i++) {
+         const GLfloat cs = UBYTE_TO_FLOAT(i);
+         if (cs <= 0.04045) {
+            table[i] = cs / 12.92f;
+         }
+         else {
+            table[i] = (GLfloat) pow((cs + 0.055) / 1.055, 2.4);
+         }
+      }
+      tableReady = GL_TRUE;
+   }
+   return table[cs8];
+}
+
+
+
+/* Texel fetch routines for all supported formats
+ */
+#define DIM 1
+#include "texfetch_tmp.h"
+
+#define DIM 2
+#include "texfetch_tmp.h"
+
+#define DIM 3
+#include "texfetch_tmp.h"
+
+/**
+ * Null texel fetch function.
+ *
+ * Have to have this so the FetchTexel function pointer is never NULL.
+ */
+static void fetch_null_texelf( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   (void) texImage; (void) i; (void) j; (void) k;
+   texel[RCOMP] = 0.0;
+   texel[GCOMP] = 0.0;
+   texel[BCOMP] = 0.0;
+   texel[ACOMP] = 0.0;
+   _mesa_warning(NULL, "fetch_null_texelf() called!");
+}
+
+static void store_null_texel(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   (void) texImage;
+   (void) i;
+   (void) j;
+   (void) k;
+   (void) texel;
+   /* no-op */
+}
+
+
+
+/**
+ * Table to map MESA_FORMAT_ to texel fetch/store funcs.
+ * XXX this is somewhat temporary.
+ */
+static struct {
+   gl_format Name;
+   FetchTexelFuncF Fetch1D;
+   FetchTexelFuncF Fetch2D;
+   FetchTexelFuncF Fetch3D;
+   StoreTexelFunc StoreTexel;
+}
+texfetch_funcs[MESA_FORMAT_COUNT] =
+{
+   {
+      MESA_FORMAT_NONE,
+      fetch_null_texelf,
+      fetch_null_texelf,
+      fetch_null_texelf,
+      store_null_texel
+   },
+
+   {
+      MESA_FORMAT_RGBA8888,
+      fetch_texel_1d_f_rgba8888,
+      fetch_texel_2d_f_rgba8888,
+      fetch_texel_3d_f_rgba8888,
+      store_texel_rgba8888
+   },
+   {
+      MESA_FORMAT_RGBA8888_REV,
+      fetch_texel_1d_f_rgba8888_rev,
+      fetch_texel_2d_f_rgba8888_rev,
+      fetch_texel_3d_f_rgba8888_rev,
+      store_texel_rgba8888_rev
+   },
+   {
+      MESA_FORMAT_ARGB8888,
+      fetch_texel_1d_f_argb8888,
+      fetch_texel_2d_f_argb8888,
+      fetch_texel_3d_f_argb8888,
+      store_texel_argb8888
+   },
+   {
+      MESA_FORMAT_ARGB8888_REV,
+      fetch_texel_1d_f_argb8888_rev,
+      fetch_texel_2d_f_argb8888_rev,
+      fetch_texel_3d_f_argb8888_rev,
+      store_texel_argb8888_rev
+   },
+   {
+      MESA_FORMAT_XRGB8888,
+      fetch_texel_1d_f_xrgb8888,
+      fetch_texel_2d_f_xrgb8888,
+      fetch_texel_3d_f_xrgb8888,
+      store_texel_xrgb8888
+   },
+   {
+      MESA_FORMAT_XRGB8888_REV,
+      fetch_texel_1d_f_xrgb8888_rev,
+      fetch_texel_2d_f_xrgb8888_rev,
+      fetch_texel_3d_f_xrgb8888_rev,
+      store_texel_xrgb8888_rev,
+   },
+   {
+      MESA_FORMAT_RGB888,
+      fetch_texel_1d_f_rgb888,
+      fetch_texel_2d_f_rgb888,
+      fetch_texel_3d_f_rgb888,
+      store_texel_rgb888
+   },
+   {
+      MESA_FORMAT_BGR888,
+      fetch_texel_1d_f_bgr888,
+      fetch_texel_2d_f_bgr888,
+      fetch_texel_3d_f_bgr888,
+      store_texel_bgr888
+   },
+   {
+      MESA_FORMAT_RGB565,
+      fetch_texel_1d_f_rgb565,
+      fetch_texel_2d_f_rgb565,
+      fetch_texel_3d_f_rgb565,
+      store_texel_rgb565
+   },
+   {
+      MESA_FORMAT_RGB565_REV,
+      fetch_texel_1d_f_rgb565_rev,
+      fetch_texel_2d_f_rgb565_rev,
+      fetch_texel_3d_f_rgb565_rev,
+      store_texel_rgb565_rev
+   },
+   {
+      MESA_FORMAT_ARGB4444,
+      fetch_texel_1d_f_argb4444,
+      fetch_texel_2d_f_argb4444,
+      fetch_texel_3d_f_argb4444,
+      store_texel_argb4444
+   },
+   {
+      MESA_FORMAT_ARGB4444_REV,
+      fetch_texel_1d_f_argb4444_rev,
+      fetch_texel_2d_f_argb4444_rev,
+      fetch_texel_3d_f_argb4444_rev,
+      store_texel_argb4444_rev
+   },
+   {
+      MESA_FORMAT_RGBA5551,
+      fetch_texel_1d_f_rgba5551,
+      fetch_texel_2d_f_rgba5551,
+      fetch_texel_3d_f_rgba5551,
+      store_texel_rgba5551
+   },
+   {
+      MESA_FORMAT_ARGB1555,
+      fetch_texel_1d_f_argb1555,
+      fetch_texel_2d_f_argb1555,
+      fetch_texel_3d_f_argb1555,
+      store_texel_argb1555
+   },
+   {
+      MESA_FORMAT_ARGB1555_REV,
+      fetch_texel_1d_f_argb1555_rev,
+      fetch_texel_2d_f_argb1555_rev,
+      fetch_texel_3d_f_argb1555_rev,
+      store_texel_argb1555_rev
+   },
+   {
+      MESA_FORMAT_AL44,
+      fetch_texel_1d_f_al44,
+      fetch_texel_2d_f_al44,
+      fetch_texel_3d_f_al44,
+      store_texel_al44
+   },
+   {
+      MESA_FORMAT_AL88,
+      fetch_texel_1d_f_al88,
+      fetch_texel_2d_f_al88,
+      fetch_texel_3d_f_al88,
+      store_texel_al88
+   },
+   {
+      MESA_FORMAT_AL88_REV,
+      fetch_texel_1d_f_al88_rev,
+      fetch_texel_2d_f_al88_rev,
+      fetch_texel_3d_f_al88_rev,
+      store_texel_al88_rev
+   },
+   {
+      MESA_FORMAT_AL1616,
+      fetch_texel_1d_f_al1616,
+      fetch_texel_2d_f_al1616,
+      fetch_texel_3d_f_al1616,
+      store_texel_al1616
+   },
+   {
+      MESA_FORMAT_AL1616_REV,
+      fetch_texel_1d_f_al1616_rev,
+      fetch_texel_2d_f_al1616_rev,
+      fetch_texel_3d_f_al1616_rev,
+      store_texel_al1616_rev
+   },
+   {
+      MESA_FORMAT_RGB332,
+      fetch_texel_1d_f_rgb332,
+      fetch_texel_2d_f_rgb332,
+      fetch_texel_3d_f_rgb332,
+      store_texel_rgb332
+   },
+   {
+      MESA_FORMAT_A8,
+      fetch_texel_1d_f_a8,
+      fetch_texel_2d_f_a8,
+      fetch_texel_3d_f_a8,
+      store_texel_a8
+   },
+   {
+      MESA_FORMAT_A16,
+      fetch_texel_1d_f_a16,
+      fetch_texel_2d_f_a16,
+      fetch_texel_3d_f_a16,
+      store_texel_a16
+   },
+   {
+      MESA_FORMAT_L8,
+      fetch_texel_1d_f_l8,
+      fetch_texel_2d_f_l8,
+      fetch_texel_3d_f_l8,
+      store_texel_l8
+   },
+   {
+      MESA_FORMAT_L16,
+      fetch_texel_1d_f_l16,
+      fetch_texel_2d_f_l16,
+      fetch_texel_3d_f_l16,
+      store_texel_l16
+   },
+   {
+      MESA_FORMAT_I8,
+      fetch_texel_1d_f_i8,
+      fetch_texel_2d_f_i8,
+      fetch_texel_3d_f_i8,
+      store_texel_i8
+   },
+   {
+      MESA_FORMAT_I16,
+      fetch_texel_1d_f_i16,
+      fetch_texel_2d_f_i16,
+      fetch_texel_3d_f_i16,
+      store_texel_i16
+   },
+   {
+      MESA_FORMAT_CI8,
+      fetch_texel_1d_f_ci8,
+      fetch_texel_2d_f_ci8,
+      fetch_texel_3d_f_ci8,
+      store_texel_ci8
+   },
+   {
+      MESA_FORMAT_YCBCR,
+      fetch_texel_1d_f_ycbcr,
+      fetch_texel_2d_f_ycbcr,
+      fetch_texel_3d_f_ycbcr,
+      store_texel_ycbcr
+   },
+   {
+      MESA_FORMAT_YCBCR_REV,
+      fetch_texel_1d_f_ycbcr_rev,
+      fetch_texel_2d_f_ycbcr_rev,
+      fetch_texel_3d_f_ycbcr_rev,
+      store_texel_ycbcr_rev
+   },
+   {
+      MESA_FORMAT_R8,
+      fetch_texel_1d_f_r8,
+      fetch_texel_2d_f_r8,
+      fetch_texel_3d_f_r8,
+      store_texel_r8,
+   },
+   {
+      MESA_FORMAT_RG88,
+      fetch_texel_1d_f_rg88,
+      fetch_texel_2d_f_rg88,
+      fetch_texel_3d_f_rg88,
+      store_texel_rg88,
+   },
+   {
+      MESA_FORMAT_RG88_REV,
+      fetch_texel_1d_f_rg88_rev,
+      fetch_texel_2d_f_rg88_rev,
+      fetch_texel_3d_f_rg88_rev,
+      store_texel_rg88_rev,
+   },
+   {
+      MESA_FORMAT_R16,
+      fetch_texel_1d_f_r16,
+      fetch_texel_2d_f_r16,
+      fetch_texel_3d_f_r16,
+      store_texel_r16,
+   },
+   {
+      MESA_FORMAT_RG1616,
+      fetch_texel_1d_f_rg1616,
+      fetch_texel_2d_f_rg1616,
+      fetch_texel_3d_f_rg1616,
+      store_texel_rg1616,
+   },
+   {
+      MESA_FORMAT_RG1616_REV,
+      fetch_texel_1d_f_rg1616_rev,
+      fetch_texel_2d_f_rg1616_rev,
+      fetch_texel_3d_f_rg1616_rev,
+      store_texel_rg1616_rev,
+   },
+   {
+      MESA_FORMAT_ARGB2101010,
+      fetch_texel_1d_f_argb2101010,
+      fetch_texel_2d_f_argb2101010,
+      fetch_texel_3d_f_argb2101010,
+      store_texel_argb2101010
+   },
+   {
+      MESA_FORMAT_Z24_S8,
+      fetch_texel_1d_f_z24_s8,
+      fetch_texel_2d_f_z24_s8,
+      fetch_texel_3d_f_z24_s8,
+      store_texel_z24_s8
+   },
+   {
+      MESA_FORMAT_S8_Z24,
+      fetch_texel_1d_f_s8_z24,
+      fetch_texel_2d_f_s8_z24,
+      fetch_texel_3d_f_s8_z24,
+      store_texel_s8_z24
+   },
+   {
+      MESA_FORMAT_Z16,
+      fetch_texel_1d_f_z16,
+      fetch_texel_2d_f_z16,
+      fetch_texel_3d_f_z16,
+      store_texel_z16
+   },
+   {
+      MESA_FORMAT_X8_Z24,
+      fetch_texel_1d_f_s8_z24,
+      fetch_texel_2d_f_s8_z24,
+      fetch_texel_3d_f_s8_z24,
+      store_texel_s8_z24
+   },
+   {
+      MESA_FORMAT_Z24_X8,
+      fetch_texel_1d_f_z24_s8,
+      fetch_texel_2d_f_z24_s8,
+      fetch_texel_3d_f_z24_s8,
+      store_texel_z24_s8
+   },
+   {
+      MESA_FORMAT_Z32,
+      fetch_texel_1d_f_z32,
+      fetch_texel_2d_f_z32,
+      fetch_texel_3d_f_z32,
+      store_texel_z32
+   },
+   {
+      MESA_FORMAT_S8,
+      NULL,
+      NULL,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SRGB8,
+      fetch_texel_1d_srgb8,
+      fetch_texel_2d_srgb8,
+      fetch_texel_3d_srgb8,
+      store_texel_srgb8
+   },
+   {
+      MESA_FORMAT_SRGBA8,
+      fetch_texel_1d_srgba8,
+      fetch_texel_2d_srgba8,
+      fetch_texel_3d_srgba8,
+      store_texel_srgba8
+   },
+   {
+      MESA_FORMAT_SARGB8,
+      fetch_texel_1d_sargb8,
+      fetch_texel_2d_sargb8,
+      fetch_texel_3d_sargb8,
+      store_texel_sargb8
+   },
+   {
+      MESA_FORMAT_SL8,
+      fetch_texel_1d_sl8,
+      fetch_texel_2d_sl8,
+      fetch_texel_3d_sl8,
+      store_texel_sl8
+   },
+   {
+      MESA_FORMAT_SLA8,
+      fetch_texel_1d_sla8,
+      fetch_texel_2d_sla8,
+      fetch_texel_3d_sla8,
+      store_texel_sla8
+   },
+   {
+      MESA_FORMAT_SRGB_DXT1,
+      NULL,
+      _mesa_fetch_texel_2d_f_srgb_dxt1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SRGBA_DXT1,
+      NULL,
+      _mesa_fetch_texel_2d_f_srgba_dxt1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SRGBA_DXT3,
+      NULL,
+      _mesa_fetch_texel_2d_f_srgba_dxt3,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SRGBA_DXT5,
+      NULL,
+      _mesa_fetch_texel_2d_f_srgba_dxt5,
+      NULL,
+      NULL
+   },
+
+   {
+      MESA_FORMAT_RGB_FXT1,
+      NULL,
+      _mesa_fetch_texel_2d_f_rgb_fxt1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_RGBA_FXT1,
+      NULL,
+      _mesa_fetch_texel_2d_f_rgba_fxt1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_RGB_DXT1,
+      NULL,
+      _mesa_fetch_texel_2d_f_rgb_dxt1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_RGBA_DXT1,
+      NULL,
+      _mesa_fetch_texel_2d_f_rgba_dxt1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_RGBA_DXT3,
+      NULL,
+      _mesa_fetch_texel_2d_f_rgba_dxt3,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_RGBA_DXT5,
+      NULL,
+      _mesa_fetch_texel_2d_f_rgba_dxt5,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_RGBA_FLOAT32,
+      fetch_texel_1d_f_rgba_f32,
+      fetch_texel_2d_f_rgba_f32,
+      fetch_texel_3d_f_rgba_f32,
+      store_texel_rgba_f32
+   },
+   {
+      MESA_FORMAT_RGBA_FLOAT16,
+      fetch_texel_1d_f_rgba_f16,
+      fetch_texel_2d_f_rgba_f16,
+      fetch_texel_3d_f_rgba_f16,
+      store_texel_rgba_f16
+   },
+   {
+      MESA_FORMAT_RGB_FLOAT32,
+      fetch_texel_1d_f_rgb_f32,
+      fetch_texel_2d_f_rgb_f32,
+      fetch_texel_3d_f_rgb_f32,
+      store_texel_rgb_f32
+   },
+   {
+      MESA_FORMAT_RGB_FLOAT16,
+      fetch_texel_1d_f_rgb_f16,
+      fetch_texel_2d_f_rgb_f16,
+      fetch_texel_3d_f_rgb_f16,
+      store_texel_rgb_f16
+   },
+   {
+      MESA_FORMAT_ALPHA_FLOAT32,
+      fetch_texel_1d_f_alpha_f32,
+      fetch_texel_2d_f_alpha_f32,
+      fetch_texel_3d_f_alpha_f32,
+      store_texel_alpha_f32
+   },
+   {
+      MESA_FORMAT_ALPHA_FLOAT16,
+      fetch_texel_1d_f_alpha_f16,
+      fetch_texel_2d_f_alpha_f16,
+      fetch_texel_3d_f_alpha_f16,
+      store_texel_alpha_f16
+   },
+   {
+      MESA_FORMAT_LUMINANCE_FLOAT32,
+      fetch_texel_1d_f_luminance_f32,
+      fetch_texel_2d_f_luminance_f32,
+      fetch_texel_3d_f_luminance_f32,
+      store_texel_luminance_f32
+   },
+   {
+      MESA_FORMAT_LUMINANCE_FLOAT16,
+      fetch_texel_1d_f_luminance_f16,
+      fetch_texel_2d_f_luminance_f16,
+      fetch_texel_3d_f_luminance_f16,
+      store_texel_luminance_f16
+   },
+   {
+      MESA_FORMAT_LUMINANCE_ALPHA_FLOAT32,
+      fetch_texel_1d_f_luminance_alpha_f32,
+      fetch_texel_2d_f_luminance_alpha_f32,
+      fetch_texel_3d_f_luminance_alpha_f32,
+      store_texel_luminance_alpha_f32
+   },
+   {
+      MESA_FORMAT_LUMINANCE_ALPHA_FLOAT16,
+      fetch_texel_1d_f_luminance_alpha_f16,
+      fetch_texel_2d_f_luminance_alpha_f16,
+      fetch_texel_3d_f_luminance_alpha_f16,
+      store_texel_luminance_alpha_f16
+   },
+   {
+      MESA_FORMAT_INTENSITY_FLOAT32,
+      fetch_texel_1d_f_intensity_f32,
+      fetch_texel_2d_f_intensity_f32,
+      fetch_texel_3d_f_intensity_f32,
+      store_texel_intensity_f32
+   },
+   {
+      MESA_FORMAT_INTENSITY_FLOAT16,
+      fetch_texel_1d_f_intensity_f16,
+      fetch_texel_2d_f_intensity_f16,
+      fetch_texel_3d_f_intensity_f16,
+      store_texel_intensity_f16
+   },
+   {
+      MESA_FORMAT_R_FLOAT32,
+      fetch_texel_1d_f_r_f32,
+      fetch_texel_2d_f_r_f32,
+      fetch_texel_3d_f_r_f32,
+      store_texel_r_f32
+   },
+   {
+      MESA_FORMAT_R_FLOAT16,
+      fetch_texel_1d_f_r_f16,
+      fetch_texel_2d_f_r_f16,
+      fetch_texel_3d_f_r_f16,
+      store_texel_r_f16
+   },
+   {
+      MESA_FORMAT_RG_FLOAT32,
+      fetch_texel_1d_f_rg_f32,
+      fetch_texel_2d_f_rg_f32,
+      fetch_texel_3d_f_rg_f32,
+      store_texel_rg_f32
+   },
+   {
+      MESA_FORMAT_RG_FLOAT16,
+      fetch_texel_1d_f_rg_f16,
+      fetch_texel_2d_f_rg_f16,
+      fetch_texel_3d_f_rg_f16,
+      store_texel_rg_f16
+   },
+
+   /* non-normalized, signed int */
+   {
+      MESA_FORMAT_RGBA_INT8,
+      fetch_texel_1d_rgba_int8,
+      fetch_texel_2d_rgba_int8,
+      fetch_texel_3d_rgba_int8,
+      store_texel_rgba_int8
+   },
+   {
+      MESA_FORMAT_RGBA_INT16,
+      fetch_texel_1d_rgba_int16,
+      fetch_texel_2d_rgba_int16,
+      fetch_texel_3d_rgba_int16,
+      store_texel_rgba_int16
+   },
+   {
+      MESA_FORMAT_RGBA_INT32,
+      fetch_texel_1d_rgba_int32,
+      fetch_texel_2d_rgba_int32,
+      fetch_texel_3d_rgba_int32,
+      store_texel_rgba_int32
+   },
+
+   /* non-normalized, unsigned int */
+   {
+      MESA_FORMAT_RGBA_UINT8,
+      fetch_texel_1d_rgba_uint8,
+      fetch_texel_2d_rgba_uint8,
+      fetch_texel_3d_rgba_uint8,
+      store_texel_rgba_uint8
+   },
+   {
+      MESA_FORMAT_RGBA_UINT16,
+      fetch_texel_1d_rgba_uint16,
+      fetch_texel_2d_rgba_uint16,
+      fetch_texel_3d_rgba_uint16,
+      store_texel_rgba_uint16
+   },
+   {
+      MESA_FORMAT_RGBA_UINT32,
+      fetch_texel_1d_rgba_uint32,
+      fetch_texel_2d_rgba_uint32,
+      fetch_texel_3d_rgba_uint32,
+      store_texel_rgba_uint32
+   },
+
+   /* dudv */
+   {
+      MESA_FORMAT_DUDV8,
+      fetch_texel_1d_dudv8,
+      fetch_texel_2d_dudv8,
+      fetch_texel_3d_dudv8,
+      NULL
+   },
+
+   /* signed, normalized */
+   {
+      MESA_FORMAT_SIGNED_R8,
+      fetch_texel_1d_signed_r8,
+      fetch_texel_2d_signed_r8,
+      fetch_texel_3d_signed_r8,
+      store_texel_signed_r8
+   },
+   {
+      MESA_FORMAT_SIGNED_RG88_REV,
+      fetch_texel_1d_signed_rg88_rev,
+      fetch_texel_2d_signed_rg88_rev,
+      fetch_texel_3d_signed_rg88_rev,
+      store_texel_signed_rg88_rev
+   },
+   {
+      MESA_FORMAT_SIGNED_RGBX8888,
+      fetch_texel_1d_signed_rgbx8888,
+      fetch_texel_2d_signed_rgbx8888,
+      fetch_texel_3d_signed_rgbx8888,
+      store_texel_signed_rgbx8888
+   },
+   {
+      MESA_FORMAT_SIGNED_RGBA8888,
+      fetch_texel_1d_signed_rgba8888,
+      fetch_texel_2d_signed_rgba8888,
+      fetch_texel_3d_signed_rgba8888,
+      store_texel_signed_rgba8888
+   },
+   {
+      MESA_FORMAT_SIGNED_RGBA8888_REV,
+      fetch_texel_1d_signed_rgba8888_rev,
+      fetch_texel_2d_signed_rgba8888_rev,
+      fetch_texel_3d_signed_rgba8888_rev,
+      store_texel_signed_rgba8888_rev
+   },
+   {
+      MESA_FORMAT_SIGNED_R16,
+      fetch_texel_1d_signed_r16,
+      fetch_texel_2d_signed_r16,
+      fetch_texel_3d_signed_r16,
+      store_texel_signed_r16
+   },
+   {
+      MESA_FORMAT_SIGNED_GR1616,
+      fetch_texel_1d_signed_rg1616,
+      fetch_texel_2d_signed_rg1616,
+      fetch_texel_3d_signed_rg1616,
+      store_texel_signed_rg1616
+   },
+   {
+      MESA_FORMAT_SIGNED_RGB_16,
+      fetch_texel_1d_signed_rgb_16,
+      fetch_texel_2d_signed_rgb_16,
+      fetch_texel_3d_signed_rgb_16,
+      store_texel_signed_rgb_16
+   },
+   {
+      MESA_FORMAT_SIGNED_RGBA_16,
+      fetch_texel_1d_signed_rgba_16,
+      fetch_texel_2d_signed_rgba_16,
+      fetch_texel_3d_signed_rgba_16,
+      store_texel_signed_rgba_16
+   },
+   {
+      MESA_FORMAT_RGBA_16,
+      fetch_texel_1d_rgba_16,
+      fetch_texel_2d_rgba_16,
+      fetch_texel_3d_rgba_16,
+      store_texel_rgba_16
+   },
+   {
+      MESA_FORMAT_RED_RGTC1,
+      NULL,
+      _mesa_fetch_texel_2d_f_red_rgtc1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SIGNED_RED_RGTC1,
+      NULL,
+      _mesa_fetch_texel_2d_f_signed_red_rgtc1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_RG_RGTC2,
+      NULL,
+      _mesa_fetch_texel_2d_f_rg_rgtc2,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SIGNED_RG_RGTC2,
+      NULL,
+      _mesa_fetch_texel_2d_f_signed_rg_rgtc2,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_L_LATC1,
+      NULL,
+      _mesa_fetch_texel_2d_f_l_latc1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SIGNED_L_LATC1,
+      NULL,
+      _mesa_fetch_texel_2d_f_signed_l_latc1,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_LA_LATC2,
+      NULL,
+      _mesa_fetch_texel_2d_f_la_latc2,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SIGNED_LA_LATC2,
+      NULL,
+      _mesa_fetch_texel_2d_f_signed_la_latc2,
+      NULL,
+      NULL
+   },
+   {
+      MESA_FORMAT_SIGNED_A8,
+      fetch_texel_1d_signed_a8,
+      fetch_texel_2d_signed_a8,
+      fetch_texel_3d_signed_a8,
+      store_texel_signed_a8
+   },
+   {
+      MESA_FORMAT_SIGNED_L8,
+      fetch_texel_1d_signed_l8,
+      fetch_texel_2d_signed_l8,
+      fetch_texel_3d_signed_l8,
+      store_texel_signed_l8
+   },
+   {
+      MESA_FORMAT_SIGNED_AL88,
+      fetch_texel_1d_signed_al88,
+      fetch_texel_2d_signed_al88,
+      fetch_texel_3d_signed_al88,
+      store_texel_signed_al88
+   },
+   {
+      MESA_FORMAT_SIGNED_I8,
+      fetch_texel_1d_signed_i8,
+      fetch_texel_2d_signed_i8,
+      fetch_texel_3d_signed_i8,
+      store_texel_signed_i8
+   },
+   {
+      MESA_FORMAT_SIGNED_A16,
+      fetch_texel_1d_signed_a16,
+      fetch_texel_2d_signed_a16,
+      fetch_texel_3d_signed_a16,
+      store_texel_signed_a16
+   },
+   {
+      MESA_FORMAT_SIGNED_L16,
+      fetch_texel_1d_signed_l16,
+      fetch_texel_2d_signed_l16,
+      fetch_texel_3d_signed_l16,
+      store_texel_signed_l16
+   },
+   {
+      MESA_FORMAT_SIGNED_AL1616,
+      fetch_texel_1d_signed_al1616,
+      fetch_texel_2d_signed_al1616,
+      fetch_texel_3d_signed_al1616,
+      store_texel_signed_al1616
+   },
+   {
+      MESA_FORMAT_SIGNED_I16,
+      fetch_texel_1d_signed_i16,
+      fetch_texel_2d_signed_i16,
+      fetch_texel_3d_signed_i16,
+      store_texel_signed_i16
+   },
+   {
+      MESA_FORMAT_RGB9_E5_FLOAT,
+      fetch_texel_1d_rgb9_e5,
+      fetch_texel_2d_rgb9_e5,
+      fetch_texel_3d_rgb9_e5,
+      store_texel_rgb9_e5
+   },
+   {
+      MESA_FORMAT_R11_G11_B10_FLOAT,
+      fetch_texel_1d_r11_g11_b10f,
+      fetch_texel_2d_r11_g11_b10f,
+      fetch_texel_3d_r11_g11_b10f,
+      store_texel_r11_g11_b10f
+   },
+   {
+      MESA_FORMAT_Z32_FLOAT,
+      fetch_texel_1d_f_r_f32, /* Reuse the R32F functions. */
+      fetch_texel_2d_f_r_f32,
+      fetch_texel_3d_f_r_f32,
+      store_texel_r_f32
+   },
+   {
+      MESA_FORMAT_Z32_FLOAT_X24S8,
+      fetch_texel_1d_z32f_x24s8,
+      fetch_texel_2d_z32f_x24s8,
+      fetch_texel_3d_z32f_x24s8,
+      store_texel_z32f_x24s8
+   }
+};
+
+
+FetchTexelFuncF
+_mesa_get_texel_fetch_func(gl_format format, GLuint dims)
+{
+#ifdef DEBUG
+   /* check that the table entries are sorted by format name */
+   gl_format fmt;
+   for (fmt = 0; fmt < MESA_FORMAT_COUNT; fmt++) {
+      assert(texfetch_funcs[fmt].Name == fmt);
+   }
+#endif
+
+   assert(Elements(texfetch_funcs) == MESA_FORMAT_COUNT);
+   assert(format < MESA_FORMAT_COUNT);
+
+   switch (dims) {
+   case 1:
+      return texfetch_funcs[format].Fetch1D;
+   case 2:
+      return texfetch_funcs[format].Fetch2D;
+   case 3:
+      return texfetch_funcs[format].Fetch3D;
+   default:
+      assert(0 && "bad dims in _mesa_get_texel_fetch_func");
+      return NULL;
+   }
+}
+
+
+StoreTexelFunc
+_mesa_get_texel_store_func(gl_format format)
+{
+   assert(format < MESA_FORMAT_COUNT);
+   return texfetch_funcs[format].StoreTexel;
+}
+
+
+/**
+ * Adaptor for fetching a GLchan texel from a float-valued texture.
+ */
+static void
+fetch_texel_float_to_chan(const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLchan *texelOut)
+{
+   GLfloat temp[4];
+   GLenum baseFormat = _mesa_get_format_base_format(texImage->TexFormat);
+
+   ASSERT(texImage->FetchTexelf);
+   texImage->FetchTexelf(texImage, i, j, k, temp);
+   if (baseFormat == GL_DEPTH_COMPONENT ||
+       baseFormat == GL_DEPTH_STENCIL_EXT) {
+      /* just one channel */
+      UNCLAMPED_FLOAT_TO_CHAN(texelOut[0], temp[0]);
+   }
+   else {
+      /* four channels */
+      UNCLAMPED_FLOAT_TO_CHAN(texelOut[0], temp[0]);
+      UNCLAMPED_FLOAT_TO_CHAN(texelOut[1], temp[1]);
+      UNCLAMPED_FLOAT_TO_CHAN(texelOut[2], temp[2]);
+      UNCLAMPED_FLOAT_TO_CHAN(texelOut[3], temp[3]);
+   }
+}
+
+
+#if 0
+/**
+ * Adaptor for fetching a float texel from a GLchan-valued texture.
+ */
+static void
+fetch_texel_chan_to_float(const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texelOut)
+{
+   GLchan temp[4];
+   GLenum baseFormat = _mesa_get_format_base_format(texImage->TexFormat);
+
+   ASSERT(texImage->FetchTexelc);
+   texImage->FetchTexelc(texImage, i, j, k, temp);
+   if (baseFormat == GL_DEPTH_COMPONENT ||
+       baseFormat == GL_DEPTH_STENCIL_EXT) {
+      /* just one channel */
+      texelOut[0] = CHAN_TO_FLOAT(temp[0]);
+   }
+   else {
+      /* four channels */
+      texelOut[0] = CHAN_TO_FLOAT(temp[0]);
+      texelOut[1] = CHAN_TO_FLOAT(temp[1]);
+      texelOut[2] = CHAN_TO_FLOAT(temp[2]);
+      texelOut[3] = CHAN_TO_FLOAT(temp[3]);
+   }
+}
+#endif
+
+
+/**
+ * Initialize the texture image's FetchTexelc and FetchTexelf methods.
+ */
+void
+_mesa_set_fetch_functions(struct gl_texture_image *texImage, GLuint dims)
+{
+   gl_format format = texImage->TexFormat;
+
+   ASSERT(dims == 1 || dims == 2 || dims == 3);
+
+   if (texImage->TexObject->Sampler.sRGBDecode == GL_SKIP_DECODE_EXT &&
+       _mesa_get_format_color_encoding(format) == GL_SRGB) {
+      format = _mesa_get_srgb_format_linear(format);
+   }
+
+   texImage->FetchTexelf = _mesa_get_texel_fetch_func(format, dims);
+
+   texImage->FetchTexelc = fetch_texel_float_to_chan;
+
+   ASSERT(texImage->FetchTexelc);
+   ASSERT(texImage->FetchTexelf);
+}
+
+void
+_mesa_update_fetch_functions(struct gl_texture_object *texObj)
+{
+   GLuint face, i;
+   GLuint dims;
+
+   dims = _mesa_get_texture_dimensions(texObj->Target);
+
+   for (face = 0; face < 6; face++) {
+      for (i = 0; i < MAX_TEXTURE_LEVELS; i++) {
+         if (texObj->Image[face][i]) {
+	    _mesa_set_fetch_functions(texObj->Image[face][i], dims);
+         }
+      }
+   }
+}
diff --git a/mesalib/src/mesa/main/texfetch_tmp.h b/mesalib/src/mesa/main/texfetch_tmp.h
index 278becc11..3b1eedf39 100644
--- a/mesalib/src/mesa/main/texfetch_tmp.h
+++ b/mesalib/src/mesa/main/texfetch_tmp.h
@@ -1,2379 +1,2402 @@
-/*

- * Mesa 3-D graphics library

- * Version:  7.7

- *

- * 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.

- */

-

-

-/**

- * \file texfetch_tmp.h

- * Texel fetch functions template.

- * 

- * This template file is used by texfetch.c to generate texel fetch functions

- * for 1-D, 2-D and 3-D texture images. 

- *

- * It should be expanded by defining \p DIM as the number texture dimensions

- * (1, 2 or 3).  According to the value of \p DIM a series of macros is defined

- * for the texel lookup in the gl_texture_image::Data.

- * 

- * \author Gareth Hughes

- * \author Brian Paul

- */

-

-

-#if DIM == 1

-

-#define TEXEL_ADDR( type, image, i, j, k, size ) \

-	((void) (j), (void) (k), ((type *)(image)->Data + (i) * (size)))

-

-#define FETCH(x) fetch_texel_1d_##x

-

-#elif DIM == 2

-

-#define TEXEL_ADDR( type, image, i, j, k, size )			\

-	((void) (k),							\

-	 ((type *)(image)->Data + ((image)->RowStride * (j) + (i)) * (size)))

-

-#define FETCH(x) fetch_texel_2d_##x

-

-#elif DIM == 3

-

-#define TEXEL_ADDR( type, image, i, j, k, size )			\

-	((type *)(image)->Data + ((image)->ImageOffsets[k]		\

-             + (image)->RowStride * (j) + (i)) * (size))

-

-#define FETCH(x) fetch_texel_3d_##x

-

-#else

-#error	illegal number of texture dimensions

-#endif

-

-

-/* MESA_FORMAT_Z32 ***********************************************************/

-

-/* Fetch depth texel from 1D, 2D or 3D 32-bit depth texture,

- * returning 1 GLfloat.

- * Note: no GLchan version of this function.

- */

-static void FETCH(f_z32)( const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[0] = src[0] * (1.0F / 0xffffffff);

-}

-

-#if DIM == 3

-static void store_texel_z32(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLuint *depth = (const GLuint *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   dst[0] = *depth;

-}

-#endif

-

-

-/* MESA_FORMAT_Z16 ***********************************************************/

-

-/* Fetch depth texel from 1D, 2D or 3D 16-bit depth texture,

- * returning 1 GLfloat.

- * Note: no GLchan version of this function.

- */

-static void FETCH(f_z16)(const struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[0] = src[0] * (1.0F / 65535.0F);

-}

-

-#if DIM == 3

-static void store_texel_z16(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *depth = (const GLushort *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   dst[0] = *depth;

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_F32 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D RGBA_FLOAT32 texture, returning 4 GLfloats.

- */

-static void FETCH(f_rgba_f32)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 4);

-   texel[RCOMP] = src[0];

-   texel[GCOMP] = src[1];

-   texel[BCOMP] = src[2];

-   texel[ACOMP] = src[3];

-}

-

-#if DIM == 3

-static void store_texel_rgba_f32(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *depth = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 4);

-   dst[0] = depth[RCOMP];

-   dst[1] = depth[GCOMP];

-   dst[2] = depth[BCOMP];

-   dst[3] = depth[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_F16 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D RGBA_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_rgba_f16)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 4);

-   texel[RCOMP] = _mesa_half_to_float(src[0]);

-   texel[GCOMP] = _mesa_half_to_float(src[1]);

-   texel[BCOMP] = _mesa_half_to_float(src[2]);

-   texel[ACOMP] = _mesa_half_to_float(src[3]);

-}

-

-#if DIM == 3

-static void store_texel_rgba_f16(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *src = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 4);

-   dst[0] = _mesa_float_to_half(src[RCOMP]);

-   dst[1] = _mesa_float_to_half(src[GCOMP]);

-   dst[2] = _mesa_float_to_half(src[BCOMP]);

-   dst[3] = _mesa_float_to_half(src[ACOMP]);

-}

-#endif

-

-/* MESA_FORMAT_RGB_F32 *******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D RGB_FLOAT32 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_rgb_f32)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 3);

-   texel[RCOMP] = src[0];

-   texel[GCOMP] = src[1];

-   texel[BCOMP] = src[2];

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgb_f32(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *src = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 3);

-   dst[0] = src[RCOMP];

-   dst[1] = src[GCOMP];

-   dst[2] = src[BCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGB_F16 *******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D RGB_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_rgb_f16)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 3);

-   texel[RCOMP] = _mesa_half_to_float(src[0]);

-   texel[GCOMP] = _mesa_half_to_float(src[1]);

-   texel[BCOMP] = _mesa_half_to_float(src[2]);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgb_f16(struct gl_texture_image *texImage,

-                                GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *src = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 3);

-   dst[0] = _mesa_float_to_half(src[RCOMP]);

-   dst[1] = _mesa_float_to_half(src[GCOMP]);

-   dst[2] = _mesa_float_to_half(src[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_ALPHA_F32 *****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D ALPHA_FLOAT32 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_alpha_f32)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = src[0];

-}

-

-#if DIM == 3

-static void store_texel_alpha_f32(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   dst[0] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_ALPHA_F32 *****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D ALPHA_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_alpha_f16)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = _mesa_half_to_float(src[0]);

-}

-

-#if DIM == 3

-static void store_texel_alpha_f16(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   dst[0] = _mesa_float_to_half(rgba[ACOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_LUMINANCE_F32 *************************************************/

-

-/* Fetch texel from 1D, 2D or 3D LUMINANCE_FLOAT32 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_luminance_f32)( const struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = src[0];

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_luminance_f32(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   dst[0] = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_LUMINANCE_F16 *************************************************/

-

-/* Fetch texel from 1D, 2D or 3D LUMINANCE_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_luminance_f16)( const struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = _mesa_half_to_float(src[0]);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_luminance_f16(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   dst[0] = _mesa_float_to_half(rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_LUMINANCE_ALPHA_F32 *******************************************/

-

-/* Fetch texel from 1D, 2D or 3D LUMINANCE_ALPHA_FLOAT32 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_luminance_alpha_f32)( const struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = src[0];

-   texel[ACOMP] = src[1];

-}

-

-#if DIM == 3

-static void store_texel_luminance_alpha_f32(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_LUMINANCE_ALPHA_F16 *******************************************/

-

-/* Fetch texel from 1D, 2D or 3D LUMINANCE_ALPHA_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_luminance_alpha_f16)( const struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = _mesa_half_to_float(src[0]);

-   texel[ACOMP] = _mesa_half_to_float(src[1]);

-}

-

-#if DIM == 3

-static void store_texel_luminance_alpha_f16(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);

-   dst[0] = _mesa_float_to_half(rgba[RCOMP]);

-   dst[1] = _mesa_float_to_half(rgba[ACOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_INTENSITY_F32 *************************************************/

-

-/* Fetch texel from 1D, 2D or 3D INTENSITY_FLOAT32 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_intensity_f32)( const struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] =

-   texel[ACOMP] = src[0];

-}

-

-#if DIM == 3

-static void store_texel_intensity_f32(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   dst[0] = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_INTENSITY_F16 *************************************************/

-

-/* Fetch texel from 1D, 2D or 3D INTENSITY_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_intensity_f16)( const struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] =

-   texel[ACOMP] = _mesa_half_to_float(src[0]);

-}

-

-#if DIM == 3

-static void store_texel_intensity_f16(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   dst[0] = _mesa_float_to_half(rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_R_FLOAT32 *****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D R_FLOAT32 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_r_f32)( const struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   texel[RCOMP] = src[0];

-   texel[GCOMP] = 0.0F;

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_r_f32(struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);

-   dst[0] = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_R_FLOAT16 *****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D R_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_r_f16)( const struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   texel[RCOMP] = _mesa_half_to_float(src[0]);

-   texel[GCOMP] = 0.0F;

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_r_f16(struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);

-   dst[0] = _mesa_float_to_half(rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RG_FLOAT32 ****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D RG_FLOAT32 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_rg_f32)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);

-   texel[RCOMP] = src[0];

-   texel[GCOMP] = src[1];

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rg_f32(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RG_FLOAT16 ****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D RG_FLOAT16 texture,

- * returning 4 GLfloats.

- */

-static void FETCH(f_rg_f16)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);

-   texel[RCOMP] = _mesa_half_to_float(src[0]);

-   texel[GCOMP] = _mesa_half_to_float(src[1]);

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rg_f16(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *rgba = (const GLfloat *) texel;

-   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);

-   dst[0] = _mesa_float_to_half(rgba[RCOMP]);

-   dst[1] = _mesa_float_to_half(rgba[GCOMP]);

-}

-#endif

-

-

-/*

- * Begin Hardware formats

- */

-

-/* MESA_FORMAT_RGBA8888 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rgba8888 texture, return 4 GLfloats */

-static void FETCH(f_rgba8888)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );

-   texel[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );

-   texel[BCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );

-   texel[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );

-}

-

-

-

-#if DIM == 3

-static void store_texel_rgba8888(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA888_REV ***************************************************/

-

-/* Fetch texel from 1D, 2D or 3D abgr8888 texture, return 4 GLchans */

-static void FETCH(f_rgba8888_rev)( const struct gl_texture_image *texImage,

-                                   GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );

-   texel[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );

-   texel[BCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );

-   texel[ACOMP] = UBYTE_TO_FLOAT( (s >> 24)        );

-}

-

-#if DIM == 3

-static void store_texel_rgba8888_rev(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888_REV(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_ARGB8888 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb8888 texture, return 4 GLchans */

-static void FETCH(f_argb8888)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );

-   texel[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );

-   texel[BCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );

-   texel[ACOMP] = UBYTE_TO_FLOAT( (s >> 24)        );

-}

-

-#if DIM == 3

-static void store_texel_argb8888(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_ARGB8888_REV **************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb8888_rev texture, return 4 GLfloats */

-static void FETCH(f_argb8888_rev)( const struct gl_texture_image *texImage,

-                                   GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );

-   texel[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );

-   texel[BCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );

-   texel[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );

-}

-

-#if DIM == 3

-static void store_texel_argb8888_rev(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[BCOMP], rgba[GCOMP], rgba[RCOMP], rgba[ACOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_XRGB8888 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D xrgb8888 texture, return 4 GLchans */

-static void FETCH(f_xrgb8888)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );

-   texel[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );

-   texel[BCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );

-   texel[ACOMP] = 1.0f;

-}

-

-#if DIM == 3

-static void store_texel_xrgb8888(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(0xff, rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_XRGB8888_REV **************************************************/

-

-/* Fetch texel from 1D, 2D or 3D xrgb8888_rev texture, return 4 GLfloats */

-static void FETCH(f_xrgb8888_rev)( const struct gl_texture_image *texImage,

-                                   GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );

-   texel[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );

-   texel[BCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );

-   texel[ACOMP] = 1.0f;

-}

-

-#if DIM == 3

-static void store_texel_xrgb8888_rev(struct gl_texture_image *texImage,

-                                     GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[BCOMP], rgba[GCOMP], rgba[RCOMP], 0xff);

-}

-#endif

-

-

-/* MESA_FORMAT_RGB888 ********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rgb888 texture, return 4 GLchans */

-static void FETCH(f_rgb888)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);

-   texel[RCOMP] = UBYTE_TO_FLOAT( src[2] );

-   texel[GCOMP] = UBYTE_TO_FLOAT( src[1] );

-   texel[BCOMP] = UBYTE_TO_FLOAT( src[0] );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgb888(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);

-   dst[0] = rgba[BCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_BGR888 ********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D bgr888 texture, return 4 GLchans */

-static void FETCH(f_bgr888)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);

-   texel[RCOMP] = UBYTE_TO_FLOAT( src[0] );

-   texel[GCOMP] = UBYTE_TO_FLOAT( src[1] );

-   texel[BCOMP] = UBYTE_TO_FLOAT( src[2] );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_bgr888(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-}

-#endif

-

-

-/* use color expansion like (g << 2) | (g >> 4) (does somewhat random rounding)

-   instead of slow (g << 2) * 255 / 252 (always rounds down) */

-

-/* MESA_FORMAT_RGB565 ********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rgb565 texture, return 4 GLchans */

-static void FETCH(f_rgb565)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   const GLushort s = *src;

-   texel[RCOMP] = ((s >> 11) & 0x1f) * (1.0F / 31.0F);

-   texel[GCOMP] = ((s >> 5 ) & 0x3f) * (1.0F / 63.0F);

-   texel[BCOMP] = ((s      ) & 0x1f) * (1.0F / 31.0F);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgb565(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_565(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RGB565_REV ****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rgb565_rev texture, return 4 GLchans */

-static void FETCH(f_rgb565_rev)( const struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   const GLushort s = (*src >> 8) | (*src << 8); /* byte swap */

-   texel[RCOMP] = UBYTE_TO_FLOAT( ((s >> 8) & 0xf8) | ((s >> 13) & 0x7) );

-   texel[GCOMP] = UBYTE_TO_FLOAT( ((s >> 3) & 0xfc) | ((s >>  9) & 0x3) );

-   texel[BCOMP] = UBYTE_TO_FLOAT( ((s << 3) & 0xf8) | ((s >>  2) & 0x7) );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgb565_rev(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_565(rgba[BCOMP], rgba[GCOMP], rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_ARGB4444 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb444 texture, return 4 GLchans */

-static void FETCH(f_argb4444)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   const GLushort s = *src;

-   texel[RCOMP] = ((s >>  8) & 0xf) * (1.0F / 15.0F);

-   texel[GCOMP] = ((s >>  4) & 0xf) * (1.0F / 15.0F);

-   texel[BCOMP] = ((s      ) & 0xf) * (1.0F / 15.0F);

-   texel[ACOMP] = ((s >> 12) & 0xf) * (1.0F / 15.0F);

-}

-

-#if DIM == 3

-static void store_texel_argb4444(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_4444(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_ARGB4444_REV **************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb4444_rev texture, return 4 GLchans */

-static void FETCH(f_argb4444_rev)( const struct gl_texture_image *texImage,

-                                   GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] = ((s      ) & 0xf) * (1.0F / 15.0F);

-   texel[GCOMP] = ((s >> 12) & 0xf) * (1.0F / 15.0F);

-   texel[BCOMP] = ((s >>  8) & 0xf) * (1.0F / 15.0F);

-   texel[ACOMP] = ((s >>  4) & 0xf) * (1.0F / 15.0F);

-}

-

-#if DIM == 3

-static void store_texel_argb4444_rev(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_4444(rgba[ACOMP], rgba[BCOMP], rgba[GCOMP], rgba[RCOMP]);

-}

-#endif

-

-/* MESA_FORMAT_RGBA5551 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb1555 texture, return 4 GLchans */

-static void FETCH(f_rgba5551)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   const GLushort s = *src;

-   texel[RCOMP] = ((s >> 11) & 0x1f) * (1.0F / 31.0F);

-   texel[GCOMP] = ((s >>  6) & 0x1f) * (1.0F / 31.0F);

-   texel[BCOMP] = ((s >>  1) & 0x1f) * (1.0F / 31.0F);

-   texel[ACOMP] = ((s      ) & 0x01) * 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgba5551(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_5551(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);

-}

-#endif

-

-/* MESA_FORMAT_ARGB1555 ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb1555 texture, return 4 GLchans */

-static void FETCH(f_argb1555)( const struct gl_texture_image *texImage,

-			     GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   const GLushort s = *src;

-   texel[RCOMP] = ((s >> 10) & 0x1f) * (1.0F / 31.0F);

-   texel[GCOMP] = ((s >>  5) & 0x1f) * (1.0F / 31.0F);

-   texel[BCOMP] = ((s >>  0) & 0x1f) * (1.0F / 31.0F);

-   texel[ACOMP] = ((s >> 15) & 0x01) * 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_argb1555(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_1555(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_ARGB1555_REV **************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb1555_rev texture, return 4 GLchans */

-static void FETCH(f_argb1555_rev)( const struct gl_texture_image *texImage,

-                                   GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   const GLushort s = (*src << 8) | (*src >> 8); /* byteswap */

-   texel[RCOMP] = UBYTE_TO_FLOAT( ((s >>  7) & 0xf8) | ((s >> 12) & 0x7) );

-   texel[GCOMP] = UBYTE_TO_FLOAT( ((s >>  2) & 0xf8) | ((s >>  7) & 0x7) );

-   texel[BCOMP] = UBYTE_TO_FLOAT( ((s <<  3) & 0xf8) | ((s >>  2) & 0x7) );

-   texel[ACOMP] = UBYTE_TO_FLOAT( ((s >> 15) & 0x01) * 255 );

-}

-

-#if DIM == 3

-static void store_texel_argb1555_rev(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_1555_REV(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_ARGB2101010 ***************************************************/

-

-/* Fetch texel from 1D, 2D or 3D argb2101010 texture, return 4 GLchans */

-static void FETCH(f_argb2101010)( const struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   const GLuint s = *src;

-   texel[RCOMP] = ((s >> 20) & 0x3ff) * (1.0F / 1023.0F);

-   texel[GCOMP] = ((s >> 10) & 0x3ff) * (1.0F / 1023.0F);

-   texel[BCOMP] = ((s >>  0) & 0x3ff) * (1.0F / 1023.0F);

-   texel[ACOMP] = ((s >> 30) & 0x03) * (1.0F / 3.0F);

-}

-

-#if DIM == 3

-static void store_texel_argb2101010(struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_2101010_UB(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RG88 **********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rg88 texture, return 4 GLchans */

-static void FETCH(f_rg88)( const struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( s & 0xff );

-   texel[GCOMP] = UBYTE_TO_FLOAT( s >> 8 );

-   texel[BCOMP] = 0.0;

-   texel[ACOMP] = 1.0;

-}

-

-#if DIM == 3

-static void store_texel_rg88(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_88(rgba[RCOMP], rgba[GCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RG88_REV ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rg88_rev texture, return 4 GLchans */

-static void FETCH(f_rg88_rev)( const struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT( s & 0xff );

-   texel[GCOMP] = UBYTE_TO_FLOAT( s >> 8 );

-   texel[BCOMP] = 0.0;

-   texel[ACOMP] = 1.0;

-}

-

-#if DIM == 3

-static void store_texel_rg88_rev(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_88(rgba[GCOMP], rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_AL44 **********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D al44 texture, return 4 GLchans */

-static void FETCH(f_al44)( const struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte s = *TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = (s & 0xf) * (1.0F / 15.0F);

-   texel[ACOMP] = ((s >> 4) & 0xf) * (1.0F / 15.0F);

-}

-

-#if DIM == 3

-static void store_texel_al44(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_44(rgba[ACOMP], rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_AL88 **********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D al88 texture, return 4 GLchans */

-static void FETCH(f_al88)( const struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] = 

-   texel[GCOMP] = 

-   texel[BCOMP] = UBYTE_TO_FLOAT( s & 0xff );

-   texel[ACOMP] = UBYTE_TO_FLOAT( s >> 8 );

-}

-

-#if DIM == 3

-static void store_texel_al88(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_88(rgba[ACOMP], rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_R8 ************************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rg88 texture, return 4 GLchans */

-static void FETCH(f_r8)(const struct gl_texture_image *texImage,

-			GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLubyte s = *TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   texel[RCOMP] = UBYTE_TO_FLOAT(s);

-   texel[GCOMP] = 0.0;

-   texel[BCOMP] = 0.0;

-   texel[ACOMP] = 1.0;

-}

-

-#if DIM == 3

-static void store_texel_r8(struct gl_texture_image *texImage,

-			   GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_R16 ***********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D r16 texture, return 4 GLchans */

-static void FETCH(f_r16)(const struct gl_texture_image *texImage,

-			GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] = USHORT_TO_FLOAT(s);

-   texel[GCOMP] = 0.0;

-   texel[BCOMP] = 0.0;

-   texel[ACOMP] = 1.0;

-}

-

-#if DIM == 3

-static void store_texel_r16(struct gl_texture_image *texImage,

-			    GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_AL88_REV ******************************************************/

-

-/* Fetch texel from 1D, 2D or 3D al88_rev texture, return 4 GLchans */

-static void FETCH(f_al88_rev)( const struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] = 

-   texel[GCOMP] = 

-   texel[BCOMP] = UBYTE_TO_FLOAT( s >> 8 );

-   texel[ACOMP] = UBYTE_TO_FLOAT( s & 0xff );

-}

-

-#if DIM == 3

-static void store_texel_al88_rev(struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_88(rgba[RCOMP], rgba[ACOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RG1616 ********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rg1616 texture, return 4 GLchans */

-static void FETCH(f_rg1616)( const struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = USHORT_TO_FLOAT( s & 0xffff );

-   texel[GCOMP] = USHORT_TO_FLOAT( s >> 16 );

-   texel[BCOMP] = 0.0;

-   texel[ACOMP] = 1.0;

-}

-

-#if DIM == 3

-static void store_texel_rg1616(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_1616(rgba[RCOMP], rgba[GCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RG1616_REV ****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rg1616_rev texture, return 4 GLchans */

-static void FETCH(f_rg1616_rev)( const struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = USHORT_TO_FLOAT( s >> 16 );

-   texel[GCOMP] = USHORT_TO_FLOAT( s & 0xffff );

-   texel[BCOMP] = 0.0;

-   texel[ACOMP] = 1.0;

-}

-

-#if DIM == 3

-static void store_texel_rg1616_rev(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_1616(rgba[GCOMP], rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_AL1616 ********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D al1616 texture, return 4 GLchans */

-static void FETCH(f_al1616)( const struct gl_texture_image *texImage,

-			     GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = USHORT_TO_FLOAT( s & 0xffff );

-   texel[ACOMP] = USHORT_TO_FLOAT( s >> 16 );

-}

-

-#if DIM == 3

-static void store_texel_al1616(struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_1616(rgba[ACOMP], rgba[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_AL1616_REV ****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D al1616_rev texture, return 4 GLchans */

-static void FETCH(f_al1616_rev)( const struct gl_texture_image *texImage,

-				 GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = USHORT_TO_FLOAT( s >> 16 );

-   texel[ACOMP] = USHORT_TO_FLOAT( s & 0xffff );

-}

-

-#if DIM == 3

-static void store_texel_al1616_rev(struct gl_texture_image *texImage,

-				   GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_1616(rgba[RCOMP], rgba[ACOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_RGB332 ********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D rgb332 texture, return 4 GLchans */

-static void FETCH(f_rgb332)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   const GLubyte s = *src;

-   texel[RCOMP] = ((s >> 5) & 0x7) * (1.0F / 7.0F);

-   texel[GCOMP] = ((s >> 2) & 0x7) * (1.0F / 7.0F);

-   texel[BCOMP] = ((s     ) & 0x3) * (1.0F / 3.0F);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgb332(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_332(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_A8 ************************************************************/

-

-/* Fetch texel from 1D, 2D or 3D a8 texture, return 4 GLchans */

-static void FETCH(f_a8)( const struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = UBYTE_TO_FLOAT( src[0] );

-}

-

-#if DIM == 3

-static void store_texel_a8(struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   *dst = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_A16 ************************************************************/

-

-/* Fetch texel from 1D, 2D or 3D a8 texture, return 4 GLchans */

-static void FETCH(f_a16)( const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = USHORT_TO_FLOAT( src[0] );

-}

-

-#if DIM == 3

-static void store_texel_a16(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_L8 ************************************************************/

-

-/* Fetch texel from 1D, 2D or 3D l8 texture, return 4 GLchans */

-static void FETCH(f_l8)( const struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = UBYTE_TO_FLOAT( src[0] );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_l8(struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_L16 ***********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D l16 texture, return 4 GLchans */

-static void FETCH(f_l16)( const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = USHORT_TO_FLOAT( src[0] );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_l16(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_I8 ************************************************************/

-

-/* Fetch texel from 1D, 2D or 3D i8 texture, return 4 GLchans */

-static void FETCH(f_i8)( const struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] =

-   texel[ACOMP] = UBYTE_TO_FLOAT( src[0] );

-}

-

-#if DIM == 3

-static void store_texel_i8(struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_I16 ***********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D i16 texture, return 4 GLchans */

-static void FETCH(f_i16)( const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] =

-   texel[ACOMP] = USHORT_TO_FLOAT( src[0] );

-}

-

-#if DIM == 3

-static void store_texel_i16(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_CI8 ***********************************************************/

-

-/* Fetch CI texel from 1D, 2D or 3D ci8 texture, lookup the index in a

- * color table, and return 4 GLchans.

- */

-static void FETCH(f_ci8)( const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   const struct gl_color_table *palette;

-   GLuint index;

-   GET_CURRENT_CONTEXT(ctx);

-

-   if (ctx->Texture.SharedPalette) {

-      palette = &ctx->Texture.Palette;

-   }

-   else {

-      palette = &texImage->TexObject->Palette;

-   }

-   if (palette->Size == 0)

-      return; /* undefined results */

-

-   /* Mask the index against size of palette to avoid going out of bounds */

-   index = (*src) & (palette->Size - 1);

-

-   {

-      const GLfloat *table = palette->TableF;

-      switch (palette->_BaseFormat) {

-      case GL_ALPHA:

-         texel[RCOMP] =

-         texel[GCOMP] =

-         texel[BCOMP] = 0.0F;

-         texel[ACOMP] = table[index];

-         break;

-      case GL_LUMINANCE:

-         texel[RCOMP] =

-         texel[GCOMP] =

-         texel[BCOMP] = table[index];

-         texel[ACOMP] = 1.0F;

-         break;

-      case GL_INTENSITY:

-         texel[RCOMP] =

-         texel[GCOMP] =

-         texel[BCOMP] =

-         texel[ACOMP] = table[index];

-         break;

-      case GL_LUMINANCE_ALPHA:

-         texel[RCOMP] =

-         texel[GCOMP] =

-         texel[BCOMP] = table[index * 2 + 0];

-         texel[ACOMP] = table[index * 2 + 1];

-         break;

-      case GL_RGB:

-         texel[RCOMP] = table[index * 3 + 0];

-         texel[GCOMP] = table[index * 3 + 1];

-         texel[BCOMP] = table[index * 3 + 2];

-         texel[ACOMP] = 1.0F;

-         break;

-      case GL_RGBA:

-         texel[RCOMP] = table[index * 4 + 0];

-         texel[GCOMP] = table[index * 4 + 1];

-         texel[BCOMP] = table[index * 4 + 2];

-         texel[ACOMP] = table[index * 4 + 3];

-         break;

-      default:

-         _mesa_problem(ctx, "Bad palette format in fetch_texel_ci8");

-         return;

-      }

-   }

-}

-

-#if DIM == 3

-static void store_texel_ci8(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *index = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   *dst = *index;

-}

-#endif

-

-

-/* Fetch texel from 1D, 2D or 3D srgb8 texture, return 4 GLfloats */

-/* Note: component order is same as for MESA_FORMAT_RGB888 */

-static void FETCH(srgb8)(const struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);

-   texel[RCOMP] = nonlinear_to_linear(src[2]);

-   texel[GCOMP] = nonlinear_to_linear(src[1]);

-   texel[BCOMP] = nonlinear_to_linear(src[0]);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_srgb8(struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);

-   dst[0] = rgba[BCOMP]; /* no conversion */

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[RCOMP];

-}

-#endif

-

-/* Fetch texel from 1D, 2D or 3D srgba8 texture, return 4 GLfloats */

-static void FETCH(srgba8)(const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = nonlinear_to_linear( (s >> 24) );

-   texel[GCOMP] = nonlinear_to_linear( (s >> 16) & 0xff );

-   texel[BCOMP] = nonlinear_to_linear( (s >>  8) & 0xff );

-   texel[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff ); /* linear! */

-}

-

-#if DIM == 3

-static void store_texel_srgba8(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);

-}

-#endif

-

-/* Fetch texel from 1D, 2D or 3D sargb8 texture, return 4 GLfloats */

-static void FETCH(sargb8)(const struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = nonlinear_to_linear( (s >> 16) & 0xff );

-   texel[GCOMP] = nonlinear_to_linear( (s >>  8) & 0xff );

-   texel[BCOMP] = nonlinear_to_linear( (s      ) & 0xff );

-   texel[ACOMP] = UBYTE_TO_FLOAT( (s >> 24) ); /* linear! */

-}

-

-#if DIM == 3

-static void store_texel_sargb8(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);

-}

-#endif

-

-/* Fetch texel from 1D, 2D or 3D sl8 texture, return 4 GLfloats */

-static void FETCH(sl8)(const struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   texel[RCOMP] = 

-   texel[GCOMP] = 

-   texel[BCOMP] = nonlinear_to_linear(src[0]);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_sl8(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);

-   dst[0] = rgba[RCOMP];

-}

-#endif

-

-/* Fetch texel from 1D, 2D or 3D sla8 texture, return 4 GLfloats */

-static void FETCH(sla8)(const struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 2);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = nonlinear_to_linear(src[0]);

-   texel[ACOMP] = UBYTE_TO_FLOAT(src[1]); /* linear */

-}

-

-#if DIM == 3

-static void store_texel_sla8(struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 2);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_INT8 **************************************************/

-

-static void

-FETCH(rgba_int8)(const struct gl_texture_image *texImage,

-                 GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLbyte *src = TEXEL_ADDR(GLbyte, texImage, i, j, k, 4);

-   texel[RCOMP] = (GLfloat) src[0];

-   texel[GCOMP] = (GLfloat) src[1];

-   texel[BCOMP] = (GLfloat) src[2];

-   texel[ACOMP] = (GLfloat) src[3];

-}

-

-#if DIM == 3

-static void

-store_texel_rgba_int8(struct gl_texture_image *texImage,

-                      GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rgba = (const GLbyte *) texel;

-   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_INT16 **************************************************/

-

-static void

-FETCH(rgba_int16)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLshort *src = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);

-   texel[RCOMP] = (GLfloat) src[0];

-   texel[GCOMP] = (GLfloat) src[1];

-   texel[BCOMP] = (GLfloat) src[2];

-   texel[ACOMP] = (GLfloat) src[3];

-}

-

-#if DIM == 3

-static void

-store_texel_rgba_int16(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_INT32 **************************************************/

-

-static void

-FETCH(rgba_int32)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLint *src = TEXEL_ADDR(GLint, texImage, i, j, k, 4);

-   texel[RCOMP] = (GLfloat) src[0];

-   texel[GCOMP] = (GLfloat) src[1];

-   texel[BCOMP] = (GLfloat) src[2];

-   texel[ACOMP] = (GLfloat) src[3];

-}

-

-#if DIM == 3

-static void

-store_texel_rgba_int32(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLint *rgba = (const GLint *) texel;

-   GLint *dst = TEXEL_ADDR(GLint, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_UINT8 **************************************************/

-

-static void

-FETCH(rgba_uint8)(const struct gl_texture_image *texImage,

-                 GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 4);

-   texel[RCOMP] = (GLfloat) src[0];

-   texel[GCOMP] = (GLfloat) src[1];

-   texel[BCOMP] = (GLfloat) src[2];

-   texel[ACOMP] = (GLfloat) src[3];

-}

-

-#if DIM == 3

-static void

-store_texel_rgba_uint8(struct gl_texture_image *texImage,

-                      GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_UINT16 **************************************************/

-

-static void

-FETCH(rgba_uint16)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);

-   texel[RCOMP] = (GLfloat) src[0];

-   texel[GCOMP] = (GLfloat) src[1];

-   texel[BCOMP] = (GLfloat) src[2];

-   texel[ACOMP] = (GLfloat) src[3];

-}

-

-#if DIM == 3

-static void

-store_texel_rgba_uint16(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_RGBA_UINT32 **************************************************/

-

-static void

-FETCH(rgba_uint32)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 4);

-   texel[RCOMP] = (GLfloat) src[0];

-   texel[GCOMP] = (GLfloat) src[1];

-   texel[BCOMP] = (GLfloat) src[2];

-   texel[ACOMP] = (GLfloat) src[3];

-}

-

-#if DIM == 3

-static void

-store_texel_rgba_uint32(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLuint *rgba = (const GLuint *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_DUDV8 ********************************************************/

-

-/* this format by definition produces 0,0,0,1 as rgba values,

-   however we'll return the dudv values as rg and fix up elsewhere */

-static void FETCH(dudv8)(const struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLbyte *src = TEXEL_ADDR(GLbyte, texImage, i, j, k, 2);

-   texel[RCOMP] = BYTE_TO_FLOAT(src[0]);

-   texel[GCOMP] = BYTE_TO_FLOAT(src[1]);

-   texel[BCOMP] = 0;

-   texel[ACOMP] = 0;

-}

-

-

-/* MESA_FORMAT_SIGNED_R8 ***********************************************/

-

-static void FETCH(signed_r8)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   texel[RCOMP] = BYTE_TO_FLOAT_TEX( s );

-   texel[GCOMP] = 0.0F;

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_signed_r8(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rgba = (const GLbyte *) texel;

-   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_A8 ***********************************************/

-

-static void FETCH(signed_a8)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   texel[RCOMP] = 0.0F;

-   texel[GCOMP] = 0.0F;

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = BYTE_TO_FLOAT_TEX( s );

-}

-

-#if DIM == 3

-static void store_texel_signed_a8(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rgba = (const GLbyte *) texel;

-   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   *dst = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_L8 ***********************************************/

-

-static void FETCH(signed_l8)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = BYTE_TO_FLOAT_TEX( s );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_signed_l8(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rgba = (const GLbyte *) texel;

-   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_I8 ***********************************************/

-

-static void FETCH(signed_i8)( const struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] =

-   texel[ACOMP] = BYTE_TO_FLOAT_TEX( s );

-}

-

-#if DIM == 3

-static void store_texel_signed_i8(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rgba = (const GLbyte *) texel;

-   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_RG88_REV ***********************************************/

-

-static void FETCH(signed_rg88_rev)( const struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s & 0xff) );

-   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 8) );

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_signed_rg88_rev(struct gl_texture_image *texImage,

-                                        GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rg = (const GLbyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   dst[0] = PACK_COLOR_88(rg[GCOMP], rg[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_AL88 ***********************************************/

-

-static void FETCH(signed_al88)( const struct gl_texture_image *texImage,

-                                GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s & 0xff) );

-   texel[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 8) );

-}

-

-#if DIM == 3

-static void store_texel_signed_al88(struct gl_texture_image *texImage,

-                                    GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rg = (const GLbyte *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);

-   dst[0] = PACK_COLOR_88(rg[ACOMP], rg[RCOMP]);

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_RGBX8888 ***********************************************/

-

-static void FETCH(signed_rgbx8888)( const struct gl_texture_image *texImage,

-			            GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );

-   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );

-   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );

-   texel[ACOMP] = 1.0f;

-}

-

-#if DIM == 3

-static void store_texel_signed_rgbx8888(struct gl_texture_image *texImage,

-                                        GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rgba = (const GLbyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], 255);

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_RGBA8888 ***********************************************/

-

-static void FETCH(signed_rgba8888)( const struct gl_texture_image *texImage,

-			            GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );

-   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );

-   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );

-   texel[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s      ) );

-}

-

-#if DIM == 3

-static void store_texel_signed_rgba8888(struct gl_texture_image *texImage,

-                                        GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLbyte *rgba = (const GLbyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);

-}

-#endif

-

-static void FETCH(signed_rgba8888_rev)( const struct gl_texture_image *texImage,

-                                        GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s      ) );

-   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );

-   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );

-   texel[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );

-}

-

-#if DIM == 3

-static void store_texel_signed_rgba8888_rev(struct gl_texture_image *texImage,

-                                            GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLubyte *rgba = (const GLubyte *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = PACK_COLOR_8888_REV(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);

-}

-#endif

-

-

-

-/* MESA_FORMAT_SIGNED_R16 ***********************************************/

-

-static void

-FETCH(signed_r16)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s );

-   texel[GCOMP] = 0.0F;

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void

-store_texel_signed_r16(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   *dst = rgba[0];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_A16 ***********************************************/

-

-static void

-FETCH(signed_a16)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   texel[RCOMP] = 0.0F;

-   texel[GCOMP] = 0.0F;

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s );

-}

-

-#if DIM == 3

-static void

-store_texel_signed_a16(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   *dst = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_L16 ***********************************************/

-

-static void

-FETCH(signed_l16)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void

-store_texel_signed_l16(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_I16 ***********************************************/

-

-static void

-FETCH(signed_i16)(const struct gl_texture_image *texImage,

-                  GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] =

-   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s );

-}

-

-#if DIM == 3

-static void

-store_texel_signed_i16(struct gl_texture_image *texImage,

-                       GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);

-   *dst = rgba[RCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_RG1616 ***********************************************/

-

-static void

-FETCH(signed_rg1616)(const struct gl_texture_image *texImage,

-                    GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);

-   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );

-   texel[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );

-   texel[BCOMP] = 0.0F;

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void

-store_texel_signed_rg1616(struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_AL1616 ***********************************************/

-

-static void

-FETCH(signed_al1616)(const struct gl_texture_image *texImage,

-                    GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);

-   texel[RCOMP] =

-   texel[GCOMP] =

-   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s[0] );

-   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s[1] );

-}

-

-#if DIM == 3

-static void

-store_texel_signed_al1616(struct gl_texture_image *texImage,

-                         GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[ACOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_RGB_16 ***********************************************/

-

-static void 

-FETCH(signed_rgb_16)(const struct gl_texture_image *texImage,

-                     GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 3);

-   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );

-   texel[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );

-   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s[2] );

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void

-store_texel_signed_rgb_16(struct gl_texture_image *texImage,

-                          GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 3);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-}

-#endif

-

-

-/* MESA_FORMAT_SIGNED_RGBA_16 ***********************************************/

-

-static void

-FETCH(signed_rgba_16)(const struct gl_texture_image *texImage,

-                      GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);

-   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );

-   texel[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );

-   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s[2] );

-   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s[3] );

-}

-

-#if DIM == 3

-static void

-store_texel_signed_rgba_16(struct gl_texture_image *texImage,

-                           GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLshort *rgba = (const GLshort *) texel;

-   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-

-/* MESA_FORMAT_RGBA_16 ***********************************************/

-

-static void

-FETCH(rgba_16)(const struct gl_texture_image *texImage,

-               GLint i, GLint j, GLint k, GLfloat *texel)

-{

-   const GLushort *s = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);

-   texel[RCOMP] = USHORT_TO_FLOAT( s[0] );

-   texel[GCOMP] = USHORT_TO_FLOAT( s[1] );

-   texel[BCOMP] = USHORT_TO_FLOAT( s[2] );

-   texel[ACOMP] = USHORT_TO_FLOAT( s[3] );

-}

-

-#if DIM == 3

-static void

-store_texel_rgba_16(struct gl_texture_image *texImage,

-                    GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLushort *rgba = (const GLushort *) texel;

-   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);

-   dst[0] = rgba[RCOMP];

-   dst[1] = rgba[GCOMP];

-   dst[2] = rgba[BCOMP];

-   dst[3] = rgba[ACOMP];

-}

-#endif

-

-

-

-/* MESA_FORMAT_YCBCR *********************************************************/

-

-/* Fetch texel from 1D, 2D or 3D ycbcr texture, return 4 GLfloats.

- * We convert YCbCr to RGB here.

- */

-static void FETCH(f_ycbcr)( const struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src0 = TEXEL_ADDR(GLushort, texImage, (i & ~1), j, k, 1); /* even */

-   const GLushort *src1 = src0 + 1;                               /* odd */

-   const GLubyte y0 = (*src0 >> 8) & 0xff;  /* luminance */

-   const GLubyte cb = *src0 & 0xff;         /* chroma U */

-   const GLubyte y1 = (*src1 >> 8) & 0xff;  /* luminance */

-   const GLubyte cr = *src1 & 0xff;         /* chroma V */

-   const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */

-   GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);

-   GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);

-   GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);

-   r *= (1.0F / 255.0F);

-   g *= (1.0F / 255.0F);

-   b *= (1.0F / 255.0F);

-   texel[RCOMP] = CLAMP(r, 0.0F, 1.0F);

-   texel[GCOMP] = CLAMP(g, 0.0F, 1.0F);

-   texel[BCOMP] = CLAMP(b, 0.0F, 1.0F);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_ycbcr(struct gl_texture_image *texImage,

-                              GLint i, GLint j, GLint k, const void *texel)

-{

-   (void) texImage;

-   (void) i;

-   (void) j;

-   (void) k;

-   (void) texel;

-   /* XXX to do */

-}

-#endif

-

-

-/* MESA_FORMAT_YCBCR_REV *****************************************************/

-

-/* Fetch texel from 1D, 2D or 3D ycbcr_rev texture, return 4 GLfloats.

- * We convert YCbCr to RGB here.

- */

-static void FETCH(f_ycbcr_rev)( const struct gl_texture_image *texImage,

-                                GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLushort *src0 = TEXEL_ADDR(GLushort, texImage, (i & ~1), j, k, 1); /* even */

-   const GLushort *src1 = src0 + 1;                               /* odd */

-   const GLubyte y0 = *src0 & 0xff;         /* luminance */

-   const GLubyte cr = (*src0 >> 8) & 0xff;  /* chroma V */

-   const GLubyte y1 = *src1 & 0xff;         /* luminance */

-   const GLubyte cb = (*src1 >> 8) & 0xff;  /* chroma U */

-   const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */

-   GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);

-   GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);

-   GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);

-   r *= (1.0F / 255.0F);

-   g *= (1.0F / 255.0F);

-   b *= (1.0F / 255.0F);

-   texel[RCOMP] = CLAMP(r, 0.0F, 1.0F);

-   texel[GCOMP] = CLAMP(g, 0.0F, 1.0F);

-   texel[BCOMP] = CLAMP(b, 0.0F, 1.0F);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_ycbcr_rev(struct gl_texture_image *texImage,

-                                  GLint i, GLint j, GLint k, const void *texel)

-{

-   (void) texImage;

-   (void) i;

-   (void) j;

-   (void) k;

-   (void) texel;

-   /* XXX to do */

-}

-#endif

-

-

-/* MESA_TEXFORMAT_Z24_S8 ***************************************************/

-

-static void FETCH(f_z24_s8)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   /* only return Z, not stencil data */

-   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   const GLfloat scale = 1.0F / (GLfloat) 0xffffff;

-   texel[0] = ((*src) >> 8) * scale;

-   ASSERT(texImage->TexFormat == MESA_FORMAT_Z24_S8);

-   ASSERT(texel[0] >= 0.0F);

-   ASSERT(texel[0] <= 1.0F);

-}

-

-#if DIM == 3

-static void store_texel_z24_s8(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   /* only store Z, not stencil */

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   GLfloat depth = *((GLfloat *) texel);

-   GLuint zi = ((GLuint) (depth * 0xffffff)) << 8;

-   *dst = zi | (*dst & 0xff);

-}

-#endif

-

-

-/* MESA_TEXFORMAT_S8_Z24 ***************************************************/

-

-static void FETCH(f_s8_z24)( const struct gl_texture_image *texImage,

-                             GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   /* only return Z, not stencil data */

-   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   const GLfloat scale = 1.0F / (GLfloat) 0xffffff;

-   texel[0] = ((*src) & 0x00ffffff) * scale;

-   ASSERT(texImage->TexFormat == MESA_FORMAT_S8_Z24);

-   ASSERT(texel[0] >= 0.0F);

-   ASSERT(texel[0] <= 1.0F);

-}

-

-#if DIM == 3

-static void store_texel_s8_z24(struct gl_texture_image *texImage,

-                               GLint i, GLint j, GLint k, const void *texel)

-{

-   /* only store Z, not stencil */

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   GLfloat depth = *((GLfloat *) texel);

-   GLuint zi = (GLuint) (depth * 0xffffff);

-   *dst = zi | (*dst & 0xff000000);

-}

-#endif

-

-

-/* MESA_FORMAT_RGB9_E5 ******************************************************/

-

-static void FETCH(rgb9_e5)( const struct gl_texture_image *texImage,

-                            GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   rgb9e5_to_float3(*src, texel);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_rgb9_e5(struct gl_texture_image *texImage,

-                                GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *src = (const GLfloat *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = float3_to_rgb9e5(src);

-}

-#endif

-

-

-/* MESA_FORMAT_R11_G11_B10_FLOAT *********************************************/

-

-static void FETCH(r11_g11_b10f)( const struct gl_texture_image *texImage,

-                                 GLint i, GLint j, GLint k, GLfloat *texel )

-{

-   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   r11g11b10f_to_float3(*src, texel);

-   texel[ACOMP] = 1.0F;

-}

-

-#if DIM == 3

-static void store_texel_r11_g11_b10f(struct gl_texture_image *texImage,

-                                     GLint i, GLint j, GLint k, const void *texel)

-{

-   const GLfloat *src = (const GLfloat *) texel;

-   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);

-   *dst = float3_to_r11g11b10f(src);

-}

-#endif

-

-

-#undef TEXEL_ADDR

-#undef DIM

-#undef FETCH

+/*
+ * Mesa 3-D graphics library
+ * Version:  7.7
+ *
+ * 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.
+ */
+
+
+/**
+ * \file texfetch_tmp.h
+ * Texel fetch functions template.
+ * 
+ * This template file is used by texfetch.c to generate texel fetch functions
+ * for 1-D, 2-D and 3-D texture images. 
+ *
+ * It should be expanded by defining \p DIM as the number texture dimensions
+ * (1, 2 or 3).  According to the value of \p DIM a series of macros is defined
+ * for the texel lookup in the gl_texture_image::Data.
+ * 
+ * \author Gareth Hughes
+ * \author Brian Paul
+ */
+
+
+#if DIM == 1
+
+#define TEXEL_ADDR( type, image, i, j, k, size ) \
+	((void) (j), (void) (k), ((type *)(image)->Data + (i) * (size)))
+
+#define FETCH(x) fetch_texel_1d_##x
+
+#elif DIM == 2
+
+#define TEXEL_ADDR( type, image, i, j, k, size )			\
+	((void) (k),							\
+	 ((type *)(image)->Data + ((image)->RowStride * (j) + (i)) * (size)))
+
+#define FETCH(x) fetch_texel_2d_##x
+
+#elif DIM == 3
+
+#define TEXEL_ADDR( type, image, i, j, k, size )			\
+	((type *)(image)->Data + ((image)->ImageOffsets[k]		\
+             + (image)->RowStride * (j) + (i)) * (size))
+
+#define FETCH(x) fetch_texel_3d_##x
+
+#else
+#error	illegal number of texture dimensions
+#endif
+
+
+/* MESA_FORMAT_Z32 ***********************************************************/
+
+/* Fetch depth texel from 1D, 2D or 3D 32-bit depth texture,
+ * returning 1 GLfloat.
+ * Note: no GLchan version of this function.
+ */
+static void FETCH(f_z32)( const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[0] = src[0] * (1.0F / 0xffffffff);
+}
+
+#if DIM == 3
+static void store_texel_z32(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLuint *depth = (const GLuint *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   dst[0] = *depth;
+}
+#endif
+
+
+/* MESA_FORMAT_Z16 ***********************************************************/
+
+/* Fetch depth texel from 1D, 2D or 3D 16-bit depth texture,
+ * returning 1 GLfloat.
+ * Note: no GLchan version of this function.
+ */
+static void FETCH(f_z16)(const struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[0] = src[0] * (1.0F / 65535.0F);
+}
+
+#if DIM == 3
+static void store_texel_z16(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *depth = (const GLushort *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   dst[0] = *depth;
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_F32 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D RGBA_FLOAT32 texture, returning 4 GLfloats.
+ */
+static void FETCH(f_rgba_f32)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 4);
+   texel[RCOMP] = src[0];
+   texel[GCOMP] = src[1];
+   texel[BCOMP] = src[2];
+   texel[ACOMP] = src[3];
+}
+
+#if DIM == 3
+static void store_texel_rgba_f32(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *depth = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 4);
+   dst[0] = depth[RCOMP];
+   dst[1] = depth[GCOMP];
+   dst[2] = depth[BCOMP];
+   dst[3] = depth[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_F16 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D RGBA_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_rgba_f16)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 4);
+   texel[RCOMP] = _mesa_half_to_float(src[0]);
+   texel[GCOMP] = _mesa_half_to_float(src[1]);
+   texel[BCOMP] = _mesa_half_to_float(src[2]);
+   texel[ACOMP] = _mesa_half_to_float(src[3]);
+}
+
+#if DIM == 3
+static void store_texel_rgba_f16(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *src = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 4);
+   dst[0] = _mesa_float_to_half(src[RCOMP]);
+   dst[1] = _mesa_float_to_half(src[GCOMP]);
+   dst[2] = _mesa_float_to_half(src[BCOMP]);
+   dst[3] = _mesa_float_to_half(src[ACOMP]);
+}
+#endif
+
+/* MESA_FORMAT_RGB_F32 *******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D RGB_FLOAT32 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_rgb_f32)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 3);
+   texel[RCOMP] = src[0];
+   texel[GCOMP] = src[1];
+   texel[BCOMP] = src[2];
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgb_f32(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *src = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 3);
+   dst[0] = src[RCOMP];
+   dst[1] = src[GCOMP];
+   dst[2] = src[BCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGB_F16 *******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D RGB_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_rgb_f16)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 3);
+   texel[RCOMP] = _mesa_half_to_float(src[0]);
+   texel[GCOMP] = _mesa_half_to_float(src[1]);
+   texel[BCOMP] = _mesa_half_to_float(src[2]);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgb_f16(struct gl_texture_image *texImage,
+                                GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *src = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 3);
+   dst[0] = _mesa_float_to_half(src[RCOMP]);
+   dst[1] = _mesa_float_to_half(src[GCOMP]);
+   dst[2] = _mesa_float_to_half(src[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_ALPHA_F32 *****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D ALPHA_FLOAT32 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_alpha_f32)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = src[0];
+}
+
+#if DIM == 3
+static void store_texel_alpha_f32(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   dst[0] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_ALPHA_F32 *****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D ALPHA_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_alpha_f16)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = _mesa_half_to_float(src[0]);
+}
+
+#if DIM == 3
+static void store_texel_alpha_f16(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   dst[0] = _mesa_float_to_half(rgba[ACOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_LUMINANCE_F32 *************************************************/
+
+/* Fetch texel from 1D, 2D or 3D LUMINANCE_FLOAT32 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_luminance_f32)( const struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = src[0];
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_luminance_f32(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   dst[0] = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_LUMINANCE_F16 *************************************************/
+
+/* Fetch texel from 1D, 2D or 3D LUMINANCE_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_luminance_f16)( const struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = _mesa_half_to_float(src[0]);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_luminance_f16(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   dst[0] = _mesa_float_to_half(rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_LUMINANCE_ALPHA_F32 *******************************************/
+
+/* Fetch texel from 1D, 2D or 3D LUMINANCE_ALPHA_FLOAT32 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_luminance_alpha_f32)( const struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = src[0];
+   texel[ACOMP] = src[1];
+}
+
+#if DIM == 3
+static void store_texel_luminance_alpha_f32(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_LUMINANCE_ALPHA_F16 *******************************************/
+
+/* Fetch texel from 1D, 2D or 3D LUMINANCE_ALPHA_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_luminance_alpha_f16)( const struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = _mesa_half_to_float(src[0]);
+   texel[ACOMP] = _mesa_half_to_float(src[1]);
+}
+
+#if DIM == 3
+static void store_texel_luminance_alpha_f16(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);
+   dst[0] = _mesa_float_to_half(rgba[RCOMP]);
+   dst[1] = _mesa_float_to_half(rgba[ACOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_INTENSITY_F32 *************************************************/
+
+/* Fetch texel from 1D, 2D or 3D INTENSITY_FLOAT32 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_intensity_f32)( const struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] =
+   texel[ACOMP] = src[0];
+}
+
+#if DIM == 3
+static void store_texel_intensity_f32(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   dst[0] = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_INTENSITY_F16 *************************************************/
+
+/* Fetch texel from 1D, 2D or 3D INTENSITY_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_intensity_f16)( const struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] =
+   texel[ACOMP] = _mesa_half_to_float(src[0]);
+}
+
+#if DIM == 3
+static void store_texel_intensity_f16(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   dst[0] = _mesa_float_to_half(rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_R_FLOAT32 *****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D R_FLOAT32 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_r_f32)( const struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   texel[RCOMP] = src[0];
+   texel[GCOMP] = 0.0F;
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_r_f32(struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 1);
+   dst[0] = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_R_FLOAT16 *****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D R_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_r_f16)( const struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   texel[RCOMP] = _mesa_half_to_float(src[0]);
+   texel[GCOMP] = 0.0F;
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_r_f16(struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 1);
+   dst[0] = _mesa_float_to_half(rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RG_FLOAT32 ****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D RG_FLOAT32 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_rg_f32)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);
+   texel[RCOMP] = src[0];
+   texel[GCOMP] = src[1];
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rg_f32(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RG_FLOAT16 ****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D RG_FLOAT16 texture,
+ * returning 4 GLfloats.
+ */
+static void FETCH(f_rg_f16)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLhalfARB *src = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);
+   texel[RCOMP] = _mesa_half_to_float(src[0]);
+   texel[GCOMP] = _mesa_half_to_float(src[1]);
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rg_f16(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *rgba = (const GLfloat *) texel;
+   GLhalfARB *dst = TEXEL_ADDR(GLhalfARB, texImage, i, j, k, 2);
+   dst[0] = _mesa_float_to_half(rgba[RCOMP]);
+   dst[1] = _mesa_float_to_half(rgba[GCOMP]);
+}
+#endif
+
+
+/*
+ * Begin Hardware formats
+ */
+
+/* MESA_FORMAT_RGBA8888 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rgba8888 texture, return 4 GLfloats */
+static void FETCH(f_rgba8888)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
+   texel[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
+   texel[BCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
+   texel[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
+}
+
+
+
+#if DIM == 3
+static void store_texel_rgba8888(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA888_REV ***************************************************/
+
+/* Fetch texel from 1D, 2D or 3D abgr8888 texture, return 4 GLchans */
+static void FETCH(f_rgba8888_rev)( const struct gl_texture_image *texImage,
+                                   GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
+   texel[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
+   texel[BCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
+   texel[ACOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
+}
+
+#if DIM == 3
+static void store_texel_rgba8888_rev(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888_REV(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_ARGB8888 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb8888 texture, return 4 GLchans */
+static void FETCH(f_argb8888)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
+   texel[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
+   texel[BCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
+   texel[ACOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
+}
+
+#if DIM == 3
+static void store_texel_argb8888(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_ARGB8888_REV **************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb8888_rev texture, return 4 GLfloats */
+static void FETCH(f_argb8888_rev)( const struct gl_texture_image *texImage,
+                                   GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
+   texel[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
+   texel[BCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
+   texel[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
+}
+
+#if DIM == 3
+static void store_texel_argb8888_rev(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[BCOMP], rgba[GCOMP], rgba[RCOMP], rgba[ACOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_XRGB8888 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D xrgb8888 texture, return 4 GLchans */
+static void FETCH(f_xrgb8888)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
+   texel[GCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
+   texel[BCOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff );
+   texel[ACOMP] = 1.0f;
+}
+
+#if DIM == 3
+static void store_texel_xrgb8888(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(0xff, rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_XRGB8888_REV **************************************************/
+
+/* Fetch texel from 1D, 2D or 3D xrgb8888_rev texture, return 4 GLfloats */
+static void FETCH(f_xrgb8888_rev)( const struct gl_texture_image *texImage,
+                                   GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( (s >>  8) & 0xff );
+   texel[GCOMP] = UBYTE_TO_FLOAT( (s >> 16) & 0xff );
+   texel[BCOMP] = UBYTE_TO_FLOAT( (s >> 24)        );
+   texel[ACOMP] = 1.0f;
+}
+
+#if DIM == 3
+static void store_texel_xrgb8888_rev(struct gl_texture_image *texImage,
+                                     GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[BCOMP], rgba[GCOMP], rgba[RCOMP], 0xff);
+}
+#endif
+
+
+/* MESA_FORMAT_RGB888 ********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rgb888 texture, return 4 GLchans */
+static void FETCH(f_rgb888)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);
+   texel[RCOMP] = UBYTE_TO_FLOAT( src[2] );
+   texel[GCOMP] = UBYTE_TO_FLOAT( src[1] );
+   texel[BCOMP] = UBYTE_TO_FLOAT( src[0] );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgb888(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);
+   dst[0] = rgba[BCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_BGR888 ********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D bgr888 texture, return 4 GLchans */
+static void FETCH(f_bgr888)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);
+   texel[RCOMP] = UBYTE_TO_FLOAT( src[0] );
+   texel[GCOMP] = UBYTE_TO_FLOAT( src[1] );
+   texel[BCOMP] = UBYTE_TO_FLOAT( src[2] );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_bgr888(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+}
+#endif
+
+
+/* use color expansion like (g << 2) | (g >> 4) (does somewhat random rounding)
+   instead of slow (g << 2) * 255 / 252 (always rounds down) */
+
+/* MESA_FORMAT_RGB565 ********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rgb565 texture, return 4 GLchans */
+static void FETCH(f_rgb565)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = *src;
+   texel[RCOMP] = ((s >> 11) & 0x1f) * (1.0F / 31.0F);
+   texel[GCOMP] = ((s >> 5 ) & 0x3f) * (1.0F / 63.0F);
+   texel[BCOMP] = ((s      ) & 0x1f) * (1.0F / 31.0F);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgb565(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_565(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RGB565_REV ****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rgb565_rev texture, return 4 GLchans */
+static void FETCH(f_rgb565_rev)( const struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = (*src >> 8) | (*src << 8); /* byte swap */
+   texel[RCOMP] = UBYTE_TO_FLOAT( ((s >> 8) & 0xf8) | ((s >> 13) & 0x7) );
+   texel[GCOMP] = UBYTE_TO_FLOAT( ((s >> 3) & 0xfc) | ((s >>  9) & 0x3) );
+   texel[BCOMP] = UBYTE_TO_FLOAT( ((s << 3) & 0xf8) | ((s >>  2) & 0x7) );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgb565_rev(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_565(rgba[BCOMP], rgba[GCOMP], rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_ARGB4444 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb444 texture, return 4 GLchans */
+static void FETCH(f_argb4444)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = *src;
+   texel[RCOMP] = ((s >>  8) & 0xf) * (1.0F / 15.0F);
+   texel[GCOMP] = ((s >>  4) & 0xf) * (1.0F / 15.0F);
+   texel[BCOMP] = ((s      ) & 0xf) * (1.0F / 15.0F);
+   texel[ACOMP] = ((s >> 12) & 0xf) * (1.0F / 15.0F);
+}
+
+#if DIM == 3
+static void store_texel_argb4444(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_4444(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_ARGB4444_REV **************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb4444_rev texture, return 4 GLchans */
+static void FETCH(f_argb4444_rev)( const struct gl_texture_image *texImage,
+                                   GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] = ((s      ) & 0xf) * (1.0F / 15.0F);
+   texel[GCOMP] = ((s >> 12) & 0xf) * (1.0F / 15.0F);
+   texel[BCOMP] = ((s >>  8) & 0xf) * (1.0F / 15.0F);
+   texel[ACOMP] = ((s >>  4) & 0xf) * (1.0F / 15.0F);
+}
+
+#if DIM == 3
+static void store_texel_argb4444_rev(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_4444(rgba[ACOMP], rgba[BCOMP], rgba[GCOMP], rgba[RCOMP]);
+}
+#endif
+
+/* MESA_FORMAT_RGBA5551 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb1555 texture, return 4 GLchans */
+static void FETCH(f_rgba5551)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = *src;
+   texel[RCOMP] = ((s >> 11) & 0x1f) * (1.0F / 31.0F);
+   texel[GCOMP] = ((s >>  6) & 0x1f) * (1.0F / 31.0F);
+   texel[BCOMP] = ((s >>  1) & 0x1f) * (1.0F / 31.0F);
+   texel[ACOMP] = ((s      ) & 0x01) * 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgba5551(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_5551(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
+
+/* MESA_FORMAT_ARGB1555 ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb1555 texture, return 4 GLchans */
+static void FETCH(f_argb1555)( const struct gl_texture_image *texImage,
+			     GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = *src;
+   texel[RCOMP] = ((s >> 10) & 0x1f) * (1.0F / 31.0F);
+   texel[GCOMP] = ((s >>  5) & 0x1f) * (1.0F / 31.0F);
+   texel[BCOMP] = ((s >>  0) & 0x1f) * (1.0F / 31.0F);
+   texel[ACOMP] = ((s >> 15) & 0x01) * 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_argb1555(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_1555(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_ARGB1555_REV **************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb1555_rev texture, return 4 GLchans */
+static void FETCH(f_argb1555_rev)( const struct gl_texture_image *texImage,
+                                   GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   const GLushort s = (*src << 8) | (*src >> 8); /* byteswap */
+   texel[RCOMP] = UBYTE_TO_FLOAT( ((s >>  7) & 0xf8) | ((s >> 12) & 0x7) );
+   texel[GCOMP] = UBYTE_TO_FLOAT( ((s >>  2) & 0xf8) | ((s >>  7) & 0x7) );
+   texel[BCOMP] = UBYTE_TO_FLOAT( ((s <<  3) & 0xf8) | ((s >>  2) & 0x7) );
+   texel[ACOMP] = UBYTE_TO_FLOAT( ((s >> 15) & 0x01) * 255 );
+}
+
+#if DIM == 3
+static void store_texel_argb1555_rev(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_1555_REV(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_ARGB2101010 ***************************************************/
+
+/* Fetch texel from 1D, 2D or 3D argb2101010 texture, return 4 GLchans */
+static void FETCH(f_argb2101010)( const struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   const GLuint s = *src;
+   texel[RCOMP] = ((s >> 20) & 0x3ff) * (1.0F / 1023.0F);
+   texel[GCOMP] = ((s >> 10) & 0x3ff) * (1.0F / 1023.0F);
+   texel[BCOMP] = ((s >>  0) & 0x3ff) * (1.0F / 1023.0F);
+   texel[ACOMP] = ((s >> 30) & 0x03) * (1.0F / 3.0F);
+}
+
+#if DIM == 3
+static void store_texel_argb2101010(struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_2101010_UB(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RG88 **********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rg88 texture, return 4 GLchans */
+static void FETCH(f_rg88)( const struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( s & 0xff );
+   texel[GCOMP] = UBYTE_TO_FLOAT( s >> 8 );
+   texel[BCOMP] = 0.0;
+   texel[ACOMP] = 1.0;
+}
+
+#if DIM == 3
+static void store_texel_rg88(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_88(rgba[RCOMP], rgba[GCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RG88_REV ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rg88_rev texture, return 4 GLchans */
+static void FETCH(f_rg88_rev)( const struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT( s & 0xff );
+   texel[GCOMP] = UBYTE_TO_FLOAT( s >> 8 );
+   texel[BCOMP] = 0.0;
+   texel[ACOMP] = 1.0;
+}
+
+#if DIM == 3
+static void store_texel_rg88_rev(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_88(rgba[GCOMP], rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_AL44 **********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D al44 texture, return 4 GLchans */
+static void FETCH(f_al44)( const struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte s = *TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = (s & 0xf) * (1.0F / 15.0F);
+   texel[ACOMP] = ((s >> 4) & 0xf) * (1.0F / 15.0F);
+}
+
+#if DIM == 3
+static void store_texel_al44(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_44(rgba[ACOMP], rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_AL88 **********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D al88 texture, return 4 GLchans */
+static void FETCH(f_al88)( const struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] = 
+   texel[GCOMP] = 
+   texel[BCOMP] = UBYTE_TO_FLOAT( s & 0xff );
+   texel[ACOMP] = UBYTE_TO_FLOAT( s >> 8 );
+}
+
+#if DIM == 3
+static void store_texel_al88(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_88(rgba[ACOMP], rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_R8 ************************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rg88 texture, return 4 GLchans */
+static void FETCH(f_r8)(const struct gl_texture_image *texImage,
+			GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLubyte s = *TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   texel[RCOMP] = UBYTE_TO_FLOAT(s);
+   texel[GCOMP] = 0.0;
+   texel[BCOMP] = 0.0;
+   texel[ACOMP] = 1.0;
+}
+
+#if DIM == 3
+static void store_texel_r8(struct gl_texture_image *texImage,
+			   GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_R16 ***********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D r16 texture, return 4 GLchans */
+static void FETCH(f_r16)(const struct gl_texture_image *texImage,
+			GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] = USHORT_TO_FLOAT(s);
+   texel[GCOMP] = 0.0;
+   texel[BCOMP] = 0.0;
+   texel[ACOMP] = 1.0;
+}
+
+#if DIM == 3
+static void store_texel_r16(struct gl_texture_image *texImage,
+			    GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_AL88_REV ******************************************************/
+
+/* Fetch texel from 1D, 2D or 3D al88_rev texture, return 4 GLchans */
+static void FETCH(f_al88_rev)( const struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort s = *TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] = 
+   texel[GCOMP] = 
+   texel[BCOMP] = UBYTE_TO_FLOAT( s >> 8 );
+   texel[ACOMP] = UBYTE_TO_FLOAT( s & 0xff );
+}
+
+#if DIM == 3
+static void store_texel_al88_rev(struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_88(rgba[RCOMP], rgba[ACOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RG1616 ********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rg1616 texture, return 4 GLchans */
+static void FETCH(f_rg1616)( const struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = USHORT_TO_FLOAT( s & 0xffff );
+   texel[GCOMP] = USHORT_TO_FLOAT( s >> 16 );
+   texel[BCOMP] = 0.0;
+   texel[ACOMP] = 1.0;
+}
+
+#if DIM == 3
+static void store_texel_rg1616(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_1616(rgba[RCOMP], rgba[GCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RG1616_REV ****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rg1616_rev texture, return 4 GLchans */
+static void FETCH(f_rg1616_rev)( const struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = USHORT_TO_FLOAT( s >> 16 );
+   texel[GCOMP] = USHORT_TO_FLOAT( s & 0xffff );
+   texel[BCOMP] = 0.0;
+   texel[ACOMP] = 1.0;
+}
+
+#if DIM == 3
+static void store_texel_rg1616_rev(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_1616(rgba[GCOMP], rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_AL1616 ********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D al1616 texture, return 4 GLchans */
+static void FETCH(f_al1616)( const struct gl_texture_image *texImage,
+			     GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = USHORT_TO_FLOAT( s & 0xffff );
+   texel[ACOMP] = USHORT_TO_FLOAT( s >> 16 );
+}
+
+#if DIM == 3
+static void store_texel_al1616(struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_1616(rgba[ACOMP], rgba[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_AL1616_REV ****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D al1616_rev texture, return 4 GLchans */
+static void FETCH(f_al1616_rev)( const struct gl_texture_image *texImage,
+				 GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = USHORT_TO_FLOAT( s >> 16 );
+   texel[ACOMP] = USHORT_TO_FLOAT( s & 0xffff );
+}
+
+#if DIM == 3
+static void store_texel_al1616_rev(struct gl_texture_image *texImage,
+				   GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_1616(rgba[RCOMP], rgba[ACOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_RGB332 ********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D rgb332 texture, return 4 GLchans */
+static void FETCH(f_rgb332)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   const GLubyte s = *src;
+   texel[RCOMP] = ((s >> 5) & 0x7) * (1.0F / 7.0F);
+   texel[GCOMP] = ((s >> 2) & 0x7) * (1.0F / 7.0F);
+   texel[BCOMP] = ((s     ) & 0x3) * (1.0F / 3.0F);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgb332(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_332(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_A8 ************************************************************/
+
+/* Fetch texel from 1D, 2D or 3D a8 texture, return 4 GLchans */
+static void FETCH(f_a8)( const struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = UBYTE_TO_FLOAT( src[0] );
+}
+
+#if DIM == 3
+static void store_texel_a8(struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   *dst = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_A16 ************************************************************/
+
+/* Fetch texel from 1D, 2D or 3D a8 texture, return 4 GLchans */
+static void FETCH(f_a16)( const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = USHORT_TO_FLOAT( src[0] );
+}
+
+#if DIM == 3
+static void store_texel_a16(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_L8 ************************************************************/
+
+/* Fetch texel from 1D, 2D or 3D l8 texture, return 4 GLchans */
+static void FETCH(f_l8)( const struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = UBYTE_TO_FLOAT( src[0] );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_l8(struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_L16 ***********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D l16 texture, return 4 GLchans */
+static void FETCH(f_l16)( const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = USHORT_TO_FLOAT( src[0] );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_l16(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_I8 ************************************************************/
+
+/* Fetch texel from 1D, 2D or 3D i8 texture, return 4 GLchans */
+static void FETCH(f_i8)( const struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] =
+   texel[ACOMP] = UBYTE_TO_FLOAT( src[0] );
+}
+
+#if DIM == 3
+static void store_texel_i8(struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_I16 ***********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D i16 texture, return 4 GLchans */
+static void FETCH(f_i16)( const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] =
+   texel[ACOMP] = USHORT_TO_FLOAT( src[0] );
+}
+
+#if DIM == 3
+static void store_texel_i16(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_CI8 ***********************************************************/
+
+/* Fetch CI texel from 1D, 2D or 3D ci8 texture, lookup the index in a
+ * color table, and return 4 GLchans.
+ */
+static void FETCH(f_ci8)( const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   const struct gl_color_table *palette;
+   GLuint index;
+   GET_CURRENT_CONTEXT(ctx);
+
+   if (ctx->Texture.SharedPalette) {
+      palette = &ctx->Texture.Palette;
+   }
+   else {
+      palette = &texImage->TexObject->Palette;
+   }
+   if (palette->Size == 0)
+      return; /* undefined results */
+
+   /* Mask the index against size of palette to avoid going out of bounds */
+   index = (*src) & (palette->Size - 1);
+
+   {
+      const GLfloat *table = palette->TableF;
+      switch (palette->_BaseFormat) {
+      case GL_ALPHA:
+         texel[RCOMP] =
+         texel[GCOMP] =
+         texel[BCOMP] = 0.0F;
+         texel[ACOMP] = table[index];
+         break;
+      case GL_LUMINANCE:
+         texel[RCOMP] =
+         texel[GCOMP] =
+         texel[BCOMP] = table[index];
+         texel[ACOMP] = 1.0F;
+         break;
+      case GL_INTENSITY:
+         texel[RCOMP] =
+         texel[GCOMP] =
+         texel[BCOMP] =
+         texel[ACOMP] = table[index];
+         break;
+      case GL_LUMINANCE_ALPHA:
+         texel[RCOMP] =
+         texel[GCOMP] =
+         texel[BCOMP] = table[index * 2 + 0];
+         texel[ACOMP] = table[index * 2 + 1];
+         break;
+      case GL_RGB:
+         texel[RCOMP] = table[index * 3 + 0];
+         texel[GCOMP] = table[index * 3 + 1];
+         texel[BCOMP] = table[index * 3 + 2];
+         texel[ACOMP] = 1.0F;
+         break;
+      case GL_RGBA:
+         texel[RCOMP] = table[index * 4 + 0];
+         texel[GCOMP] = table[index * 4 + 1];
+         texel[BCOMP] = table[index * 4 + 2];
+         texel[ACOMP] = table[index * 4 + 3];
+         break;
+      default:
+         _mesa_problem(ctx, "Bad palette format in fetch_texel_ci8");
+         return;
+      }
+   }
+}
+
+#if DIM == 3
+static void store_texel_ci8(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *index = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   *dst = *index;
+}
+#endif
+
+
+/* Fetch texel from 1D, 2D or 3D srgb8 texture, return 4 GLfloats */
+/* Note: component order is same as for MESA_FORMAT_RGB888 */
+static void FETCH(srgb8)(const struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);
+   texel[RCOMP] = nonlinear_to_linear(src[2]);
+   texel[GCOMP] = nonlinear_to_linear(src[1]);
+   texel[BCOMP] = nonlinear_to_linear(src[0]);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_srgb8(struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 3);
+   dst[0] = rgba[BCOMP]; /* no conversion */
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[RCOMP];
+}
+#endif
+
+/* Fetch texel from 1D, 2D or 3D srgba8 texture, return 4 GLfloats */
+static void FETCH(srgba8)(const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = nonlinear_to_linear( (s >> 24) );
+   texel[GCOMP] = nonlinear_to_linear( (s >> 16) & 0xff );
+   texel[BCOMP] = nonlinear_to_linear( (s >>  8) & 0xff );
+   texel[ACOMP] = UBYTE_TO_FLOAT( (s      ) & 0xff ); /* linear! */
+}
+
+#if DIM == 3
+static void store_texel_srgba8(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
+
+/* Fetch texel from 1D, 2D or 3D sargb8 texture, return 4 GLfloats */
+static void FETCH(sargb8)(const struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = nonlinear_to_linear( (s >> 16) & 0xff );
+   texel[GCOMP] = nonlinear_to_linear( (s >>  8) & 0xff );
+   texel[BCOMP] = nonlinear_to_linear( (s      ) & 0xff );
+   texel[ACOMP] = UBYTE_TO_FLOAT( (s >> 24) ); /* linear! */
+}
+
+#if DIM == 3
+static void store_texel_sargb8(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[ACOMP], rgba[RCOMP], rgba[GCOMP], rgba[BCOMP]);
+}
+#endif
+
+/* Fetch texel from 1D, 2D or 3D sl8 texture, return 4 GLfloats */
+static void FETCH(sl8)(const struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   texel[RCOMP] = 
+   texel[GCOMP] = 
+   texel[BCOMP] = nonlinear_to_linear(src[0]);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_sl8(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 1);
+   dst[0] = rgba[RCOMP];
+}
+#endif
+
+/* Fetch texel from 1D, 2D or 3D sla8 texture, return 4 GLfloats */
+static void FETCH(sla8)(const struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 2);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = nonlinear_to_linear(src[0]);
+   texel[ACOMP] = UBYTE_TO_FLOAT(src[1]); /* linear */
+}
+
+#if DIM == 3
+static void store_texel_sla8(struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 2);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_INT8 **************************************************/
+
+static void
+FETCH(rgba_int8)(const struct gl_texture_image *texImage,
+                 GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLbyte *src = TEXEL_ADDR(GLbyte, texImage, i, j, k, 4);
+   texel[RCOMP] = (GLfloat) src[0];
+   texel[GCOMP] = (GLfloat) src[1];
+   texel[BCOMP] = (GLfloat) src[2];
+   texel[ACOMP] = (GLfloat) src[3];
+}
+
+#if DIM == 3
+static void
+store_texel_rgba_int8(struct gl_texture_image *texImage,
+                      GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rgba = (const GLbyte *) texel;
+   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_INT16 **************************************************/
+
+static void
+FETCH(rgba_int16)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLshort *src = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);
+   texel[RCOMP] = (GLfloat) src[0];
+   texel[GCOMP] = (GLfloat) src[1];
+   texel[BCOMP] = (GLfloat) src[2];
+   texel[ACOMP] = (GLfloat) src[3];
+}
+
+#if DIM == 3
+static void
+store_texel_rgba_int16(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_INT32 **************************************************/
+
+static void
+FETCH(rgba_int32)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLint *src = TEXEL_ADDR(GLint, texImage, i, j, k, 4);
+   texel[RCOMP] = (GLfloat) src[0];
+   texel[GCOMP] = (GLfloat) src[1];
+   texel[BCOMP] = (GLfloat) src[2];
+   texel[ACOMP] = (GLfloat) src[3];
+}
+
+#if DIM == 3
+static void
+store_texel_rgba_int32(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLint *rgba = (const GLint *) texel;
+   GLint *dst = TEXEL_ADDR(GLint, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_UINT8 **************************************************/
+
+static void
+FETCH(rgba_uint8)(const struct gl_texture_image *texImage,
+                 GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLubyte *src = TEXEL_ADDR(GLubyte, texImage, i, j, k, 4);
+   texel[RCOMP] = (GLfloat) src[0];
+   texel[GCOMP] = (GLfloat) src[1];
+   texel[BCOMP] = (GLfloat) src[2];
+   texel[ACOMP] = (GLfloat) src[3];
+}
+
+#if DIM == 3
+static void
+store_texel_rgba_uint8(struct gl_texture_image *texImage,
+                      GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLubyte *dst = TEXEL_ADDR(GLubyte, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_UINT16 **************************************************/
+
+static void
+FETCH(rgba_uint16)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);
+   texel[RCOMP] = (GLfloat) src[0];
+   texel[GCOMP] = (GLfloat) src[1];
+   texel[BCOMP] = (GLfloat) src[2];
+   texel[ACOMP] = (GLfloat) src[3];
+}
+
+#if DIM == 3
+static void
+store_texel_rgba_uint16(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_RGBA_UINT32 **************************************************/
+
+static void
+FETCH(rgba_uint32)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 4);
+   texel[RCOMP] = (GLfloat) src[0];
+   texel[GCOMP] = (GLfloat) src[1];
+   texel[BCOMP] = (GLfloat) src[2];
+   texel[ACOMP] = (GLfloat) src[3];
+}
+
+#if DIM == 3
+static void
+store_texel_rgba_uint32(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLuint *rgba = (const GLuint *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_DUDV8 ********************************************************/
+
+/* this format by definition produces 0,0,0,1 as rgba values,
+   however we'll return the dudv values as rg and fix up elsewhere */
+static void FETCH(dudv8)(const struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLbyte *src = TEXEL_ADDR(GLbyte, texImage, i, j, k, 2);
+   texel[RCOMP] = BYTE_TO_FLOAT(src[0]);
+   texel[GCOMP] = BYTE_TO_FLOAT(src[1]);
+   texel[BCOMP] = 0;
+   texel[ACOMP] = 0;
+}
+
+
+/* MESA_FORMAT_SIGNED_R8 ***********************************************/
+
+static void FETCH(signed_r8)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   texel[RCOMP] = BYTE_TO_FLOAT_TEX( s );
+   texel[GCOMP] = 0.0F;
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_signed_r8(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rgba = (const GLbyte *) texel;
+   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_A8 ***********************************************/
+
+static void FETCH(signed_a8)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   texel[RCOMP] = 0.0F;
+   texel[GCOMP] = 0.0F;
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = BYTE_TO_FLOAT_TEX( s );
+}
+
+#if DIM == 3
+static void store_texel_signed_a8(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rgba = (const GLbyte *) texel;
+   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   *dst = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_L8 ***********************************************/
+
+static void FETCH(signed_l8)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = BYTE_TO_FLOAT_TEX( s );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_signed_l8(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rgba = (const GLbyte *) texel;
+   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_I8 ***********************************************/
+
+static void FETCH(signed_i8)( const struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLbyte s = *TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] =
+   texel[ACOMP] = BYTE_TO_FLOAT_TEX( s );
+}
+
+#if DIM == 3
+static void store_texel_signed_i8(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rgba = (const GLbyte *) texel;
+   GLbyte *dst = TEXEL_ADDR(GLbyte, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_RG88_REV ***********************************************/
+
+static void FETCH(signed_rg88_rev)( const struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s & 0xff) );
+   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 8) );
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_signed_rg88_rev(struct gl_texture_image *texImage,
+                                        GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rg = (const GLbyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   dst[0] = PACK_COLOR_88(rg[GCOMP], rg[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_AL88 ***********************************************/
+
+static void FETCH(signed_al88)( const struct gl_texture_image *texImage,
+                                GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s & 0xff) );
+   texel[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 8) );
+}
+
+#if DIM == 3
+static void store_texel_signed_al88(struct gl_texture_image *texImage,
+                                    GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rg = (const GLbyte *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 1);
+   dst[0] = PACK_COLOR_88(rg[ACOMP], rg[RCOMP]);
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_RGBX8888 ***********************************************/
+
+static void FETCH(signed_rgbx8888)( const struct gl_texture_image *texImage,
+			            GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );
+   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );
+   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );
+   texel[ACOMP] = 1.0f;
+}
+
+#if DIM == 3
+static void store_texel_signed_rgbx8888(struct gl_texture_image *texImage,
+                                        GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rgba = (const GLbyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], 255);
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_RGBA8888 ***********************************************/
+
+static void FETCH(signed_rgba8888)( const struct gl_texture_image *texImage,
+			            GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );
+   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );
+   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );
+   texel[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s      ) );
+}
+
+#if DIM == 3
+static void store_texel_signed_rgba8888(struct gl_texture_image *texImage,
+                                        GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLbyte *rgba = (const GLbyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
+
+static void FETCH(signed_rgba8888_rev)( const struct gl_texture_image *texImage,
+                                        GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint s = *TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   texel[RCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s      ) );
+   texel[GCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >>  8) );
+   texel[BCOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 16) );
+   texel[ACOMP] = BYTE_TO_FLOAT_TEX( (GLbyte) (s >> 24) );
+}
+
+#if DIM == 3
+static void store_texel_signed_rgba8888_rev(struct gl_texture_image *texImage,
+                                            GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLubyte *rgba = (const GLubyte *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = PACK_COLOR_8888_REV(rgba[RCOMP], rgba[GCOMP], rgba[BCOMP], rgba[ACOMP]);
+}
+#endif
+
+
+
+/* MESA_FORMAT_SIGNED_R16 ***********************************************/
+
+static void
+FETCH(signed_r16)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s );
+   texel[GCOMP] = 0.0F;
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void
+store_texel_signed_r16(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   *dst = rgba[0];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_A16 ***********************************************/
+
+static void
+FETCH(signed_a16)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   texel[RCOMP] = 0.0F;
+   texel[GCOMP] = 0.0F;
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s );
+}
+
+#if DIM == 3
+static void
+store_texel_signed_a16(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   *dst = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_L16 ***********************************************/
+
+static void
+FETCH(signed_l16)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void
+store_texel_signed_l16(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_I16 ***********************************************/
+
+static void
+FETCH(signed_i16)(const struct gl_texture_image *texImage,
+                  GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort s = *TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] =
+   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s );
+}
+
+#if DIM == 3
+static void
+store_texel_signed_i16(struct gl_texture_image *texImage,
+                       GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 1);
+   *dst = rgba[RCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_RG1616 ***********************************************/
+
+static void
+FETCH(signed_rg1616)(const struct gl_texture_image *texImage,
+                    GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);
+   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );
+   texel[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void
+store_texel_signed_rg1616(struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_AL1616 ***********************************************/
+
+static void
+FETCH(signed_al1616)(const struct gl_texture_image *texImage,
+                    GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);
+   texel[RCOMP] =
+   texel[GCOMP] =
+   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s[0] );
+   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s[1] );
+}
+
+#if DIM == 3
+static void
+store_texel_signed_al1616(struct gl_texture_image *texImage,
+                         GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 2);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[ACOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_RGB_16 ***********************************************/
+
+static void 
+FETCH(signed_rgb_16)(const struct gl_texture_image *texImage,
+                     GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 3);
+   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );
+   texel[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );
+   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s[2] );
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void
+store_texel_signed_rgb_16(struct gl_texture_image *texImage,
+                          GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 3);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+}
+#endif
+
+
+/* MESA_FORMAT_SIGNED_RGBA_16 ***********************************************/
+
+static void
+FETCH(signed_rgba_16)(const struct gl_texture_image *texImage,
+                      GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLshort *s = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);
+   texel[RCOMP] = SHORT_TO_FLOAT_TEX( s[0] );
+   texel[GCOMP] = SHORT_TO_FLOAT_TEX( s[1] );
+   texel[BCOMP] = SHORT_TO_FLOAT_TEX( s[2] );
+   texel[ACOMP] = SHORT_TO_FLOAT_TEX( s[3] );
+}
+
+#if DIM == 3
+static void
+store_texel_signed_rgba_16(struct gl_texture_image *texImage,
+                           GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLshort *rgba = (const GLshort *) texel;
+   GLshort *dst = TEXEL_ADDR(GLshort, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+
+/* MESA_FORMAT_RGBA_16 ***********************************************/
+
+static void
+FETCH(rgba_16)(const struct gl_texture_image *texImage,
+               GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLushort *s = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);
+   texel[RCOMP] = USHORT_TO_FLOAT( s[0] );
+   texel[GCOMP] = USHORT_TO_FLOAT( s[1] );
+   texel[BCOMP] = USHORT_TO_FLOAT( s[2] );
+   texel[ACOMP] = USHORT_TO_FLOAT( s[3] );
+}
+
+#if DIM == 3
+static void
+store_texel_rgba_16(struct gl_texture_image *texImage,
+                    GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLushort *rgba = (const GLushort *) texel;
+   GLushort *dst = TEXEL_ADDR(GLushort, texImage, i, j, k, 4);
+   dst[0] = rgba[RCOMP];
+   dst[1] = rgba[GCOMP];
+   dst[2] = rgba[BCOMP];
+   dst[3] = rgba[ACOMP];
+}
+#endif
+
+
+
+/* MESA_FORMAT_YCBCR *********************************************************/
+
+/* Fetch texel from 1D, 2D or 3D ycbcr texture, return 4 GLfloats.
+ * We convert YCbCr to RGB here.
+ */
+static void FETCH(f_ycbcr)( const struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src0 = TEXEL_ADDR(GLushort, texImage, (i & ~1), j, k, 1); /* even */
+   const GLushort *src1 = src0 + 1;                               /* odd */
+   const GLubyte y0 = (*src0 >> 8) & 0xff;  /* luminance */
+   const GLubyte cb = *src0 & 0xff;         /* chroma U */
+   const GLubyte y1 = (*src1 >> 8) & 0xff;  /* luminance */
+   const GLubyte cr = *src1 & 0xff;         /* chroma V */
+   const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */
+   GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);
+   GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);
+   GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);
+   r *= (1.0F / 255.0F);
+   g *= (1.0F / 255.0F);
+   b *= (1.0F / 255.0F);
+   texel[RCOMP] = CLAMP(r, 0.0F, 1.0F);
+   texel[GCOMP] = CLAMP(g, 0.0F, 1.0F);
+   texel[BCOMP] = CLAMP(b, 0.0F, 1.0F);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_ycbcr(struct gl_texture_image *texImage,
+                              GLint i, GLint j, GLint k, const void *texel)
+{
+   (void) texImage;
+   (void) i;
+   (void) j;
+   (void) k;
+   (void) texel;
+   /* XXX to do */
+}
+#endif
+
+
+/* MESA_FORMAT_YCBCR_REV *****************************************************/
+
+/* Fetch texel from 1D, 2D or 3D ycbcr_rev texture, return 4 GLfloats.
+ * We convert YCbCr to RGB here.
+ */
+static void FETCH(f_ycbcr_rev)( const struct gl_texture_image *texImage,
+                                GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLushort *src0 = TEXEL_ADDR(GLushort, texImage, (i & ~1), j, k, 1); /* even */
+   const GLushort *src1 = src0 + 1;                               /* odd */
+   const GLubyte y0 = *src0 & 0xff;         /* luminance */
+   const GLubyte cr = (*src0 >> 8) & 0xff;  /* chroma V */
+   const GLubyte y1 = *src1 & 0xff;         /* luminance */
+   const GLubyte cb = (*src1 >> 8) & 0xff;  /* chroma U */
+   const GLubyte y = (i & 1) ? y1 : y0;     /* choose even/odd luminance */
+   GLfloat r = 1.164F * (y - 16) + 1.596F * (cr - 128);
+   GLfloat g = 1.164F * (y - 16) - 0.813F * (cr - 128) - 0.391F * (cb - 128);
+   GLfloat b = 1.164F * (y - 16) + 2.018F * (cb - 128);
+   r *= (1.0F / 255.0F);
+   g *= (1.0F / 255.0F);
+   b *= (1.0F / 255.0F);
+   texel[RCOMP] = CLAMP(r, 0.0F, 1.0F);
+   texel[GCOMP] = CLAMP(g, 0.0F, 1.0F);
+   texel[BCOMP] = CLAMP(b, 0.0F, 1.0F);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_ycbcr_rev(struct gl_texture_image *texImage,
+                                  GLint i, GLint j, GLint k, const void *texel)
+{
+   (void) texImage;
+   (void) i;
+   (void) j;
+   (void) k;
+   (void) texel;
+   /* XXX to do */
+}
+#endif
+
+
+/* MESA_TEXFORMAT_Z24_S8 ***************************************************/
+
+static void FETCH(f_z24_s8)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   /* only return Z, not stencil data */
+   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   const GLfloat scale = 1.0F / (GLfloat) 0xffffff;
+   texel[0] = ((*src) >> 8) * scale;
+   ASSERT(texImage->TexFormat == MESA_FORMAT_Z24_S8);
+   ASSERT(texel[0] >= 0.0F);
+   ASSERT(texel[0] <= 1.0F);
+}
+
+#if DIM == 3
+static void store_texel_z24_s8(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   /* only store Z, not stencil */
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   GLfloat depth = *((GLfloat *) texel);
+   GLuint zi = ((GLuint) (depth * 0xffffff)) << 8;
+   *dst = zi | (*dst & 0xff);
+}
+#endif
+
+
+/* MESA_TEXFORMAT_S8_Z24 ***************************************************/
+
+static void FETCH(f_s8_z24)( const struct gl_texture_image *texImage,
+                             GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   /* only return Z, not stencil data */
+   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   const GLfloat scale = 1.0F / (GLfloat) 0xffffff;
+   texel[0] = ((*src) & 0x00ffffff) * scale;
+   ASSERT(texImage->TexFormat == MESA_FORMAT_S8_Z24);
+   ASSERT(texel[0] >= 0.0F);
+   ASSERT(texel[0] <= 1.0F);
+}
+
+#if DIM == 3
+static void store_texel_s8_z24(struct gl_texture_image *texImage,
+                               GLint i, GLint j, GLint k, const void *texel)
+{
+   /* only store Z, not stencil */
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   GLfloat depth = *((GLfloat *) texel);
+   GLuint zi = (GLuint) (depth * 0xffffff);
+   *dst = zi | (*dst & 0xff000000);
+}
+#endif
+
+
+/* MESA_FORMAT_RGB9_E5 ******************************************************/
+
+static void FETCH(rgb9_e5)( const struct gl_texture_image *texImage,
+                            GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   rgb9e5_to_float3(*src, texel);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_rgb9_e5(struct gl_texture_image *texImage,
+                                GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *src = (const GLfloat *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = float3_to_rgb9e5(src);
+}
+#endif
+
+
+/* MESA_FORMAT_R11_G11_B10_FLOAT *********************************************/
+
+static void FETCH(r11_g11_b10f)( const struct gl_texture_image *texImage,
+                                 GLint i, GLint j, GLint k, GLfloat *texel )
+{
+   const GLuint *src = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   r11g11b10f_to_float3(*src, texel);
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_r11_g11_b10f(struct gl_texture_image *texImage,
+                                     GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *src = (const GLfloat *) texel;
+   GLuint *dst = TEXEL_ADDR(GLuint, texImage, i, j, k, 1);
+   *dst = float3_to_r11g11b10f(src);
+}
+#endif
+
+
+/* MESA_FORMAT_Z32_FLOAT_X24S8 ***********************************************/
+
+static void FETCH(z32f_x24s8)(const struct gl_texture_image *texImage,
+			      GLint i, GLint j, GLint k, GLfloat *texel)
+{
+   const GLfloat *src = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);
+   texel[RCOMP] = src[0];
+   texel[GCOMP] = 0.0F;
+   texel[BCOMP] = 0.0F;
+   texel[ACOMP] = 1.0F;
+}
+
+#if DIM == 3
+static void store_texel_z32f_x24s8(struct gl_texture_image *texImage,
+                                   GLint i, GLint j, GLint k, const void *texel)
+{
+   const GLfloat *src = (const GLfloat *) texel;
+   GLfloat *dst = TEXEL_ADDR(GLfloat, texImage, i, j, k, 2);
+   dst[0] = src[0];
+}
+#endif
+
+
+#undef TEXEL_ADDR
+#undef DIM
+#undef FETCH
diff --git a/mesalib/src/mesa/main/texformat.c b/mesalib/src/mesa/main/texformat.c
index 8cbb021d8..c919a74e0 100644
--- a/mesalib/src/mesa/main/texformat.c
+++ b/mesalib/src/mesa/main/texformat.c
@@ -416,6 +416,19 @@ _mesa_choose_tex_format( struct gl_context *ctx, GLint internalFormat,
       }
    }
 
+   if (ctx->Extensions.ARB_depth_buffer_float) {
+      switch (internalFormat) {
+         case GL_DEPTH_COMPONENT32F:
+            ASSERT(ctx->TextureFormatSupported[MESA_FORMAT_Z32_FLOAT]);
+            return MESA_FORMAT_Z32_FLOAT;
+         case GL_DEPTH32F_STENCIL8:
+            ASSERT(ctx->TextureFormatSupported[MESA_FORMAT_Z32_FLOAT_X24S8]);
+            return MESA_FORMAT_Z32_FLOAT_X24S8;
+         default:
+            ; /* fallthrough */
+      }
+   }
+
    if (ctx->Extensions.ATI_envmap_bumpmap) {
       switch (internalFormat) {
          case GL_DUDV_ATI:
diff --git a/mesalib/src/mesa/main/texobj.c b/mesalib/src/mesa/main/texobj.c
index 565a3a2d8..0e84b8744 100644
--- a/mesalib/src/mesa/main/texobj.c
+++ b/mesalib/src/mesa/main/texobj.c
@@ -899,7 +899,7 @@ unbind_texobj_from_texunits(struct gl_context *ctx,
 {
    GLuint u, tex;
 
-   for (u = 0; u < MAX_TEXTURE_IMAGE_UNITS; u++) {
+   for (u = 0; u < Elements(ctx->Texture.Unit); u++) {
       struct gl_texture_unit *unit = &ctx->Texture.Unit[u];
       for (tex = 0; tex < NUM_TEXTURE_TARGETS; tex++) {
          if (texObj == unit->CurrentTex[tex]) {
diff --git a/mesalib/src/mesa/main/texstore.c b/mesalib/src/mesa/main/texstore.c
index e527981ff..6e1e63bdf 100644
--- a/mesalib/src/mesa/main/texstore.c
+++ b/mesalib/src/mesa/main/texstore.c
@@ -1002,15 +1002,17 @@ memcpy_texture(struct gl_context *ctx,
 
 
 /**
- * Store a 32-bit integer depth component texture image.
+ * Store a 32-bit integer or float depth component texture image.
  */
 static GLboolean
 _mesa_texstore_z32(TEXSTORE_PARAMS)
 {
    const GLuint depthScale = 0xffffffff;
    const GLuint texelBytes = _mesa_get_format_bytes(dstFormat);
+   const GLenum dstType = _mesa_get_format_datatype(dstFormat);
    (void) dims;
-   ASSERT(dstFormat == MESA_FORMAT_Z32);
+   ASSERT(dstFormat == MESA_FORMAT_Z32 ||
+          dstFormat == MESA_FORMAT_Z32_FLOAT);
    ASSERT(texelBytes == sizeof(GLuint));
 
    if (ctx->Pixel.DepthScale == 1.0f &&
@@ -1018,7 +1020,7 @@ _mesa_texstore_z32(TEXSTORE_PARAMS)
        !srcPacking->SwapBytes &&
        baseInternalFormat == GL_DEPTH_COMPONENT &&
        srcFormat == GL_DEPTH_COMPONENT &&
-       srcType == GL_UNSIGNED_INT) {
+       srcType == dstType) {
       /* simple memcpy path */
       memcpy_texture(ctx, dims,
                      dstFormat, dstAddr, dstXoffset, dstYoffset, dstZoffset,
@@ -1039,7 +1041,7 @@ _mesa_texstore_z32(TEXSTORE_PARAMS)
             const GLvoid *src = _mesa_image_address(dims, srcPacking,
                 srcAddr, srcWidth, srcHeight, srcFormat, srcType, img, row, 0);
             _mesa_unpack_depth_span(ctx, srcWidth,
-                                    GL_UNSIGNED_INT, (GLuint *) dstRow,
+                                    dstType, dstRow,
                                     depthScale, srcType, src, srcPacking);
             dstRow += dstRowStride;
          }
@@ -4285,6 +4287,72 @@ _mesa_texstore_r11_g11_b10f(TEXSTORE_PARAMS)
 }
 
 
+static GLboolean
+_mesa_texstore_z32f_x24s8(TEXSTORE_PARAMS)
+{
+   ASSERT(dstFormat == MESA_FORMAT_Z32_FLOAT_X24S8);
+   ASSERT(srcFormat == GL_DEPTH_STENCIL ||
+          srcFormat == GL_DEPTH_COMPONENT ||
+          srcFormat == GL_STENCIL_INDEX);
+   ASSERT(srcFormat != GL_DEPTH_STENCIL ||
+          srcType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
+
+   if (srcFormat == GL_DEPTH_STENCIL &&
+       ctx->Pixel.DepthScale == 1.0f &&
+       ctx->Pixel.DepthBias == 0.0f &&
+       !srcPacking->SwapBytes) {
+      /* simple path */
+      memcpy_texture(ctx, dims,
+                     dstFormat, dstAddr, dstXoffset, dstYoffset, dstZoffset,
+                     dstRowStride,
+                     dstImageOffsets,
+                     srcWidth, srcHeight, srcDepth, srcFormat, srcType,
+                     srcAddr, srcPacking);
+   }
+   else if (srcFormat == GL_DEPTH_COMPONENT ||
+            srcFormat == GL_STENCIL_INDEX) {
+      GLint img, row;
+      const GLint srcRowStride
+         = _mesa_image_row_stride(srcPacking, srcWidth, srcFormat, srcType)
+         / sizeof(uint64_t);
+
+      /* In case we only upload depth we need to preserve the stencil */
+      for (img = 0; img < srcDepth; img++) {
+         uint64_t *dstRow = (uint64_t *) dstAddr
+            + dstImageOffsets[dstZoffset + img]
+            + dstYoffset * dstRowStride / sizeof(uint64_t)
+            + dstXoffset;
+         const uint64_t *src
+            = (const uint64_t *) _mesa_image_address(dims, srcPacking, srcAddr,
+                  srcWidth, srcHeight,
+                  srcFormat, srcType,
+                  img, 0, 0);
+         for (row = 0; row < srcHeight; row++) {
+            /* The unpack functions with:
+             *    dstType = GL_FLOAT_32_UNSIGNED_INT_24_8_REV
+             * only write their own dword, so the other dword (stencil
+             * or depth) is preserved. */
+            if (srcFormat != GL_STENCIL_INDEX)
+               _mesa_unpack_depth_span(ctx, srcWidth,
+                                       GL_FLOAT_32_UNSIGNED_INT_24_8_REV, /* dst type */
+                                       dstRow, /* dst addr */
+                                       1.0f, srcType, src, srcPacking);
+
+            if (srcFormat != GL_DEPTH_COMPONENT)
+               _mesa_unpack_stencil_span(ctx, srcWidth,
+                                         GL_FLOAT_32_UNSIGNED_INT_24_8_REV, /* dst type */
+                                         dstRow, /* dst addr */
+                                         srcType, src, srcPacking,
+                                         ctx->_ImageTransferState);
+
+            src += srcRowStride;
+            dstRow += dstRowStride / sizeof(uint64_t);
+         }
+      }
+   }
+   return GL_TRUE;
+}
+
 
 /**
  * Table mapping MESA_FORMAT_* to _mesa_texstore_*()
@@ -4419,6 +4487,9 @@ texstore_funcs[MESA_FORMAT_COUNT] =
 
    { MESA_FORMAT_RGB9_E5_FLOAT, _mesa_texstore_rgb9_e5 },
    { MESA_FORMAT_R11_G11_B10_FLOAT, _mesa_texstore_r11_g11_b10f },
+
+   { MESA_FORMAT_Z32_FLOAT, _mesa_texstore_z32 },
+   { MESA_FORMAT_Z32_FLOAT_X24S8, _mesa_texstore_z32f_x24s8 },
 };
 
 
diff --git a/mesalib/src/mesa/main/uniforms.c b/mesalib/src/mesa/main/uniforms.c
index 1c4fd82ba..dd069a3a4 100644
--- a/mesalib/src/mesa/main/uniforms.c
+++ b/mesalib/src/mesa/main/uniforms.c
@@ -580,7 +580,7 @@ _mesa_update_shader_textures_used(struct gl_program *prog)
       if (prog->SamplersUsed & (1 << s)) {
          GLuint unit = prog->SamplerUnits[s];
          GLuint tgt = prog->SamplerTargets[s];
-         assert(unit < MAX_TEXTURE_IMAGE_UNITS);
+         assert(unit < Elements(prog->TexturesUsed));
          assert(tgt < NUM_TEXTURE_TARGETS);
          prog->TexturesUsed[unit] |= (1 << tgt);
       }
@@ -674,7 +674,7 @@ set_program_uniform(struct gl_context *ctx, struct gl_program *program,
          GLuint texUnit = ((GLuint *) values)[i];
 
          /* check that the sampler (tex unit index) is legal */
-         if (texUnit >= ctx->Const.MaxTextureImageUnits) {
+         if (texUnit >= ctx->Const.MaxCombinedTextureImageUnits) {
             _mesa_error(ctx, GL_INVALID_VALUE,
                         "glUniform1(invalid sampler/tex unit index for '%s')",
                         param->Name);
diff --git a/mesalib/src/mesa/state_tracker/st_cb_clear.c b/mesalib/src/mesa/state_tracker/st_cb_clear.c
index 181fedd2b..117000ba7 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_clear.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_clear.c
@@ -381,7 +381,8 @@ check_clear_depth_stencil_with_quad(struct gl_context *ctx, struct gl_renderbuff
 
    assert(rb->Format == MESA_FORMAT_S8 ||
           rb->Format == MESA_FORMAT_Z24_S8 ||
-          rb->Format == MESA_FORMAT_S8_Z24);
+          rb->Format == MESA_FORMAT_S8_Z24 ||
+          rb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
 
    if (ctx->Scissor.Enabled &&
        (ctx->Scissor.X != 0 ||
@@ -436,7 +437,8 @@ check_clear_stencil_with_quad(struct gl_context *ctx, struct gl_renderbuffer *rb
 
    assert(rb->Format == MESA_FORMAT_S8 ||
           rb->Format == MESA_FORMAT_Z24_S8 ||
-          rb->Format == MESA_FORMAT_S8_Z24);
+          rb->Format == MESA_FORMAT_S8_Z24 ||
+          rb->Format == MESA_FORMAT_Z32_FLOAT_X24S8);
 
    if (maskStencil) 
       return GL_TRUE;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
index d61d7ac22..dca332464 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
@@ -812,6 +812,7 @@ draw_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
       for (row = 0; row < height; row++) {
          GLubyte sValues[MAX_WIDTH];
          GLuint zValues[MAX_WIDTH];
+         GLfloat *zValuesFloat = (GLfloat*)zValues;
          GLenum destType = GL_UNSIGNED_BYTE;
          const GLvoid *source = _mesa_image_address2d(&clippedUnpack, pixels,
                                                       width, height,
@@ -822,7 +823,11 @@ draw_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
                                    ctx->_ImageTransferState);
 
          if (format == GL_DEPTH_STENCIL) {
-            _mesa_unpack_depth_span(ctx, spanWidth, GL_UNSIGNED_INT, zValues,
+            GLenum ztype =
+               pt->resource->format == PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED ?
+               GL_FLOAT : GL_UNSIGNED_INT;
+
+            _mesa_unpack_depth_span(ctx, spanWidth, ztype, zValues,
                                     (1 << 24) - 1, type, source,
                                     &clippedUnpack);
          }
@@ -887,6 +892,26 @@ draw_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
                   }
                }
                break;
+            case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+               if (format == GL_DEPTH_STENCIL) {
+                  uint *dest = (uint *) (stmap + spanY * pt->stride + spanX*4);
+                  GLfloat *destf = (GLfloat*)dest;
+                  GLint k;
+                  assert(usage == PIPE_TRANSFER_WRITE);
+                  for (k = 0; k < spanWidth; k++) {
+                     destf[k*2] = zValuesFloat[k];
+                     dest[k*2+1] = sValues[k] & 0xff;
+                  }
+               }
+               else {
+                  uint *dest = (uint *) (stmap + spanY * pt->stride + spanX*4);
+                  GLint k;
+                  assert(usage == PIPE_TRANSFER_READ_WRITE);
+                  for (k = 0; k < spanWidth; k++) {
+                     dest[k*2+1] = sValues[k] & 0xff;
+                  }
+               }
+               break;
             default:
                assert(0);
             }
@@ -994,14 +1019,23 @@ st_DrawPixels(struct gl_context *ctx, GLint x, GLint y,
                                     GL_NONE, GL_NONE,
                                     PIPE_TEXTURE_2D,
 				    0, PIPE_BIND_SAMPLER_VIEW);
-      if (tex_format == PIPE_FORMAT_Z24_UNORM_S8_USCALED)
-	 stencil_format = PIPE_FORMAT_X24S8_USCALED;
-      else if (tex_format == PIPE_FORMAT_S8_USCALED_Z24_UNORM)
-	 stencil_format = PIPE_FORMAT_S8X24_USCALED;
-      else
-	 stencil_format = PIPE_FORMAT_S8_USCALED;
-      if (stencil_format == PIPE_FORMAT_NONE)
-	 goto stencil_fallback;
+
+      switch (tex_format) {
+      case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+         stencil_format = PIPE_FORMAT_X24S8_USCALED;
+         break;
+      case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+         stencil_format = PIPE_FORMAT_S8X24_USCALED;
+         break;
+      case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+         stencil_format = PIPE_FORMAT_X32_S8X24_USCALED;
+         break;
+      case PIPE_FORMAT_S8_USCALED:
+         stencil_format = PIPE_FORMAT_S8_USCALED;
+         break;
+      default:
+         goto stencil_fallback;
+      }
    }
 
    /* Mesa state should be up to date by now */
@@ -1188,6 +1222,18 @@ copy_stencil_pixels(struct gl_context *ctx, GLint srcx, GLint srcy,
          assert(usage == PIPE_TRANSFER_WRITE);
          memcpy(dst, src, width);
          break;
+      case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+         {
+            uint *dst4 = (uint *) dst;
+            int j;
+            dst4++;
+            assert(usage == PIPE_TRANSFER_READ_WRITE);
+            for (j = 0; j < width; j++) {
+               *dst4 = src[j] & 0xff;
+               dst4 += 2;
+            }
+         }
+         break;
       default:
          assert(0);
       }
diff --git a/mesalib/src/mesa/state_tracker/st_cb_readpixels.c b/mesalib/src/mesa/state_tracker/st_cb_readpixels.c
index 67926e392..e2b29fe30 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_readpixels.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_readpixels.c
@@ -151,6 +151,24 @@ st_read_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
             }
          }
          break;
+      case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+         if (format == GL_DEPTH_STENCIL) {
+            const uint *src = (uint *) (stmap + srcY * pt->stride);
+            const GLfloat *srcf = (const GLfloat*)src;
+            GLint k;
+            for (k = 0; k < width; k++) {
+               zValues[k] = srcf[k*2];
+               sValues[k] = src[k*2+1] & 0xff;
+            }
+         }
+         else {
+            const uint *src = (uint *) (stmap + srcY * pt->stride);
+            GLint k;
+            for (k = 0; k < width; k++) {
+               sValues[k] = src[k*2+1] & 0xff;
+            }
+         }
+         break;
       default:
          assert(0);
       }
@@ -159,7 +177,7 @@ st_read_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
       dest = _mesa_image_address2d(packing, pixels, width, height,
                                    format, type, j, 0);
       if (format == GL_DEPTH_STENCIL) {
-         _mesa_pack_depth_stencil_span(ctx, width, dest,
+         _mesa_pack_depth_stencil_span(ctx, width, type, dest,
                                        zValues, sValues, packing);
       }
       else {
@@ -568,6 +586,31 @@ st_readpixels(struct gl_context *ctx, GLint x, GLint y, GLsizei width, GLsizei h
             dst += dstStride;
          }
       }
+      else if (pformat == PIPE_FORMAT_Z32_FLOAT) {
+         for (i = 0; i < height; i++) {
+            GLfloat zfloat[MAX_WIDTH];
+            pipe_get_tile_raw(pipe, trans, 0, y, width, 1, zfloat, 0);
+            y += yStep;
+            _mesa_pack_depth_span(ctx, width, dst, type,
+                                  zfloat, &clippedPacking);
+            dst += dstStride;
+         }
+      }
+      else if (pformat == PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED) {
+         assert(format == GL_DEPTH_COMPONENT);
+         for (i = 0; i < height; i++) {
+            GLfloat zfloat[MAX_WIDTH];    /* Z32 */
+            GLfloat zfloat2[MAX_WIDTH*2]; /* Z32X32 */
+            pipe_get_tile_raw(pipe, trans, 0, y, width, 1, zfloat2, 0);
+            y += yStep;
+            for (j = 0; j < width; j++) {
+               zfloat[j] = zfloat2[j*2];
+            }
+            _mesa_pack_depth_span(ctx, width, dst, type,
+                                  zfloat, &clippedPacking);
+            dst += dstStride;
+         }
+      }
       else {
          /* RGBA format */
          /* Do a row at a time to flip image data vertically */
diff --git a/mesalib/src/mesa/state_tracker/st_extensions.c b/mesalib/src/mesa/state_tracker/st_extensions.c
index d3aebe526..99b231d97 100644
--- a/mesalib/src/mesa/state_tracker/st_extensions.c
+++ b/mesalib/src/mesa/state_tracker/st_extensions.c
@@ -607,4 +607,15 @@ void st_init_extensions(struct st_context *st)
    if (screen->get_param(screen, PIPE_CAP_SM3)) {
       ctx->Extensions.ARB_shader_texture_lod = GL_TRUE;
    }
+
+   if (screen->is_format_supported(screen, PIPE_FORMAT_Z32_FLOAT,
+                                   PIPE_TEXTURE_2D, 0,
+                                   PIPE_BIND_DEPTH_STENCIL |
+                                   PIPE_BIND_SAMPLER_VIEW) &&
+       screen->is_format_supported(screen, PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED,
+                                   PIPE_TEXTURE_2D, 0,
+                                   PIPE_BIND_DEPTH_STENCIL |
+                                   PIPE_BIND_SAMPLER_VIEW)) {
+      ctx->Extensions.ARB_depth_buffer_float = GL_TRUE;
+   }
 }
diff --git a/mesalib/src/mesa/state_tracker/st_format.c b/mesalib/src/mesa/state_tracker/st_format.c
index fa5d8f505..bd4f0860c 100644
--- a/mesalib/src/mesa/state_tracker/st_format.c
+++ b/mesalib/src/mesa/state_tracker/st_format.c
@@ -68,42 +68,70 @@ GLenum
 st_format_datatype(enum pipe_format format)
 {
    const struct util_format_description *desc;
+   int i;
 
    desc = util_format_description(format);
    assert(desc);
 
+   /* Find the first non-VOID channel. */
+   for (i = 0; i < 4; i++) {
+       if (desc->channel[i].type != UTIL_FORMAT_TYPE_VOID) {
+           break;
+       }
+   }
+
    if (desc->layout == UTIL_FORMAT_LAYOUT_PLAIN) {
       if (format == PIPE_FORMAT_B5G5R5A1_UNORM ||
           format == PIPE_FORMAT_B5G6R5_UNORM) {
          return GL_UNSIGNED_SHORT;
       }
+      else if (format == PIPE_FORMAT_R11G11B10_FLOAT ||
+               format == PIPE_FORMAT_R9G9B9E5_FLOAT) {
+         return GL_FLOAT;
+      }
       else if (format == PIPE_FORMAT_Z24_UNORM_S8_USCALED ||
                format == PIPE_FORMAT_S8_USCALED_Z24_UNORM ||
                format == PIPE_FORMAT_Z24X8_UNORM ||
                format == PIPE_FORMAT_X8Z24_UNORM) {
          return GL_UNSIGNED_INT_24_8;
       }
+      else if (format == PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED) {
+         return GL_FLOAT_32_UNSIGNED_INT_24_8_REV;
+      }
       else {
          const GLuint size = format_max_bits(format);
+
+         assert(i < 4);
+         if (i == 4)
+            return GL_NONE;
+
          if (size == 8) {
-            if (desc->channel[0].type == UTIL_FORMAT_TYPE_UNSIGNED)
+            if (desc->channel[i].type == UTIL_FORMAT_TYPE_UNSIGNED)
                return GL_UNSIGNED_BYTE;
             else
                return GL_BYTE;
          }
          else if (size == 16) {
-            if (desc->channel[0].type == UTIL_FORMAT_TYPE_UNSIGNED)
+            if (desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT)
+               return GL_HALF_FLOAT;
+            if (desc->channel[i].type == UTIL_FORMAT_TYPE_UNSIGNED)
                return GL_UNSIGNED_SHORT;
             else
                return GL_SHORT;
          }
-         else {
-            assert( size <= 32 );
-            if (desc->channel[0].type == UTIL_FORMAT_TYPE_UNSIGNED)
+         else if (size <= 32) {
+            if (desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT)
+               return GL_FLOAT;
+            if (desc->channel[i].type == UTIL_FORMAT_TYPE_UNSIGNED)
                return GL_UNSIGNED_INT;
             else
                return GL_INT;
          }
+         else {
+            assert(size == 64);
+            assert(desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT);
+            return GL_DOUBLE;
+         }
       }
    }
    else if (format == PIPE_FORMAT_UYVY) {
@@ -180,6 +208,10 @@ st_mesa_format_to_pipe_format(gl_format mesaFormat)
       return PIPE_FORMAT_Z24X8_UNORM;
    case MESA_FORMAT_S8:
       return PIPE_FORMAT_S8_USCALED;
+   case MESA_FORMAT_Z32_FLOAT:
+      return PIPE_FORMAT_Z32_FLOAT;
+   case MESA_FORMAT_Z32_FLOAT_X24S8:
+      return PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED;
    case MESA_FORMAT_YCBCR:
       return PIPE_FORMAT_UYVY;
 #if FEATURE_texture_s3tc
@@ -402,6 +434,10 @@ st_pipe_format_to_mesa_format(enum pipe_format format)
       return MESA_FORMAT_X8_Z24;
    case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
       return MESA_FORMAT_S8_Z24;
+   case PIPE_FORMAT_Z32_FLOAT:
+      return MESA_FORMAT_Z32_FLOAT;
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      return MESA_FORMAT_Z32_FLOAT_X24S8;
 
    case PIPE_FORMAT_UYVY:
       return MESA_FORMAT_YCBCR;
@@ -759,6 +795,10 @@ static const struct format_mapping format_map[] = {
       { GL_DEPTH_COMPONENT, 0 },
       { DEFAULT_DEPTH_FORMATS }
    },
+   {
+      { GL_DEPTH_COMPONENT32F, 0 },
+      { PIPE_FORMAT_Z32_FLOAT, 0 }
+   },
 
    /* stencil formats */
    {
@@ -775,6 +815,10 @@ static const struct format_mapping format_map[] = {
       { GL_DEPTH_STENCIL_EXT, GL_DEPTH24_STENCIL8_EXT, 0 },
       { PIPE_FORMAT_Z24_UNORM_S8_USCALED, PIPE_FORMAT_S8_USCALED_Z24_UNORM, 0 }
    },
+   {
+      { GL_DEPTH32F_STENCIL8, 0 },
+      { PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED, 0 }
+   },
 
    /* sRGB formats */
    {
diff --git a/mesalib/src/mesa/state_tracker/st_manager.c b/mesalib/src/mesa/state_tracker/st_manager.c
index a8c4b5c3f..7bd82aae2 100644
--- a/mesalib/src/mesa/state_tracker/st_manager.c
+++ b/mesalib/src/mesa/state_tracker/st_manager.c
@@ -918,6 +918,15 @@ st_manager_add_color_renderbuffer(struct st_context *st,
       return FALSE;
 
    st_framebuffer_update_attachments(stfb);
+
+   /*
+    * Force a call to the state tracker manager to validate the
+    * new renderbuffer. It might be that there is a window system
+    * renderbuffer available.
+    */
+   if(stfb->iface)
+      stfb->iface_stamp = p_atomic_read(&stfb->iface->stamp) - 1;
+
    st_invalidate_state(st->ctx, _NEW_BUFFERS);
 
    return TRUE;
diff --git a/mesalib/src/mesa/swrast/s_blit.c b/mesalib/src/mesa/swrast/s_blit.c
index 3516a41bf..7f53f19eb 100644
--- a/mesalib/src/mesa/swrast/s_blit.c
+++ b/mesalib/src/mesa/swrast/s_blit.c
@@ -568,9 +568,6 @@ _swrast_BlitFramebuffer(struct gl_context *ctx,
    };
    GLint i;
 
-   if (!ctx->DrawBuffer->_NumColorDrawBuffers)
-      return;
-
    if (!_mesa_clip_blit(ctx, &srcX0, &srcY0, &srcX1, &srcY1,
                         &dstX0, &dstY0, &dstX1, &dstY1)) {
       return;
diff --git a/mesalib/src/mesa/swrast/s_readpix.c b/mesalib/src/mesa/swrast/s_readpix.c
index 214f2ea1a..66ca39293 100644
--- a/mesalib/src/mesa/swrast/s_readpix.c
+++ b/mesalib/src/mesa/swrast/s_readpix.c
@@ -446,7 +446,7 @@ read_depth_stencil_pixels(struct gl_context *ctx,
             GLfloat depthVals[MAX_WIDTH];
             _swrast_read_depth_span_float(ctx, depthRb, width, x, y + i,
                                           depthVals);
-            _mesa_pack_depth_stencil_span(ctx, width, depthStencilDst,
+            _mesa_pack_depth_stencil_span(ctx, width, type, depthStencilDst,
                                           depthVals, stencilVals, packing);
          }
       }