util: Add dedicated depth-stencil packing/unpacking functions.

Depth-stencil manually written given that each one is very close to
be a special case.

u_format_zs.c's still untested.

15 files changed, 1437 insertions, 261 deletions

diff --git a/src/gallium/auxiliary/Makefile b/src/gallium/auxiliary/Makefile
index 8d0ed67d94..050c7043c3 100644
--- a/src/gallium/auxiliary/Makefile
+++ b/src/gallium/auxiliary/Makefile
@@ -112,6 +112,7 @@ C_SOURCES = \
 	util/u_format_table.c \
 	util/u_format_tests.c \
 	util/u_format_yuv.c \
+	util/u_format_zs.c \
 	util/u_gen_mipmap.c \
 	util/u_half.c \
 	util/u_handle_table.c \
diff --git a/src/gallium/auxiliary/SConscript b/src/gallium/auxiliary/SConscript
index cadf246058..c23fbdf70a 100644
--- a/src/gallium/auxiliary/SConscript
+++ b/src/gallium/auxiliary/SConscript
@@ -161,6 +161,7 @@ source = [
     'util/u_format_table.c',
     'util/u_format_tests.c',
     'util/u_format_yuv.c',
+    'util/u_format_zs.c',
     'util/u_gen_mipmap.c',
     'util/u_half.c',
     'util/u_handle_table.c',
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_format_soa.c b/src/gallium/auxiliary/gallivm/lp_bld_format_soa.c
index 459a6bfe3f..2b66162eb4 100644
--- a/src/gallium/auxiliary/gallivm/lp_bld_format_soa.c
+++ b/src/gallium/auxiliary/gallivm/lp_bld_format_soa.c
@@ -307,7 +307,7 @@ lp_build_fetch_rgba_soa(LLVMBuilderRef builder,
    }
    else {
       /*
-       * Fallback to calling util_format_description::fetch_float for each
+       * Fallback to calling util_format_description::fetch_rgba_float for each
        * pixel.
        *
        * This is definitely not the most efficient way of fetching pixels, as
@@ -324,10 +324,10 @@ lp_build_fetch_rgba_soa(LLVMBuilderRef builder,
 
       assert(type.floating);
 
-      util_snprintf(name, sizeof name, "util_format_%s_fetch_float", format_desc->short_name);
+      util_snprintf(name, sizeof name, "util_format_%s_fetch_rgba_float", format_desc->short_name);
 
       /*
-       * Declare and bind format_desc->fetch_float().
+       * Declare and bind format_desc->fetch_rgba_float().
        */
 
       function = LLVMGetNamedFunction(module, name);
@@ -348,7 +348,7 @@ lp_build_fetch_rgba_soa(LLVMBuilderRef builder,
 
          assert(LLVMIsDeclaration(function));
 
-         LLVMAddGlobalMapping(lp_build_engine, function, format_desc->fetch_float);
+         LLVMAddGlobalMapping(lp_build_engine, function, format_desc->fetch_rgba_float);
       }
 
       for (chan = 0; chan < 4; ++chan) {
@@ -361,7 +361,7 @@ lp_build_fetch_rgba_soa(LLVMBuilderRef builder,
                                  "");
 
       /*
-       * Invoke format_desc->fetch_float() for each pixel and insert the result
+       * Invoke format_desc->fetch_rgba_float() for each pixel and insert the result
        * in the SoA vectors.
        */
 
diff --git a/src/gallium/auxiliary/util/u_format.c b/src/gallium/auxiliary/util/u_format.c
index 418b354a96..663325764e 100644
--- a/src/gallium/auxiliary/util/u_format.c
+++ b/src/gallium/auxiliary/util/u_format.c
@@ -56,7 +56,7 @@ util_format_read_4f(enum pipe_format format,
    src_row = (const uint8_t *)src + y*src_stride + x*(format_desc->block.bits/8);
    dst_row = dst;
 
-   format_desc->unpack_float(dst_row, dst_stride, src_row, src_stride, w, h);
+   format_desc->unpack_rgba_float(dst_row, dst_stride, src_row, src_stride, w, h);
 }
 
 
@@ -78,7 +78,7 @@ util_format_write_4f(enum pipe_format format,
    dst_row = (uint8_t *)dst + y*dst_stride + x*(format_desc->block.bits/8);
    src_row = src;
 
-   format_desc->pack_float(dst_row, dst_stride, src_row, src_stride, w, h);
+   format_desc->pack_rgba_float(dst_row, dst_stride, src_row, src_stride, w, h);
 }
 
 
@@ -97,7 +97,7 @@ util_format_read_4ub(enum pipe_format format, uint8_t *dst, unsigned dst_stride,
    src_row = (const uint8_t *)src + y*src_stride + x*(format_desc->block.bits/8);
    dst_row = dst;
 
-   format_desc->unpack_8unorm(dst_row, dst_stride, src_row, src_stride, w, h);
+   format_desc->unpack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, w, h);
 }
 
 
@@ -116,7 +116,7 @@ util_format_write_4ub(enum pipe_format format, const uint8_t *src, unsigned src_
    dst_row = (uint8_t *)dst + y*dst_stride + x*(format_desc->block.bits/8);
    src_row = src;
 
-   format_desc->pack_8unorm(dst_row, dst_stride, src_row, src_stride, w, h);
+   format_desc->pack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, w, h);
 }
 
 
@@ -242,8 +242,8 @@ util_format_translate(enum pipe_format dst_format,
          return;
 
       while (height >= y_step) {
-         src_format_desc->unpack_8unorm(tmp_row, tmp_stride, src_row, src_stride, width, y_step);
-         dst_format_desc->pack_8unorm(dst_row, dst_stride, tmp_row, tmp_stride, width, y_step);
+         src_format_desc->unpack_rgba_8unorm(tmp_row, tmp_stride, src_row, src_stride, width, y_step);
+         dst_format_desc->pack_rgba_8unorm(dst_row, dst_stride, tmp_row, tmp_stride, width, y_step);
 
          dst_row += dst_step;
          src_row += src_step;
@@ -251,8 +251,8 @@ util_format_translate(enum pipe_format dst_format,
       }
 
       if (height) {
-         src_format_desc->unpack_8unorm(tmp_row, tmp_stride, src_row, src_stride, width, height);
-         dst_format_desc->pack_8unorm(dst_row, dst_stride, tmp_row, tmp_stride, width, height);
+         src_format_desc->unpack_rgba_8unorm(tmp_row, tmp_stride, src_row, src_stride, width, height);
+         dst_format_desc->pack_rgba_8unorm(dst_row, dst_stride, tmp_row, tmp_stride, width, height);
       }
 
       FREE(tmp_row);
@@ -267,8 +267,8 @@ util_format_translate(enum pipe_format dst_format,
          return;
 
       while (height >= y_step) {
-         src_format_desc->unpack_float(tmp_row, tmp_stride, src_row, src_stride, width, y_step);
-         dst_format_desc->pack_float(dst_row, dst_stride, tmp_row, tmp_stride, width, y_step);
+         src_format_desc->unpack_rgba_float(tmp_row, tmp_stride, src_row, src_stride, width, y_step);
+         dst_format_desc->pack_rgba_float(dst_row, dst_stride, tmp_row, tmp_stride, width, y_step);
 
          dst_row += dst_step;
          src_row += src_step;
@@ -276,8 +276,8 @@ util_format_translate(enum pipe_format dst_format,
       }
 
       if (height) {
-         src_format_desc->unpack_float(tmp_row, tmp_stride, src_row, src_stride, width, height);
-         dst_format_desc->pack_float(dst_row, dst_stride, tmp_row, tmp_stride, width, height);
+         src_format_desc->unpack_rgba_float(tmp_row, tmp_stride, src_row, src_stride, width, height);
+         dst_format_desc->pack_rgba_float(dst_row, dst_stride, tmp_row, tmp_stride, width, height);
       }
 
       FREE(tmp_row);
diff --git a/src/gallium/auxiliary/util/u_format.h b/src/gallium/auxiliary/util/u_format.h
index a7c24f0c00..c96a39dc09 100644
--- a/src/gallium/auxiliary/util/u_format.h
+++ b/src/gallium/auxiliary/util/u_format.h
@@ -192,43 +192,114 @@ struct util_format_description
 
    /**
     * Unpack pixel blocks to R8G8B8A8_UNORM.
+    *
+    * Only defined for non-depth-stencil formats.
     */
    void
-   (*unpack_8unorm)(uint8_t *dst, unsigned dst_stride,
-                    const uint8_t *src, unsigned src_stride,
-                    unsigned width, unsigned height);
+   (*unpack_rgba_8unorm)(uint8_t *dst, unsigned dst_stride,
+                         const uint8_t *src, unsigned src_stride,
+                         unsigned width, unsigned height);
 
    /**
     * Pack pixel blocks from R8G8B8A8_UNORM.
+    *
+    * Only defined for non-depth-stencil formats.
     */
    void
-   (*pack_8unorm)(uint8_t *dst, unsigned dst_stride,
-                  const uint8_t *src, unsigned src_stride,
-                  unsigned width, unsigned height);
+   (*pack_rgba_8unorm)(uint8_t *dst, unsigned dst_stride,
+                       const uint8_t *src, unsigned src_stride,
+                       unsigned width, unsigned height);
 
    /**
     * Unpack pixel blocks to R32G32B32A32_FLOAT.
+    *
+    * Only defined for non-depth-stencil formats.
     */
    void
-   (*unpack_float)(float *dst, unsigned dst_stride,
-                   const uint8_t *src, unsigned src_stride,
-                   unsigned width, unsigned height);
+   (*unpack_rgba_float)(float *dst, unsigned dst_stride,
+                        const uint8_t *src, unsigned src_stride,
+                        unsigned width, unsigned height);
 
    /**
     * Pack pixel blocks from R32G32B32A32_FLOAT.
+    *
+    * Only defined for non-depth-stencil formats.
     */
    void
-   (*pack_float)(uint8_t *dst, unsigned dst_stride,
-                 const float *src, unsigned src_stride,
-                 unsigned width, unsigned height);
+   (*pack_rgba_float)(uint8_t *dst, unsigned dst_stride,
+                      const float *src, unsigned src_stride,
+                      unsigned width, unsigned height);
 
    /**
     * Fetch a single pixel (i, j) from a block.
+    *
+    * Only defined for non-depth-stencil formats.
+    */
+   void
+   (*fetch_rgba_float)(float *dst,
+                       const uint8_t *src,
+                       unsigned i, unsigned j);
+
+   /**
+    * Unpack pixels to Z32_UNORM.
+    *
+    * Only defined for depth formats.
+    */
+   void
+   (*unpack_z_32unorm)(uint32_t *dst, unsigned dst_stride,
+                       const uint8_t *src, unsigned src_stride,
+                       unsigned width, unsigned height);
+
+   /**
+    * Pack pixels from Z32_FLOAT.
+    *
+    * Only defined for depth formats.
+    */
+   void
+   (*pack_z_32unorm)(uint8_t *dst, unsigned dst_stride,
+                     const uint32_t *src, unsigned src_stride,
+                     unsigned width, unsigned height);
+
+   /**
+    * Unpack pixels to Z32_FLOAT.
+    *
+    * Only defined for depth formats.
+    */
+   void
+   (*unpack_z_float)(float *dst, unsigned dst_stride,
+                     const uint8_t *src, unsigned src_stride,
+                     unsigned width, unsigned height);
+
+   /**
+    * Pack pixels from Z32_FLOAT.
+    *
+    * Only defined for depth formats.
+    */
+   void
+   (*pack_z_float)(uint8_t *dst, unsigned dst_stride,
+                   const float *src, unsigned src_stride,
+                   unsigned width, unsigned height);
+
+   /**
+    * Unpack pixels to S8_USCALED.
+    *
+    * Only defined for stencil formats.
+    */
+   void
+   (*unpack_s_32unorm)(uint8_t *dst, unsigned dst_stride,
+                       const uint8_t *src, unsigned src_stride,
+                       unsigned width, unsigned height);
+
+   /**
+    * Pack pixels from S8_USCALED.
+    *
+    * Only defined for stencil formats.
     */
    void
-   (*fetch_float)(float *dst,
-                  const uint8_t *src,
-                  unsigned i, unsigned j);
+   (*pack_s_8uscaled)(uint8_t *dst, unsigned dst_stride,
+                      const uint8_t *src, unsigned src_stride,
+                      unsigned width, unsigned height);
+
 };
 
 
diff --git a/src/gallium/auxiliary/util/u_format_other.c b/src/gallium/auxiliary/util/u_format_other.c
index cecbe5e5a1..723fa8c3bf 100644
--- a/src/gallium/auxiliary/util/u_format_other.c
+++ b/src/gallium/auxiliary/util/u_format_other.c
@@ -31,7 +31,7 @@
 
 
 void
-util_format_r9g9b9e5_float_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                         const uint8_t *src_row, unsigned src_stride,
                                         unsigned width, unsigned height)
 {
@@ -39,7 +39,7 @@ util_format_r9g9b9e5_float_unpack_float(float *dst_row, unsigned dst_stride,
 }
 
 void
-util_format_r9g9b9e5_float_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                       const float *src_row, unsigned src_stride,
                                       unsigned width, unsigned height)
 {
@@ -47,7 +47,7 @@ util_format_r9g9b9e5_float_pack_float(uint8_t *dst_row, unsigned dst_stride,
 }
 
 void
-util_format_r9g9b9e5_float_fetch_float(float *dst, const uint8_t *src,
+util_format_r9g9b9e5_float_fetch_rgba_float(float *dst, const uint8_t *src,
                                        unsigned i, unsigned j)
 {
 
@@ -55,7 +55,7 @@ util_format_r9g9b9e5_float_fetch_float(float *dst, const uint8_t *src,
 
 
 void
-util_format_r9g9b9e5_float_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                          const uint8_t *src_row, unsigned src_stride,
                                          unsigned width, unsigned height)
 {
@@ -64,7 +64,7 @@ util_format_r9g9b9e5_float_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r9g9b9e5_float_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                        const uint8_t *src_row, unsigned src_stride,
                                        unsigned width, unsigned height)
 {
@@ -73,7 +73,7 @@ util_format_r9g9b9e5_float_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r1_unorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r1_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                   const uint8_t *src_row, unsigned src_stride,
                                   unsigned width, unsigned height)
 {
@@ -82,7 +82,7 @@ util_format_r1_unorm_unpack_float(float *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r1_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r1_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                 const float *src_row, unsigned src_stride,
                                 unsigned width, unsigned height)
 {
@@ -91,7 +91,7 @@ util_format_r1_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r1_unorm_fetch_float(float *dst, const uint8_t *src,
+util_format_r1_unorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                  unsigned i, unsigned j)
 {
 
@@ -99,7 +99,7 @@ util_format_r1_unorm_fetch_float(float *dst, const uint8_t *src,
 
 
 void
-util_format_r1_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r1_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                    const uint8_t *src_row, unsigned src_stride,
                                    unsigned width, unsigned height)
 {
@@ -108,7 +108,7 @@ util_format_r1_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r1_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r1_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                  const uint8_t *src_row, unsigned src_stride,
                                  unsigned width, unsigned height)
 {
@@ -123,7 +123,7 @@ util_format_r1_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8bx_snorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                       const uint8_t *src_row, unsigned src_stride,
                                       unsigned width, unsigned height)
 {
@@ -156,7 +156,7 @@ util_format_r8g8bx_snorm_unpack_float(float *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8bx_snorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                        const uint8_t *src_row, unsigned src_stride,
                                        unsigned width, unsigned height)
 {
@@ -188,7 +188,7 @@ util_format_r8g8bx_snorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8bx_snorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                     const float *src_row, unsigned src_stride,
                                     unsigned width, unsigned height)
 {
@@ -217,7 +217,7 @@ util_format_r8g8bx_snorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8bx_snorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                      const uint8_t *src_row, unsigned src_stride,
                                      unsigned width, unsigned height)
 {
@@ -247,7 +247,7 @@ util_format_r8g8bx_snorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8bx_snorm_fetch_float(float *dst, const uint8_t *src,
+util_format_r8g8bx_snorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                      unsigned i, unsigned j)
 {
    uint16_t value = *(const uint16_t *)src;
diff --git a/src/gallium/auxiliary/util/u_format_other.h b/src/gallium/auxiliary/util/u_format_other.h
index a46d698188..98c97be33f 100644
--- a/src/gallium/auxiliary/util/u_format_other.h
+++ b/src/gallium/auxiliary/util/u_format_other.h
@@ -34,74 +34,74 @@
 
 
 void
-util_format_r9g9b9e5_float_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                         const uint8_t *src_row, unsigned src_stride,
                                         unsigned width, unsigned height);
 
 void
-util_format_r9g9b9e5_float_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                       const float *src_row, unsigned src_stride,
                                       unsigned width, unsigned height);
 
 void
-util_format_r9g9b9e5_float_fetch_float(float *dst, const uint8_t *src,
+util_format_r9g9b9e5_float_fetch_rgba_float(float *dst, const uint8_t *src,
                                        unsigned i, unsigned j);
 
 void
-util_format_r9g9b9e5_float_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                          const uint8_t *src_row, unsigned src_stride,
                                          unsigned width, unsigned height);
 
 void
-util_format_r9g9b9e5_float_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r9g9b9e5_float_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                        const uint8_t *src_row, unsigned src_stride,
                                        unsigned width, unsigned height);
 
 void
-util_format_r1_unorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r1_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                   const uint8_t *src_row, unsigned src_stride,
                                   unsigned width, unsigned height);
 
 void
-util_format_r1_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r1_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                 const float *src_row, unsigned src_stride,
                                 unsigned width, unsigned height);
 
 void
-util_format_r1_unorm_fetch_float(float *dst, const uint8_t *src,
+util_format_r1_unorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                  unsigned i, unsigned j);
 
 void
-util_format_r1_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r1_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                    const uint8_t *src_row, unsigned src_stride,
                                    unsigned width, unsigned height);
 
 void
-util_format_r1_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r1_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                  const uint8_t *src_row, unsigned src_stride,
                                  unsigned width, unsigned height);
 
 void
-util_format_r8g8bx_snorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                       const uint8_t *src_row, unsigned src_stride,
                                       unsigned width, unsigned height);
 
 void
-util_format_r8g8bx_snorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                     const float *src_row, unsigned src_stride,
                                     unsigned width, unsigned height);
 
 void
-util_format_r8g8bx_snorm_fetch_float(float *dst, const uint8_t *src,
+util_format_r8g8bx_snorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                      unsigned i, unsigned j);
 
 void
-util_format_r8g8bx_snorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                        const uint8_t *src_row, unsigned src_stride,
                                        unsigned width, unsigned height);
 
 void
-util_format_r8g8bx_snorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8bx_snorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                      const uint8_t *src_row, unsigned src_stride,
                                      unsigned width, unsigned height);
 
diff --git a/src/gallium/auxiliary/util/u_format_pack.py b/src/gallium/auxiliary/util/u_format_pack.py
index 06831d1e19..7c5fa9c00a 100644
--- a/src/gallium/auxiliary/util/u_format_pack.py
+++ b/src/gallium/auxiliary/util/u_format_pack.py
@@ -46,12 +46,21 @@ from u_format_parse import *
 def generate_format_type(format):
     '''Generate a structure that describes the format.'''
 
+    assert format.layout == PLAIN
+    
     print 'union util_format_%s {' % format.short_name()
-    if format.is_bitmask() or format.short_name() == "r11g11b10_float":
+    
+    if format.block_size() in (8, 16, 32, 64):
         print '   uint%u_t value;' % (format.block_size(),)
+
+    use_bitfields = False
+    for channel in format.channels:
+        if channel.size % 8 or not is_pot(channel.size):
+            use_bitfields = True
+
     print '   struct {'
     for channel in format.channels:
-        if (format.is_bitmask() or format.is_mixed()) and not format.is_array() or format.short_name() == "r11g11b10_float":
+        if use_bitfields:
             if channel.type == VOID:
                 if channel.size:
                     print '      unsigned %s:%u;' % (channel.name, channel.size)
@@ -60,7 +69,9 @@ def generate_format_type(format):
             elif channel.type in (SIGNED, FIXED):
                 print '      int %s:%u;' % (channel.name, channel.size)
             elif channel.type == FLOAT:
-                if channel.size == 32:
+                if channel.size == 64:
+                    print '      double %s;' % (channel.name)
+                elif channel.size == 32:
                     print '      float %s;' % (channel.name)
                 else:
                     print '      unsigned %s:%u;' % (channel.name, channel.size)
@@ -116,10 +127,6 @@ def is_format_supported(format):
         if channel.type == FLOAT and channel.size not in (16, 32, 64):
             return False
 
-    # We can only read a color from a depth/stencil format if the depth channel is present
-    if format.colorspace == 'zs' and format.swizzles[0] == SWIZZLE_NONE:
-        return False
-
     return True
 
 
@@ -442,11 +449,6 @@ def generate_unpack_kernel(format, dst_channel, dst_native_type):
                 value = '0'
             else:
                 assert False
-            if format.colorspace == ZS:
-                if i == 3:
-                    value = get_one(dst_channel)
-                elif i >= 1:
-                    value = 'dst[0]'
             print '         dst[%u] = %s; /* %s */' % (i, value, 'rgba'[i])
         
     else:
@@ -475,11 +477,6 @@ def generate_unpack_kernel(format, dst_channel, dst_native_type):
                 value = '0'
             else:
                 assert False
-            if format.colorspace == ZS:
-                if i == 3:
-                    value = get_one(dst_channel)
-                elif i >= 1:
-                    value = 'dst[0]'
             print '         dst[%u] = %s; /* %s */' % (i, value, 'rgba'[i])
     
 
@@ -511,11 +508,6 @@ def generate_pack_kernel(format, src_channel, src_native_type):
                                         dst_channel, dst_native_type, 
                                         value,
                                         dst_colorspace = dst_colorspace)
-                if format.colorspace == ZS:
-                    if i == 3:
-                        value = get_one(dst_channel)
-                    elif i >= 1:
-                        value = '0'
                 if dst_channel.type in (UNSIGNED, SIGNED):
                     if shift + dst_channel.size < depth:
                         value = '(%s) & 0x%x' % (value, (1 << dst_channel.size) - 1)
@@ -555,11 +547,6 @@ def generate_pack_kernel(format, src_channel, src_native_type):
                                     dst_channel, dst_native_type, 
                                     value, 
                                     dst_colorspace = dst_colorspace)
-            if format.colorspace == ZS:
-                if i == 3:
-                    value = get_one(dst_channel)
-                elif i >= 1:
-                    value = '0'
             print '         pixel.chan.%s = %s;' % (dst_channel.name, value)
     
         bswap_format(format)
@@ -641,27 +628,30 @@ def generate_format_fetch(format, dst_channel, dst_native_type, dst_suffix):
 
 
 def is_format_hand_written(format):
-    return format.layout in ('s3tc', 'subsampled', 'other')
+    return format.layout in ('s3tc', 'subsampled', 'other') or format.colorspace == ZS
+
 
 def generate(formats):
     print
     print '#include "pipe/p_compiler.h"'
     print '#include "u_math.h"'
+    print '#include "u_half.h"'
     print '#include "u_format.h"'
     print '#include "u_format_other.h"'
     print '#include "u_format_srgb.h"'
     print '#include "u_format_yuv.h"'
-    print '#include "u_half.h"'
+    print '#include "u_format_zs.h"'
     print
 
     for format in formats:
-        if is_format_supported(format):
-            generate_format_type(format)
-
         if not is_format_hand_written(format):
+            
+            if is_format_supported(format):
+                generate_format_type(format)
+
             channel = Channel(FLOAT, False, 32)
             native_type = 'float'
-            suffix = 'float'
+            suffix = 'rgba_float'
 
             generate_format_unpack(format, channel, native_type, suffix)
             generate_format_pack(format, channel, native_type, suffix)
@@ -669,7 +659,8 @@ def generate(formats):
 
             channel = Channel(UNSIGNED, True, 8)
             native_type = 'uint8_t'
-            suffix = '8unorm'
+            suffix = 'rgba_8unorm'
 
             generate_format_unpack(format, channel, native_type, suffix)
             generate_format_pack(format, channel, native_type, suffix)
+
diff --git a/src/gallium/auxiliary/util/u_format_s3tc.c b/src/gallium/auxiliary/util/u_format_s3tc.c
index 35950ee47a..79dee2b423 100644
--- a/src/gallium/auxiliary/util/u_format_s3tc.c
+++ b/src/gallium/auxiliary/util/u_format_s3tc.c
@@ -161,31 +161,31 @@ util_format_s3tc_init(void)
  */
 
 void
-util_format_dxt1_rgb_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_rgb_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    util_format_dxt1_rgb_fetch(0, src, i, j, dst);
 }
 
 void
-util_format_dxt1_rgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_rgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    util_format_dxt1_rgba_fetch(0, src, i, j, dst);
 }
 
 void
-util_format_dxt3_rgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt3_rgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    util_format_dxt3_rgba_fetch(0, src, i, j, dst);
 }
 
 void
-util_format_dxt5_rgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt5_rgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    util_format_dxt5_rgba_fetch(0, src, i, j, dst);
 }
 
 void
-util_format_dxt1_rgb_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_rgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    uint8_t tmp[4];
    util_format_dxt1_rgb_fetch(0, src, i, j, tmp);
@@ -196,7 +196,7 @@ util_format_dxt1_rgb_fetch_float(float *dst, const uint8_t *src, unsigned i, uns
 }
 
 void
-util_format_dxt1_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_rgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    uint8_t tmp[4];
    util_format_dxt1_rgba_fetch(0, src, i, j, tmp);
@@ -207,7 +207,7 @@ util_format_dxt1_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, un
 }
 
 void
-util_format_dxt3_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt3_rgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    uint8_t tmp[4];
    util_format_dxt3_rgba_fetch(0, src, i, j, tmp);
@@ -218,7 +218,7 @@ util_format_dxt3_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, un
 }
 
 void
-util_format_dxt5_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt5_rgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
    uint8_t tmp[4];
    util_format_dxt5_rgba_fetch(0, src, i, j, tmp);
@@ -234,7 +234,7 @@ util_format_dxt5_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, un
  */
 
 void
-util_format_dxt1_rgb_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgb_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -253,7 +253,7 @@ util_format_dxt1_rgb_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const
 }
 
 void
-util_format_dxt1_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -272,7 +272,7 @@ util_format_dxt1_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const
 }
 
 void
-util_format_dxt3_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_rgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -291,7 +291,7 @@ util_format_dxt3_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const
 }
 
 void
-util_format_dxt5_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_rgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -310,7 +310,7 @@ util_format_dxt5_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const
 }
 
 void
-util_format_dxt1_rgb_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgb_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -334,7 +334,7 @@ util_format_dxt1_rgb_unpack_float(float *dst_row, unsigned dst_stride, const uin
 }
 
 void
-util_format_dxt1_rgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -358,7 +358,7 @@ util_format_dxt1_rgba_unpack_float(float *dst_row, unsigned dst_stride, const ui
 }
 
 void
-util_format_dxt3_rgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_rgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -382,7 +382,7 @@ util_format_dxt3_rgba_unpack_float(float *dst_row, unsigned dst_stride, const ui
 }
 
 void
-util_format_dxt5_rgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_rgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j;
    for(y = 0; y < height; y += 4) {
@@ -411,7 +411,7 @@ util_format_dxt5_rgba_unpack_float(float *dst_row, unsigned dst_stride, const ui
  */
 
 void
-util_format_dxt1_rgb_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgb_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -436,7 +436,7 @@ util_format_dxt1_rgb_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const ui
 }
 
 void
-util_format_dxt1_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -461,7 +461,7 @@ util_format_dxt1_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const u
 }
 
 void
-util_format_dxt3_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_rgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -486,7 +486,7 @@ util_format_dxt3_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const u
 }
 
 void
-util_format_dxt5_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_rgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -511,7 +511,7 @@ util_format_dxt5_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const u
 }
 
 void
-util_format_dxt1_rgb_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgb_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -536,7 +536,7 @@ util_format_dxt1_rgb_pack_float(uint8_t *dst_row, unsigned dst_stride, const flo
 }
 
 void
-util_format_dxt1_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_rgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -561,7 +561,7 @@ util_format_dxt1_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const fl
 }
 
 void
-util_format_dxt3_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_rgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -586,7 +586,7 @@ util_format_dxt3_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const fl
 }
 
 void
-util_format_dxt5_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_rgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
    unsigned x, y, i, j, k;
    for(y = 0; y < height; y += 4) {
@@ -618,146 +618,146 @@ util_format_dxt5_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const fl
  */
 
 void
-util_format_dxt1_srgb_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgb_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgb_unpack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgb_unpack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgb_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgb_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgb_pack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgb_pack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgb_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_srgb_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt1_rgb_fetch_8unorm(dst, src, i, j);
+   util_format_dxt1_rgb_fetch_rgba_8unorm(dst, src, i, j);
 }
 
 void
-util_format_dxt1_srgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgba_unpack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgba_unpack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgba_pack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgba_pack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_srgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt1_rgba_fetch_8unorm(dst, src, i, j);
+   util_format_dxt1_rgba_fetch_rgba_8unorm(dst, src, i, j);
 }
 
 void
-util_format_dxt3_srgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_srgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt3_rgba_unpack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt3_rgba_unpack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt3_srgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_srgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt3_rgba_pack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt3_rgba_pack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt3_srgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt3_srgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt3_rgba_fetch_8unorm(dst, src, i, j);
+   util_format_dxt3_rgba_fetch_rgba_8unorm(dst, src, i, j);
 }
 
 void
-util_format_dxt5_srgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_srgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt5_rgba_unpack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt5_rgba_unpack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt5_srgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_srgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt5_rgba_pack_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt5_rgba_pack_rgba_8unorm(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt5_srgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt5_srgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt5_rgba_fetch_8unorm(dst, src, i, j);
+   util_format_dxt5_rgba_fetch_rgba_8unorm(dst, src, i, j);
 }
 
 void
-util_format_dxt1_srgb_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgb_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgb_unpack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgb_unpack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgb_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgb_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgb_pack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgb_pack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgb_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt1_rgb_fetch_float(dst, src, i, j);
+   util_format_dxt1_rgb_fetch_rgba_float(dst, src, i, j);
 }
 
 void
-util_format_dxt1_srgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgba_unpack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgba_unpack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt1_srgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt1_rgba_pack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt1_rgba_pack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt1_srgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt1_srgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt1_rgba_fetch_float(dst, src, i, j);
+   util_format_dxt1_rgba_fetch_rgba_float(dst, src, i, j);
 }
 
 void
-util_format_dxt3_srgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_srgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt3_rgba_unpack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt3_rgba_unpack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt3_srgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt3_srgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt3_rgba_pack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt3_rgba_pack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt3_srgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt3_srgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt3_rgba_fetch_float(dst, src, i, j);
+   util_format_dxt3_rgba_fetch_rgba_float(dst, src, i, j);
 }
 
 void
-util_format_dxt5_srgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_srgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt5_rgba_unpack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt5_rgba_unpack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt5_srgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+util_format_dxt5_srgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
 {
-   util_format_dxt5_rgba_pack_float(dst_row, dst_stride, src_row, src_stride, width, height);
+   util_format_dxt5_rgba_pack_rgba_float(dst_row, dst_stride, src_row, src_stride, width, height);
 }
 
 void
-util_format_dxt5_srgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+util_format_dxt5_srgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
 {
-   util_format_dxt5_rgba_fetch_float(dst, src, i, j);
+   util_format_dxt5_rgba_fetch_rgba_float(dst, src, i, j);
 }
 
diff --git a/src/gallium/auxiliary/util/u_format_s3tc.h b/src/gallium/auxiliary/util/u_format_s3tc.h
index 5c3a9dd22c..97770abd42 100644
--- a/src/gallium/auxiliary/util/u_format_s3tc.h
+++ b/src/gallium/auxiliary/util/u_format_s3tc.h
@@ -70,149 +70,149 @@ util_format_s3tc_init(void);
 
 
 void
-util_format_dxt1_rgb_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgb_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgb_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgb_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgb_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_rgb_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt1_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_rgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt3_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_rgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_rgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_rgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt3_rgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt5_rgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_rgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_rgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_rgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_rgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt5_rgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt1_srgb_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgb_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgb_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgb_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgb_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_srgb_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt1_srgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_srgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt3_srgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_srgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_srgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_srgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_srgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt3_srgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt5_srgba_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_srgba_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_srgba_pack_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_srgba_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_srgba_fetch_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt5_srgba_fetch_rgba_8unorm(uint8_t *dst, const uint8_t *src, unsigned i, unsigned j);
 
 
 void
-util_format_dxt1_rgb_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgb_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgb_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgb_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgb_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_rgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt1_rgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_rgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_rgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt3_rgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_rgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_rgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt3_rgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt5_rgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_rgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_rgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_rgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_rgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt5_rgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt1_srgb_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgb_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgb_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgb_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgb_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt1_srgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt1_srgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt1_srgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt1_srgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt3_srgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_srgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_srgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt3_srgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt3_srgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt3_srgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 void
-util_format_dxt5_srgba_unpack_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_srgba_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_srgba_pack_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+util_format_dxt5_srgba_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
 
 void
-util_format_dxt5_srgba_fetch_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
+util_format_dxt5_srgba_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j);
 
 
 #endif /* U_FORMAT_S3TC_H_ */
diff --git a/src/gallium/auxiliary/util/u_format_table.py b/src/gallium/auxiliary/util/u_format_table.py
index 94a4331b15..5440a8b7b8 100755
--- a/src/gallium/auxiliary/util/u_format_table.py
+++ b/src/gallium/auxiliary/util/u_format_table.py
@@ -129,11 +129,34 @@ def write_format_table(formats):
             print "      %s%s\t/* %s */" % (swizzle_map[swizzle], sep, comment)
         print "   },"
         print "   %s," % (colorspace_map(format.colorspace),)
-        print "   &util_format_%s_unpack_8unorm," % format.short_name() 
-        print "   &util_format_%s_pack_8unorm," % format.short_name() 
-        print "   &util_format_%s_unpack_float," % format.short_name() 
-        print "   &util_format_%s_pack_float," % format.short_name() 
-        print "   &util_format_%s_fetch_float" % format.short_name() 
+        if format.colorspace != ZS:
+            print "   &util_format_%s_unpack_rgba_8unorm," % format.short_name() 
+            print "   &util_format_%s_pack_rgba_8unorm," % format.short_name() 
+            print "   &util_format_%s_unpack_rgba_float," % format.short_name() 
+            print "   &util_format_%s_pack_rgba_float," % format.short_name() 
+            print "   &util_format_%s_fetch_rgba_float," % format.short_name()
+        else:
+            print "   NULL, /* unpack_rgba_8unorm */" 
+            print "   NULL, /* pack_rgba_8unorm */" 
+            print "   NULL, /* unpack_rgba_float */" 
+            print "   NULL, /* pack_rgba_float */" 
+            print "   NULL, /* fetch_rgba_float */" 
+        if format.colorspace == ZS and format.swizzles[0] != SWIZZLE_NONE:
+            print "   &util_format_%s_unpack_z_32unorm," % format.short_name() 
+            print "   &util_format_%s_pack_z_32unorm," % format.short_name() 
+            print "   &util_format_%s_unpack_z_float," % format.short_name() 
+            print "   &util_format_%s_pack_z_float," % format.short_name() 
+        else:
+            print "   NULL, /* unpack_z_32unorm */" 
+            print "   NULL, /* pack_z_32unorm */" 
+            print "   NULL, /* unpack_z_float */" 
+            print "   NULL, /* pack_z_float */" 
+        if format.colorspace == ZS and format.swizzles[1] != SWIZZLE_NONE:
+            print "   &util_format_%s_unpack_s_8uscaled," % format.short_name() 
+            print "   &util_format_%s_pack_s_8uscaled" % format.short_name() 
+        else:
+            print "   NULL, /* unpack_s_8uscaled */" 
+            print "   NULL /* pack_s_8uscaled */" 
         print "};"
         print
         
diff --git a/src/gallium/auxiliary/util/u_format_yuv.c b/src/gallium/auxiliary/util/u_format_yuv.c
index 0847476e2e..ab8bf29c97 100644
--- a/src/gallium/auxiliary/util/u_format_yuv.c
+++ b/src/gallium/auxiliary/util/u_format_yuv.c
@@ -38,7 +38,7 @@
 
 
 void
-util_format_r8g8_b8g8_unorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                          const uint8_t *src_row, unsigned src_stride,
                                          unsigned width, unsigned height)
 {
@@ -100,7 +100,7 @@ util_format_r8g8_b8g8_unorm_unpack_float(float *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8_b8g8_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                           const uint8_t *src_row, unsigned src_stride,
                                           unsigned width, unsigned height)
 {
@@ -162,7 +162,7 @@ util_format_r8g8_b8g8_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8_b8g8_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                        const float *src_row, unsigned src_stride,
                                        unsigned width, unsigned height)
 {
@@ -219,7 +219,7 @@ util_format_r8g8_b8g8_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8_b8g8_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                         const uint8_t *src_row, unsigned src_stride,
                                         unsigned width, unsigned height)
 {
@@ -276,7 +276,7 @@ util_format_r8g8_b8g8_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_r8g8_b8g8_unorm_fetch_float(float *dst, const uint8_t *src,
+util_format_r8g8_b8g8_unorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                         unsigned i, unsigned j)
 {
    assert(i < 2);
@@ -290,7 +290,7 @@ util_format_r8g8_b8g8_unorm_fetch_float(float *dst, const uint8_t *src,
 
 
 void
-util_format_g8r8_g8b8_unorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                          const uint8_t *src_row, unsigned src_stride,
                                          unsigned width, unsigned height)
 {
@@ -352,7 +352,7 @@ util_format_g8r8_g8b8_unorm_unpack_float(float *dst_row, unsigned dst_stride,
 
 
 void
-util_format_g8r8_g8b8_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                           const uint8_t *src_row, unsigned src_stride,
                                           unsigned width, unsigned height)
 {
@@ -414,7 +414,7 @@ util_format_g8r8_g8b8_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_g8r8_g8b8_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                        const float *src_row, unsigned src_stride,
                                        unsigned width, unsigned height)
 {
@@ -471,7 +471,7 @@ util_format_g8r8_g8b8_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_g8r8_g8b8_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                         const uint8_t *src_row, unsigned src_stride,
                                         unsigned width, unsigned height)
 {
@@ -528,7 +528,7 @@ util_format_g8r8_g8b8_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_g8r8_g8b8_unorm_fetch_float(float *dst, const uint8_t *src,
+util_format_g8r8_g8b8_unorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                         unsigned i, unsigned j)
 {
    assert(i < 2);
@@ -542,7 +542,7 @@ util_format_g8r8_g8b8_unorm_fetch_float(float *dst, const uint8_t *src,
 
 
 void
-util_format_uyvy_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_uyvy_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                               const uint8_t *src_row, unsigned src_stride,
                               unsigned width, unsigned height)
 {
@@ -598,7 +598,7 @@ util_format_uyvy_unpack_float(float *dst_row, unsigned dst_stride,
 
 
 void
-util_format_uyvy_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_uyvy_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                const uint8_t *src_row, unsigned src_stride,
                                unsigned width, unsigned height)
 {
@@ -654,7 +654,7 @@ util_format_uyvy_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_uyvy_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_uyvy_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                             const float *src_row, unsigned src_stride,
                             unsigned width, unsigned height)
 {
@@ -715,7 +715,7 @@ util_format_uyvy_pack_float(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_uyvy_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_uyvy_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                              const uint8_t *src_row, unsigned src_stride,
                              unsigned width, unsigned height)
 {
@@ -776,7 +776,7 @@ util_format_uyvy_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_uyvy_fetch_float(float *dst, const uint8_t *src,
+util_format_uyvy_fetch_rgba_float(float *dst, const uint8_t *src,
                              unsigned i, unsigned j)
 {
    uint8_t y, u, v;
@@ -795,7 +795,7 @@ util_format_uyvy_fetch_float(float *dst, const uint8_t *src,
 
 
 void
-util_format_yuyv_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_yuyv_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                               const uint8_t *src_row, unsigned src_stride,
                               unsigned width, unsigned height)
 {
@@ -851,7 +851,7 @@ util_format_yuyv_unpack_float(float *dst_row, unsigned dst_stride,
 
 
 void
-util_format_yuyv_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_yuyv_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                const uint8_t *src_row, unsigned src_stride,
                                unsigned width, unsigned height)
 {
@@ -907,7 +907,7 @@ util_format_yuyv_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_yuyv_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_yuyv_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                             const float *src_row, unsigned src_stride,
                             unsigned width, unsigned height)
 {
@@ -968,7 +968,7 @@ util_format_yuyv_pack_float(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_yuyv_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_yuyv_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                              const uint8_t *src_row, unsigned src_stride,
                              unsigned width, unsigned height)
 {
@@ -1029,7 +1029,7 @@ util_format_yuyv_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
 
 
 void
-util_format_yuyv_fetch_float(float *dst, const uint8_t *src,
+util_format_yuyv_fetch_rgba_float(float *dst, const uint8_t *src,
                              unsigned i, unsigned j)
 {
    uint8_t y, u, v;
diff --git a/src/gallium/auxiliary/util/u_format_yuv.h b/src/gallium/auxiliary/util/u_format_yuv.h
index 10cf678480..dc9632346d 100644
--- a/src/gallium/auxiliary/util/u_format_yuv.h
+++ b/src/gallium/auxiliary/util/u_format_yuv.h
@@ -122,100 +122,100 @@ util_format_yuv_to_rgb_8unorm(uint8_t y, uint8_t u, uint8_t v,
 
 
 void
-util_format_uyvy_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_uyvy_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                               const uint8_t *src_row, unsigned src_stride,
                               unsigned width, unsigned height);
 
 void
-util_format_uyvy_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_uyvy_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                const uint8_t *src_row, unsigned src_stride,
                                unsigned width, unsigned height);
 
 void
-util_format_uyvy_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_uyvy_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                             const float *src_row, unsigned src_stride,
                             unsigned width, unsigned height);
 
 void
-util_format_uyvy_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_uyvy_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                              const uint8_t *src_row, unsigned src_stride,
                              unsigned width, unsigned height);
 
 void
-util_format_uyvy_fetch_float(float *dst, const uint8_t *src,
+util_format_uyvy_fetch_rgba_float(float *dst, const uint8_t *src,
                              unsigned i, unsigned j);
 
 void
-util_format_yuyv_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_yuyv_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                               const uint8_t *src_row, unsigned src_stride,
                               unsigned width, unsigned height);
 
 void
-util_format_yuyv_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_yuyv_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                const uint8_t *src_row, unsigned src_stride,
                                unsigned width, unsigned height);
 
 void
-util_format_yuyv_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_yuyv_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                             const float *src_row, unsigned src_stride,
                             unsigned width, unsigned height);
 
 void
-util_format_yuyv_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_yuyv_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                              const uint8_t *src_row, unsigned src_stride,
                              unsigned width, unsigned height);
 
 void
-util_format_yuyv_fetch_float(float *dst, const uint8_t *src,
+util_format_yuyv_fetch_rgba_float(float *dst, const uint8_t *src,
                              unsigned i, unsigned j);
 
 
 void
-util_format_r8g8_b8g8_unorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                          const uint8_t *src_row, unsigned src_stride,
                                          unsigned width, unsigned height);
 
 void
-util_format_r8g8_b8g8_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                           const uint8_t *src_row, unsigned src_stride,
                                           unsigned width, unsigned height);
 
 void
-util_format_r8g8_b8g8_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                        const float *src_row, unsigned src_stride,
                                        unsigned width, unsigned height);
 
 void
-util_format_r8g8_b8g8_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_r8g8_b8g8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                         const uint8_t *src_row, unsigned src_stride,
                                         unsigned width, unsigned height);
 
 void
-util_format_r8g8_b8g8_unorm_fetch_float(float *dst, const uint8_t *src,
+util_format_r8g8_b8g8_unorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                         unsigned i, unsigned j);
 
 void
-util_format_g8r8_g8b8_unorm_unpack_float(float *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
                                          const uint8_t *src_row, unsigned src_stride,
                                          unsigned width, unsigned height);
 
 void
-util_format_g8r8_g8b8_unorm_unpack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                           const uint8_t *src_row, unsigned src_stride,
                                           unsigned width, unsigned height);
 
 void
-util_format_g8r8_g8b8_unorm_pack_float(uint8_t *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
                                        const float *src_row, unsigned src_stride,
                                        unsigned width, unsigned height);
 
 void
-util_format_g8r8_g8b8_unorm_pack_8unorm(uint8_t *dst_row, unsigned dst_stride,
+util_format_g8r8_g8b8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
                                         const uint8_t *src_row, unsigned src_stride,
                                         unsigned width, unsigned height);
 
 void
-util_format_g8r8_g8b8_unorm_fetch_float(float *dst, const uint8_t *src,
+util_format_g8r8_g8b8_unorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                         unsigned i, unsigned j);
 
 
diff --git a/src/gallium/auxiliary/util/u_format_zs.c b/src/gallium/auxiliary/util/u_format_zs.c
new file mode 100644
index 0000000000..68c88e520c
--- /dev/null
+++ b/src/gallium/auxiliary/util/u_format_zs.c
@@ -0,0 +1,893 @@
+/**************************************************************************
+ *
+ * Copyright 2010 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+
+
+#include "u_debug.h"
+#include "u_math.h"
+#include "u_format_zs.h"
+
+
+void
+util_format_s8_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                         const uint8_t *src_row, unsigned src_stride,
+                                         unsigned width, unsigned height)
+{
+   unsigned y;
+   for(y = 0; y < height; ++y) {
+      memcpy(dst_row, src_row, width);
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_s8_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                       const uint8_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned y;
+   for(y = 0; y < height; ++y) {
+      memcpy(dst_row, src_row, width);
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z16_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                     const uint8_t *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      float *dst = dst_row;
+      const uint16_t *src = (const uint16_t *)src_row;
+      for(x = 0; x < width; ++x) {
+         uint16_t value = *src++;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         dst[0] = (float)(value * (1.0f/0xffff));
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z16_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                   const float *src_row, unsigned src_stride,
+                                   unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const float *src = src_row;
+      uint16_t *dst = (uint16_t *)dst_row;
+      for(x = 0; x < width; ++x) {
+         uint16_t value;
+         value = (uint16_t)(*src * 0xffff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *dst++ = value;
+         src += 1;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z16_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                       const uint8_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint32_t *dst = dst_row;
+      const uint16_t *src = (const uint16_t *)src_row;
+      for(x = 0; x < width; ++x) {
+         uint16_t value = *src++;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         /* value * 0xffffffff / 0xffff */
+         *dst++ = (value << 16) | value;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z16_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                     const uint32_t *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint32_t *src = src_row;
+      uint16_t *dst = (uint16_t *)dst_row;
+      for(x = 0; x < width; ++x) {
+         uint16_t value;
+         value = (uint16_t)(*src++ >> 16);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *dst++ = value;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z32_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                     const uint8_t *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      float *dst = dst_row;
+      const uint32_t *src = (const uint32_t *)src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *src++;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *dst++ = (float)(value * (1.0/0xffffffff));
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                   const float *src_row, unsigned src_stride,
+                                   unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const float *src = src_row;
+      uint32_t *dst = (uint32_t *)dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = (uint32_t)(*src * (double)0xffffffff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *dst++ = value;
+         ++src;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z32_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                       const uint8_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned y;
+   for(y = 0; y < height; ++y) {
+      memcpy(dst_row, src_row, width * 4);
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                     const uint32_t *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned y;
+   for(y = 0; y < height; ++y) {
+      memcpy(dst_row, src_row, width * 4);
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_float_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                     const uint8_t *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned y;
+   for(y = 0; y < height; ++y) {
+      memcpy(dst_row, src_row, width * 4);
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_float_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                   const float *src_row, unsigned src_stride,
+                                   unsigned width, unsigned height)
+{
+   unsigned y;
+   for(y = 0; y < height; ++y) {
+      memcpy(dst_row, src_row, width * 4);
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_float_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                       const uint8_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint32_t *dst = dst_row;
+      const float *src = (const float *)src_row;
+      for(x = 0; x < width; ++x) {
+         *dst++ = (uint32_t)(*src++ * (double)0xffffffff);
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_float_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                     const uint32_t *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint32_t *src = src_row;
+      float *dst = (float *)dst_row;
+      for(x = 0; x < width; ++x) {
+         *dst++ = (float)(*src++ * (1.0/0xffffffff));
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z24_unorm_s8_uscaled_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                                const uint8_t *src_row, unsigned src_stride,
+                                                unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      float *dst = dst_row;
+      const uint32_t *src = (const uint32_t *)src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *src++;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         z = (value) & 0xffffff;
+         *dst++ = (float)(z * (1.0/0xffffff));
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z24_unorm_s8_uscaled_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                              const float *src_row, unsigned src_stride,
+                                              unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const float *src = src_row;
+      uint32_t *dst = (uint32_t *)dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *dst;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         value &= 0xff000000;
+         value |= ((uint32_t)(*src++ * (double)0xffffff)) & 0xffffff;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *dst++ = value;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z24_unorm_s8_uscaled_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                                  const uint8_t *src_row, unsigned src_stride,
+                                                  unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint32_t *dst = dst_row;
+      const uint32_t *src = (const uint32_t *)src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *src++;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         z = value & 0xffffff;
+         *dst++ = (z << 8) | (z >> 16); /* z * 0xffffffff / 0xffffff */;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z24_unorm_s8_uscaled_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                                const uint32_t *src_row, unsigned src_stride,
+                                                unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint32_t *src = src_row;
+      uint32_t *dst = (uint32_t *)dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)(*src >> 8)) & 0xffffff;
+         value = ((uint32_t)(((uint64_t)src[1]) * 0x1 / 0xffffffff)) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *dst++ = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z24_unorm_s8_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                                   const uint8_t *src_row, unsigned src_stride,
+                                                   unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t z;
+         uint32_t s;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         z = (value) & 0xffffff;
+         s = value >> 24;
+         dst[1] = s;
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z24_unorm_s8_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                                 const uint8_t *src_row, unsigned src_stride,
+                                                 unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)(((uint32_t)MIN2(*src, 1)) * 0xffffff / 0x1)) & 0xffffff;
+         value = (src[1]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_s8_uscaled_z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                                const uint8_t *src_row, unsigned src_stride,
+                                                unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t s;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         s = (value) & 0xff;
+         z = value >> 8;
+         dst[0] = (float)(z * (1.0/0xffffff));
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_s8_uscaled_z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                              const float *src_row, unsigned src_stride,
+                                              unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)CLAMP(src[1], 0, 255)) & 0xff;
+         value = ((uint32_t)(*src * (double)0xffffff)) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_s8_uscaled_z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                                  const uint8_t *src_row, unsigned src_stride,
+                                                  unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint32_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t s;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         s = (value) & 0xff;
+         z = value >> 8;
+         dst[0] = (uint32_t)(((uint64_t)z) * 0xffffffff / 0xffffff);
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_s8_uscaled_z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                                const uint32_t *src_row, unsigned src_stride,
+                                                unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint32_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)(((uint64_t)src[1]) * 0x1 / 0xffffffff)) & 0xff;
+         value = ((uint32_t)(*src >> 8)) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_s8_uscaled_z24_unorm_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                                   const uint8_t *src_row, unsigned src_stride,
+                                                   unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t s;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         s = (value) & 0xff;
+         z = value >> 8;
+         dst[1] = s;
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_s8_uscaled_z24_unorm_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                                 const uint8_t *src_row, unsigned src_stride,
+                                                 unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = (src[1]) & 0xff;
+         value = ((uint32_t)(((uint32_t)MIN2(*src, 1)) * 0xffffff / 0x1)) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z24x8_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                       const uint8_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         z = (value) & 0xffffff;
+         dst[0] = (float)(z * (1.0/0xffffff));
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z24x8_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                     const float *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)(*src * (double)0xffffff)) & 0xffffff;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z24x8_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                         const uint8_t *src_row, unsigned src_stride,
+                                         unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint32_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         z = (value) & 0xffffff;
+         dst[0] = (uint32_t)(((uint64_t)z) * 0xffffffff / 0xffffff);
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z24x8_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                       const uint32_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint32_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)(*src >> 8)) & 0xffffff;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_x8z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                       const uint8_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         z = value >> 8;
+         dst[0] = (float)(z * (1.0/0xffffff));
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_x8z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                     const float *src_row, unsigned src_stride,
+                                     unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)(*src * (double)0xffffff)) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_x8z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                         const uint8_t *src_row, unsigned src_stride,
+                                         unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint32_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t z;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         z = value >> 8;
+         dst[0] = (uint32_t)(((uint64_t)z) * 0xffffffff / 0xffffff);
+         src += 4;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_x8z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                       const uint32_t *src_row, unsigned src_stride,
+                                       unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint32_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; ++x) {
+         uint32_t value;
+         value = ((uint32_t)(*src >> 8)) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 1;
+         dst += 4;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z32_float_s8x24_uscaled_unpack_z_float(float *dst_row, unsigned dst_stride,
+                                                   const uint8_t *src_row, unsigned src_stride,
+                                                   unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      float *dst = dst_row;
+      const float *src = (const float *)src_row;
+      for(x = 0; x < width; ++x) {
+         *dst = *src;
+         src += 2;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_float_s8x24_uscaled_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
+                                                 const float *src_row, unsigned src_stride,
+                                                 unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const float *src = src_row;
+      float *dst = (float *)dst_row;
+      for(x = 0; x < width; ++x) {
+         *dst = *src;
+         src += 1;
+         dst += 2;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z32_float_s8x24_uscaled_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
+                                                     const uint8_t *src_row, unsigned src_stride,
+                                                     unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint32_t *dst = dst_row;
+      const float *src = (const float *)src_row;
+      for(x = 0; x < width; ++x) {
+         *dst = (uint32_t)(*src * (double)0xffffffff);
+         src += 2;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_float_s8x24_uscaled_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
+                                                   const uint32_t *src_row, unsigned src_stride,
+                                                   unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint32_t *src = src_row;
+      float *dst = (float *)dst_row;
+      for(x = 0; x < width; ++x) {
+         *dst = (float)(*src * (1.0/0xffffffff));
+         src += 2;
+         dst += 1;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+void
+util_format_z32_float_s8x24_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                                      const uint8_t *src_row, unsigned src_stride,
+                                                      unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row + 4;
+      for(x = 0; x < width; ++x) {
+         *dst = *src;
+         src += 8;
+         dst += 1;
+      }
+      src_row += src_stride/sizeof(*src_row);
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+void
+util_format_z32_float_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride,
+                                                    const uint8_t *src_row, unsigned src_stride,
+                                                    unsigned width, unsigned height)
+{
+   unsigned x, y;
+   for(y = 0; y < height; ++y) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row + 4;
+      for(x = 0; x < width; ++x) {
+         *dst = *src;
+         src += 1;
+         dst += 8;
+      }
+      dst_row += dst_stride/sizeof(*dst_row);
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
diff --git a/src/gallium/auxiliary/util/u_format_zs.h b/src/gallium/auxiliary/util/u_format_zs.h
new file mode 100644
index 0000000000..650db4b95f
--- /dev/null
+++ b/src/gallium/auxiliary/util/u_format_zs.h
@@ -0,0 +1,196 @@
+/**************************************************************************
+ *
+ * Copyright 2010 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+
+
+#ifndef U_FORMAT_ZS_H_
+#define U_FORMAT_ZS_H_
+
+
+#include "pipe/p_compiler.h"
+
+
+void
+util_format_s8_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_s8_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z16_unorm_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z16_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z16_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z16_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_unorm_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24_unorm_s8_uscaled_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24_unorm_s8_uscaled_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24_unorm_s8_uscaled_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24_unorm_s8_uscaled_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24_unorm_s8_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24_unorm_s8_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_s8_uscaled_z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_s8_uscaled_z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_s8_uscaled_z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_s8_uscaled_z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_s8_uscaled_z24_unorm_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_s8_uscaled_z24_unorm_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24x8_unorm_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24x8_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24x8_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z24x8_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_x8z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_x8z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_x8z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_x8z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_s8x24_uscaled_unpack_z_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_s8x24_uscaled_pack_z_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_s8x24_uscaled_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_s8x24_uscaled_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride, const uint32_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_s8x24_uscaled_unpack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+void
+util_format_z32_float_s8x24_uscaled_pack_s_8uscaled(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height);
+
+
+#endif /* U_FORMAT_ZS_H_ */