Import the infrastructure for building Gallium3D graphics drivers.

Currently only software drivers are supported, looks ok to matthieu.

22 files changed, 40127 insertions, 0 deletions

diff --git a/lib/libGL/gallium/Makefile b/lib/libGL/gallium/Makefile
new file mode 100644
index 000000000..6acdc673c
--- /dev/null
+++ b/lib/libGL/gallium/Makefile
@@ -0,0 +1,7 @@
+# $OpenBSD: Makefile,v 1.1 2012/07/13 12:43:30 mpi Exp $
+
+SUBDIR= libgallium libmesagallium drivers
+
+build depend all install clean cleandir: _SUBDIRUSE
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/Makefile.inc b/lib/libGL/gallium/Makefile.inc
new file mode 100644
index 000000000..311655ad9
--- /dev/null
+++ b/lib/libGL/gallium/Makefile.inc
@@ -0,0 +1,49 @@
+# $OpenBSD: Makefile.inc,v 1.1.1.1 2012/07/13 12:43:30 mpi Exp $
+
+CPP=	cpp -notraditional
+
+OBJS+=  ${SRCS:N*.h:R:S/$/.so/g}
+
+.SUFFIXES:
+.SUFFIXES: .a .c .cpp .o .so
+
+.c.so:
+	@echo "${COMPILE.c} ${PICFLAG} -DPIC ${.IMPSRC} -o ${.TARGET}"
+	@${COMPILE.c} ${PICFLAG} -DPIC ${.IMPSRC} -o ${.TARGET}.o
+	@${LD} -X -r ${.TARGET}.o -o ${.TARGET}
+	@rm -f ${.TARGET}.o
+
+.cpp.so:
+	@echo "${COMPILE.cc} ${PICFLAG} -DPIC ${.IMPSRC} -o ${.TARGET}"
+	@${COMPILE.cc} ${PICFLAG} -DPIC ${.IMPSRC} -o ${.TARGET}.o
+	@${LD} -X -r ${.TARGET}.o -o ${.TARGET}
+	@rm -f ${.TARGET}.o
+
+lib${LIB}_pic.a: ${OBJS} ${DPADD}
+	@rm -f lib${LIB}_pic.a
+	@${AR} cq lib${LIB}_pic.a `${LORDER} ${OBJS} | tsort -q`
+	${RANLIB} lib${LIB}_pic.a
+
+
+.ifdef XENOCARA_HAVE_LLVM
+LLVM_VERSION=	0x301	# 3.1 
+LLVM_CONFIG?=	/usr/local/bin/llvm-config
+
+LLVM_CFLAGS!=	${LLVM_CONFIG} --cflags
+LLVM_CPPFLAGS!=	${LLVM_CONFIG} --cxxflags
+LLVM_LDFLAGS!=	${LLVM_CONFIG} --ldflags
+LLVM_LIBS!=	${LLVM_CONFIG} --libs engine
+
+CFLAGS+=	${LLVM_CFLAGS} -std=gnu99
+CXXFLAGS=	${LLVM_CPPFLAGS}
+LDADD+=		${LLVM_LDFLAGS} ${LLVM_LIBS}
+
+CPPFLAGS+= \
+	-DGALLIUM_LLVMPIPE \
+	-DHAVE_LLVM=${LLVM_VERSION}
+.endif
+
+
+.if ${MACHINE_ARCH} == powerpc || ${MACHINE_ARCH} == sparc
+PICFLAG=-fPIC
+.endif
diff --git a/lib/libGL/gallium/drivers/Makefile b/lib/libGL/gallium/drivers/Makefile
new file mode 100644
index 000000000..11265f2a7
--- /dev/null
+++ b/lib/libGL/gallium/drivers/Makefile
@@ -0,0 +1,11 @@
+# $OpenBSD: Makefile,v 1.1.1.1 2012/07/13 12:43:34 mpi Exp $
+
+SUBDIR= rbug softpipe trace
+
+.ifdef XENOCARA_HAVE_LLVM
+SUBDIR+= llvmpipe
+.endif
+
+build depend all install clean cleandir: _SUBDIRUSE
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/drivers/Makefile.inc b/lib/libGL/gallium/drivers/Makefile.inc
new file mode 100644
index 000000000..974cc3dab
--- /dev/null
+++ b/lib/libGL/gallium/drivers/Makefile.inc
@@ -0,0 +1,23 @@
+# $OpenBSD: Makefile.inc,v 1.1.1.1 2012/07/13 12:43:34 mpi Exp $
+
+GALLIUM=	${.CURDIR}/../../../../../dist/Mesa/src/gallium
+
+CPPFLAGS+= \
+	-I${GALLIUM}/auxiliary \
+	-I${GALLIUM}/include 
+
+all: lib${LIB}_pic.a
+
+obj: _xenocara_obj
+
+install:
+	@echo "Not installing lib${LIB}"
+
+clean: 
+	rm -f lib${LIB}_pic.a ${OBJS}
+
+cleandir:	clean
+
+.PATH: ${GALLIUM}/drivers/${LIB}
+
+.include "../Makefile.inc"
diff --git a/lib/libGL/gallium/drivers/llvmpipe/Makefile b/lib/libGL/gallium/drivers/llvmpipe/Makefile
new file mode 100644
index 000000000..5953cebaa
--- /dev/null
+++ b/lib/libGL/gallium/drivers/llvmpipe/Makefile
@@ -0,0 +1,56 @@
+# $OpenBSD: Makefile,v 1.1.1.1 2012/07/13 12:43:34 mpi Exp $
+
+LIB = llvmpipe
+
+SRCS = \
+	lp_bld_alpha.c \
+	lp_bld_blend_aos.c \
+	lp_bld_blend_logicop.c \
+	lp_bld_blend_soa.c \
+	lp_bld_depth.c \
+	lp_bld_interp.c \
+	lp_clear.c \
+	lp_context.c \
+	lp_draw_arrays.c \
+	lp_fence.c \
+	lp_flush.c \
+	lp_jit.c \
+	lp_memory.c \
+	lp_perf.c \
+	lp_query.c \
+	lp_rast.c \
+	lp_rast_debug.c \
+	lp_rast_tri.c \
+	lp_scene.c \
+	lp_scene_queue.c \
+	lp_screen.c \
+	lp_setup.c \
+	lp_setup_line.c \
+	lp_setup_point.c \
+	lp_setup_tri.c \
+	lp_setup_vbuf.c \
+	lp_state_blend.c \
+	lp_state_clip.c \
+	lp_state_derived.c \
+	lp_state_fs.c \
+	lp_state_setup.c \
+	lp_state_gs.c \
+	lp_state_rasterizer.c \
+	lp_state_sampler.c \
+        lp_state_so.c \
+	lp_state_surface.c \
+	lp_state_vertex.c \
+	lp_state_vs.c \
+	lp_surface.c \
+	lp_tex_sample.c \
+	lp_texture.c \
+	lp_tile_image.c \
+	lp_tile_soa.c
+
+CPPFLAGS+= -I${GALLIUM}/drivers/llvmpipe
+
+lp_test_sincos.o: sse_mathfun.h
+
+.include <bsd.xorg.mk>
+
+.PATH: ${.CURDIR}/generated
diff --git a/lib/libGL/gallium/drivers/llvmpipe/generated/Makefile b/lib/libGL/gallium/drivers/llvmpipe/generated/Makefile
new file mode 100644
index 000000000..afd593ed7
--- /dev/null
+++ b/lib/libGL/gallium/drivers/llvmpipe/generated/Makefile
@@ -0,0 +1,30 @@
+# $OpenBSD: Makefile,v 1.1.1.1 2012/07/13 12:43:34 mpi Exp $
+
+AUX=	${.CURDIR}/../../../../../../dist/Mesa/src/gallium/auxiliary
+DEV=	${.CURDIR}/../../../../../../dist/Mesa/src/gallium/drivers
+
+PYTHON=		python${PYTHON_VERSION}
+PYTHON_FLAGS=	-t -O -O
+
+GENERATED=	lp_tile_soa.c	
+
+all: ${GENERATED}
+
+obj:
+
+depend:
+
+install:
+
+clean distclean:
+	rm -f ${GENERATED}
+
+lp_tile_soa.c: lp_tile_soa.py u_format_parse.py u_format_pack.py u_format.csv
+	$(PYTHON) $(PYTHON_FLAGS) ${DEV}/llvmpipe/lp_tile_soa.py ${AUX}/util/u_format.csv > $@
+
+.SUFFIXES: .py .csv
+
+.PATH: ${DEV}/llvmpipe
+.PATH: ${AUX}/util
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/drivers/llvmpipe/generated/lp_tile_soa.c b/lib/libGL/gallium/drivers/llvmpipe/generated/lp_tile_soa.c
new file mode 100644
index 000000000..936bad915
--- /dev/null
+++ b/lib/libGL/gallium/drivers/llvmpipe/generated/lp_tile_soa.c
@@ -0,0 +1,4936 @@
+/* This file is autogenerated by lp_tile_soa.py from u_format.csv. Do not edit directly. */
+
+/**************************************************************************
+ *
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * Pixel format accessor functions.
+ *
+ * @author Jose Fonseca <jfonseca@vmware.com>
+ */
+
+#include "pipe/p_compiler.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
+#include "util/u_half.h"
+#include "util/u_cpu_detect.h"
+#include "lp_tile_soa.h"
+
+#ifdef DEBUG
+unsigned lp_tile_unswizzle_count = 0;
+unsigned lp_tile_swizzle_count = 0;
+#endif
+
+const unsigned char
+tile_offset[TILE_VECTOR_HEIGHT][TILE_VECTOR_WIDTH] = {
+   {  0,  1,  4,  5},
+   {  2,  3,  6,  7},
+   {  8,  9, 12, 13},
+   { 10, 11, 14, 15}
+};
+
+/* Note: these lookup tables could be replaced with some
+ * bit-twiddling code, but this is a little faster.
+ */
+static unsigned tile_x_offset[TILE_VECTOR_WIDTH * TILE_VECTOR_HEIGHT] = {
+   0, 1, 0, 1, 2, 3, 2, 3,
+   0, 1, 0, 1, 2, 3, 2, 3
+};
+
+static unsigned tile_y_offset[TILE_VECTOR_WIDTH * TILE_VECTOR_HEIGHT] = {
+   0, 0, 1, 1, 0, 0, 1, 1,
+   2, 2, 3, 3, 2, 2, 3, 3
+};
+
+
+#if defined(PIPE_ARCH_SSE)
+
+#include "util/u_sse.h"
+
+static ALWAYS_INLINE void 
+swz4( const __m128i * restrict x, 
+      const __m128i * restrict y, 
+      const __m128i * restrict z, 
+      const __m128i * restrict w, 
+      __m128i * restrict a, 
+      __m128i * restrict b, 
+      __m128i * restrict c, 
+      __m128i * restrict d)
+{
+   __m128i i, j, k, l;
+   __m128i m, n, o, p;
+   __m128i e, f, g, h;
+
+   m = _mm_unpacklo_epi8(*x,*y);
+   n = _mm_unpackhi_epi8(*x,*y);
+   o = _mm_unpacklo_epi8(*z,*w);
+   p = _mm_unpackhi_epi8(*z,*w);
+
+   i = _mm_unpacklo_epi16(m,n);
+   j = _mm_unpackhi_epi16(m,n);
+   k = _mm_unpacklo_epi16(o,p);
+   l = _mm_unpackhi_epi16(o,p);
+
+   e = _mm_unpacklo_epi8(i,j);
+   f = _mm_unpackhi_epi8(i,j);
+   g = _mm_unpacklo_epi8(k,l);
+   h = _mm_unpackhi_epi8(k,l);
+
+   *a = _mm_unpacklo_epi64(e,g);
+   *b = _mm_unpackhi_epi64(e,g);
+   *c = _mm_unpacklo_epi64(f,h);
+   *d = _mm_unpackhi_epi64(f,h);
+}
+
+static ALWAYS_INLINE void
+unswz4( const __m128i * restrict a, 
+        const __m128i * restrict b, 
+        const __m128i * restrict c, 
+        const __m128i * restrict d, 
+        __m128i * restrict x, 
+        __m128i * restrict y, 
+        __m128i * restrict z, 
+        __m128i * restrict w)
+{
+   __m128i i, j, k, l;
+   __m128i m, n, o, p;
+
+   i = _mm_unpacklo_epi8(*a,*b);
+   j = _mm_unpackhi_epi8(*a,*b);
+   k = _mm_unpacklo_epi8(*c,*d);
+   l = _mm_unpackhi_epi8(*c,*d);
+
+   m = _mm_unpacklo_epi16(i,k);
+   n = _mm_unpackhi_epi16(i,k);
+   o = _mm_unpacklo_epi16(j,l);
+   p = _mm_unpackhi_epi16(j,l);
+
+   *x = _mm_unpacklo_epi64(m,n);
+   *y = _mm_unpackhi_epi64(m,n);
+   *z = _mm_unpacklo_epi64(o,p);
+   *w = _mm_unpackhi_epi64(o,p);
+}
+
+static void
+lp_tile_b8g8r8a8_unorm_swizzle_4ub_sse2(uint8_t * restrict dst,
+                                        const uint8_t * restrict src, unsigned src_stride,
+                                        unsigned x0, unsigned y0)
+{
+   __m128i *dst128 = (__m128i *) dst;
+   unsigned x, y;
+   
+   src += y0 * src_stride;
+   src += x0 * sizeof(uint32_t);
+
+   for (y = 0; y < TILE_SIZE; y += 4) {
+      const uint8_t *src_row = src;
+
+      for (x = 0; x < TILE_SIZE; x += 4) {
+         swz4((const __m128i *) (src_row + 0 * src_stride),
+              (const __m128i *) (src_row + 1 * src_stride),
+              (const __m128i *) (src_row + 2 * src_stride),
+              (const __m128i *) (src_row + 3 * src_stride),
+              dst128 + 2,     /* b */
+              dst128 + 1,     /* g */
+              dst128 + 0,     /* r */
+              dst128 + 3);    /* a */
+
+         dst128 += 4;
+         src_row += sizeof(__m128i);
+      }
+
+      src += 4 * src_stride;
+   }
+}
+
+static void
+lp_tile_b8g8r8a8_unorm_unswizzle_4ub_sse2(const uint8_t * restrict src,
+                                          uint8_t * restrict dst, unsigned dst_stride,
+                                          unsigned x0, unsigned y0)
+{
+   unsigned int x, y;
+   const __m128i *src128 = (const __m128i *) src;
+   
+   dst += y0 * dst_stride;
+   dst += x0 * sizeof(uint32_t);
+   
+   for (y = 0; y < TILE_SIZE; y += 4) {
+      const uint8_t *dst_row = dst;
+
+      for (x = 0; x < TILE_SIZE; x += 4) {
+         unswz4( &src128[2],     /* b */
+                 &src128[1],     /* g */
+                 &src128[0],     /* r */
+                 &src128[3],     /* a */
+                 (__m128i *) (dst_row + 0 * dst_stride),
+                 (__m128i *) (dst_row + 1 * dst_stride),
+                 (__m128i *) (dst_row + 2 * dst_stride),
+                 (__m128i *) (dst_row + 3 * dst_stride));
+
+         src128 += 4;
+         dst_row += sizeof(__m128i);;
+      }
+
+      dst += 4 * dst_stride;
+   }
+}
+
+#endif /* PIPE_ARCH_SSE */
+
+static void
+lp_tile_none_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b8g8r8a8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         uint8_t a;
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         a = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b8g8r8x8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         ++src_pixel;
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_a8r8g8b8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t a;
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         a = (*src_pixel++);
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_x8r8g8b8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         ++src_pixel;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_a8b8g8r8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t a;
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         a = (*src_pixel++);
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_x8b8g8r8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         ++src_pixel;
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8x8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         ++src_pixel;
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b5g5r5x1_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint16_t pixel = *src_pixel++;
+         uint8_t b = (uint8_t)(((uint32_t)(pixel & 0x1f)) * 0xff / 0x1f);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 5) & 0x1f)) * 0xff / 0x1f);
+         uint8_t r = (uint8_t)(((uint32_t)((pixel >> 10) & 0x1f)) * 0xff / 0x1f);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b5g5r5a1_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint16_t pixel = *src_pixel++;
+         uint8_t b = (uint8_t)(((uint32_t)(pixel & 0x1f)) * 0xff / 0x1f);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 5) & 0x1f)) * 0xff / 0x1f);
+         uint8_t r = (uint8_t)(((uint32_t)((pixel >> 10) & 0x1f)) * 0xff / 0x1f);
+         uint8_t a = (uint8_t)(((uint32_t)(pixel >> 15)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b4g4r4a4_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint16_t pixel = *src_pixel++;
+         uint8_t b = (uint8_t)(((uint32_t)(pixel & 0xf)) * 0xff / 0xf);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 4) & 0xf)) * 0xff / 0xf);
+         uint8_t r = (uint8_t)(((uint32_t)((pixel >> 8) & 0xf)) * 0xff / 0xf);
+         uint8_t a = (uint8_t)(((uint32_t)(pixel >> 12)) * 0xff / 0xf);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b4g4r4x4_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint16_t pixel = *src_pixel++;
+         uint8_t b = (uint8_t)(((uint32_t)(pixel & 0xf)) * 0xff / 0xf);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 4) & 0xf)) * 0xff / 0xf);
+         uint8_t r = (uint8_t)(((uint32_t)((pixel >> 8) & 0xf)) * 0xff / 0xf);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b5g6r5_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint16_t pixel = *src_pixel++;
+         uint8_t b = (uint8_t)(((uint32_t)(pixel & 0x1f)) * 0xff / 0x1f);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 5) & 0x3f)) * 0xff / 0x3f);
+         uint8_t r = (uint8_t)(((uint32_t)(pixel >> 11)) * 0xff / 0x1f);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r10g10b10a2_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = *src_pixel++;
+         uint8_t r = (uint8_t)((pixel & 0x3ff) >> 2);
+         uint8_t g = (uint8_t)(((pixel >> 10) & 0x3ff) >> 2);
+         uint8_t b = (uint8_t)(((pixel >> 20) & 0x3ff) >> 2);
+         uint8_t a = (uint8_t)(((uint32_t)(pixel >> 30)) * 0xff / 0x3);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b10g10r10a2_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = *src_pixel++;
+         uint8_t b = (uint8_t)((pixel & 0x3ff) >> 2);
+         uint8_t g = (uint8_t)(((pixel >> 10) & 0x3ff) >> 2);
+         uint8_t r = (uint8_t)(((pixel >> 20) & 0x3ff) >> 2);
+         uint8_t a = (uint8_t)(((uint32_t)(pixel >> 30)) * 0xff / 0x3);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_l8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t rgb;
+         rgb = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = rgb; /* r */
+         TILE_PIXEL(dst, x, y, 1) = rgb; /* g */
+         TILE_PIXEL(dst, x, y, 2) = rgb; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_a8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t a;
+         a = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = 0; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_i8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t rgba;
+         rgba = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = rgba; /* r */
+         TILE_PIXEL(dst, x, y, 1) = rgba; /* g */
+         TILE_PIXEL(dst, x, y, 2) = rgba; /* b */
+         TILE_PIXEL(dst, x, y, 3) = rgba; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_l4a4_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t pixel = *src_pixel++;
+         uint8_t rgb = (uint8_t)(((uint32_t)(pixel & 0xf)) * 0xff / 0xf);
+         uint8_t a = (uint8_t)(((uint32_t)(pixel >> 4)) * 0xff / 0xf);
+         TILE_PIXEL(dst, x, y, 0) = rgb; /* r */
+         TILE_PIXEL(dst, x, y, 1) = rgb; /* g */
+         TILE_PIXEL(dst, x, y, 2) = rgb; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_l8a8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t rgb;
+         uint8_t a;
+         rgb = (*src_pixel++);
+         a = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = rgb; /* r */
+         TILE_PIXEL(dst, x, y, 1) = rgb; /* g */
+         TILE_PIXEL(dst, x, y, 2) = rgb; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_l16_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t rgb;
+         rgb = (uint8_t)((*src_pixel++) >> 8);
+         TILE_PIXEL(dst, x, y, 0) = rgb; /* r */
+         TILE_PIXEL(dst, x, y, 1) = rgb; /* g */
+         TILE_PIXEL(dst, x, y, 2) = rgb; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_l8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t rgb;
+         rgb = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = rgb; /* r */
+         TILE_PIXEL(dst, x, y, 1) = rgb; /* g */
+         TILE_PIXEL(dst, x, y, 2) = rgb; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_l8a8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t rgb;
+         uint8_t a;
+         rgb = (*src_pixel++);
+         a = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = rgb; /* r */
+         TILE_PIXEL(dst, x, y, 1) = rgb; /* g */
+         TILE_PIXEL(dst, x, y, 2) = rgb; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         a = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_a8b8g8r8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t a;
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         a = (*src_pixel++);
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_x8b8g8r8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         ++src_pixel;
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b8g8r8a8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         uint8_t a;
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         a = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_b8g8r8x8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t b;
+         uint8_t g;
+         uint8_t r;
+         b = (*src_pixel++);
+         g = (*src_pixel++);
+         r = (*src_pixel++);
+         ++src_pixel;
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_a8r8g8b8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t a;
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         a = (*src_pixel++);
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_x8r8g8b8_srgb_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         ++src_pixel;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8sg8sb8ux8u_norm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         b = (*src_pixel++);
+         ++src_pixel;
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r10sg10sb10sa2u_norm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = *src_pixel++;
+         uint8_t r = (uint8_t)((pixel & 0x3ff) >> 1);
+         uint8_t g = (uint8_t)(((pixel >> 10) & 0x3ff) >> 1);
+         uint8_t b = (uint8_t)(((pixel >> 20) & 0x3ff) >> 1);
+         uint8_t a = (uint8_t)(((uint32_t)(pixel >> 30)) * 0xff / 0x3);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r5sg5sb6u_norm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint16_t pixel = *src_pixel++;
+         uint8_t r = (uint8_t)(((uint32_t)(pixel & 0x1f)) * 0xff / 0xf);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 5) & 0x1f)) * 0xff / 0xf);
+         uint8_t b = (uint8_t)(((uint32_t)(pixel >> 10)) * 0xff / 0x3f);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r10g10b10a2_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = *src_pixel++;
+         uint8_t r = (uint8_t)(((uint32_t)(pixel & 0x3ff)) * 0xff / 0x1);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 10) & 0x3ff)) * 0xff / 0x1);
+         uint8_t b = (uint8_t)(((uint32_t)((pixel >> 20) & 0x3ff)) * 0xff / 0x1);
+         uint8_t a = (uint8_t)(((uint32_t)(pixel >> 30)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r64_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const double *src_pixel = (const double *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)((*src_pixel++) * 0xff);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r64g64_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const double *src_pixel = (const double *)(src_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)((*src_pixel++) * 0xff);
+         g = (uint8_t)((*src_pixel++) * 0xff);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r64g64b64_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const double *src_pixel = (const double *)(src_row + x0*24);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)((*src_pixel++) * 0xff);
+         g = (uint8_t)((*src_pixel++) * 0xff);
+         b = (uint8_t)((*src_pixel++) * 0xff);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r64g64b64a64_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const double *src_pixel = (const double *)(src_row + x0*32);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)((*src_pixel++) * 0xff);
+         g = (uint8_t)((*src_pixel++) * 0xff);
+         b = (uint8_t)((*src_pixel++) * 0xff);
+         a = (uint8_t)((*src_pixel++) * 0xff);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const float *src_pixel = (const float *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = float_to_ubyte((*src_pixel++));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const float *src_pixel = (const float *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = float_to_ubyte((*src_pixel++));
+         g = float_to_ubyte((*src_pixel++));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const float *src_pixel = (const float *)(src_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = float_to_ubyte((*src_pixel++));
+         g = float_to_ubyte((*src_pixel++));
+         b = float_to_ubyte((*src_pixel++));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const float *src_pixel = (const float *)(src_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = float_to_ubyte((*src_pixel++));
+         g = float_to_ubyte((*src_pixel++));
+         b = float_to_ubyte((*src_pixel++));
+         a = float_to_ubyte((*src_pixel++));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)((*src_pixel++) >> 24);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)((*src_pixel++) >> 24);
+         g = (uint8_t)((*src_pixel++) >> 24);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)((*src_pixel++) >> 24);
+         g = (uint8_t)((*src_pixel++) >> 24);
+         b = (uint8_t)((*src_pixel++) >> 24);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)((*src_pixel++) >> 24);
+         g = (uint8_t)((*src_pixel++) >> 24);
+         b = (uint8_t)((*src_pixel++) >> 24);
+         a = (uint8_t)((*src_pixel++) >> 24);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         a = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)((*src_pixel++) >> 23);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)((*src_pixel++) >> 23);
+         g = (uint8_t)((*src_pixel++) >> 23);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)((*src_pixel++) >> 23);
+         g = (uint8_t)((*src_pixel++) >> 23);
+         b = (uint8_t)((*src_pixel++) >> 23);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)((*src_pixel++) >> 23);
+         g = (uint8_t)((*src_pixel++) >> 23);
+         b = (uint8_t)((*src_pixel++) >> 23);
+         a = (uint8_t)((*src_pixel++) >> 23);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int32_t *src_pixel = (const int32_t *)(src_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         a = (uint8_t)(((uint64_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         g = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         g = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         b = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_float_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         g = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         b = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         a = float_to_ubyte(util_half_to_float((*src_pixel++)));
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)((*src_pixel++) >> 8);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)((*src_pixel++) >> 8);
+         g = (uint8_t)((*src_pixel++) >> 8);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)((*src_pixel++) >> 8);
+         g = (uint8_t)((*src_pixel++) >> 8);
+         b = (uint8_t)((*src_pixel++) >> 8);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)((*src_pixel++) >> 8);
+         g = (uint8_t)((*src_pixel++) >> 8);
+         b = (uint8_t)((*src_pixel++) >> 8);
+         a = (uint8_t)((*src_pixel++) >> 8);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint16_t *src_pixel = (const uint16_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         a = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)((*src_pixel++) >> 7);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)((*src_pixel++) >> 7);
+         g = (uint8_t)((*src_pixel++) >> 7);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)((*src_pixel++) >> 7);
+         g = (uint8_t)((*src_pixel++) >> 7);
+         b = (uint8_t)((*src_pixel++) >> 7);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)((*src_pixel++) >> 7);
+         g = (uint8_t)((*src_pixel++) >> 7);
+         b = (uint8_t)((*src_pixel++) >> 7);
+         a = (uint8_t)((*src_pixel++) >> 7);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int16_t *src_pixel = (const int16_t *)(src_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         a = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_unorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (*src_pixel++);
+         g = (*src_pixel++);
+         b = (*src_pixel++);
+         a = (*src_pixel++);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint8_t *src_pixel = (const uint8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         a = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         a = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x7f);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = 0; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = 0; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_sscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const int8_t *src_pixel = (const int8_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint8_t r;
+         uint8_t g;
+         uint8_t b;
+         uint8_t a;
+         r = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         g = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         b = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         a = (uint8_t)(((uint32_t)(*src_pixel++)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = a; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r10g10b10x2_uscaled_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = *src_pixel++;
+         uint8_t r = (uint8_t)(((uint32_t)(pixel & 0x3ff)) * 0xff / 0x1);
+         uint8_t g = (uint8_t)(((uint32_t)((pixel >> 10) & 0x3ff)) * 0xff / 0x1);
+         uint8_t b = (uint8_t)(((uint32_t)((pixel >> 20) & 0x3ff)) * 0xff / 0x1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+static void
+lp_tile_r10g10b10x2_snorm_swizzle_4ub(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   const uint8_t *src_row = src + y0*src_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      const uint32_t *src_pixel = (const uint32_t *)(src_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = *src_pixel++;
+         uint8_t r = (uint8_t)((pixel & 0x3ff) >> 1);
+         uint8_t g = (uint8_t)(((pixel >> 10) & 0x3ff) >> 1);
+         uint8_t b = (uint8_t)(((pixel >> 20) & 0x3ff) >> 1);
+         TILE_PIXEL(dst, x, y, 0) = r; /* r */
+         TILE_PIXEL(dst, x, y, 1) = g; /* g */
+         TILE_PIXEL(dst, x, y, 2) = b; /* b */
+         TILE_PIXEL(dst, x, y, 3) = 255; /* a */
+      }
+      src_row += src_stride;
+   }
+}
+
+void
+lp_tile_swizzle_4ub(enum pipe_format format, uint8_t *dst, const void *src, unsigned src_stride, unsigned x, unsigned y)
+{
+   void (*func)(uint8_t *dst, const uint8_t *src, unsigned src_stride, unsigned x0, unsigned y0);
+#ifdef DEBUG
+   lp_tile_swizzle_count += 1;
+#endif
+   switch(format) {
+   case PIPE_FORMAT_NONE:
+      func = lp_tile_none_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B8G8R8A8_UNORM:
+#ifdef PIPE_ARCH_SSE
+      func = util_cpu_caps.has_sse2 ? lp_tile_b8g8r8a8_unorm_swizzle_4ub_sse2 : lp_tile_b8g8r8a8_unorm_swizzle_4ub;
+#else
+      func = lp_tile_b8g8r8a8_unorm_swizzle_4ub;
+#endif
+      break;
+   case PIPE_FORMAT_B8G8R8X8_UNORM:
+      func = lp_tile_b8g8r8x8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8R8G8B8_UNORM:
+      func = lp_tile_a8r8g8b8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8R8G8B8_UNORM:
+      func = lp_tile_x8r8g8b8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8B8G8R8_UNORM:
+      func = lp_tile_a8b8g8r8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8B8G8R8_UNORM:
+      func = lp_tile_x8b8g8r8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8X8_UNORM:
+      func = lp_tile_r8g8b8x8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B5G5R5X1_UNORM:
+      func = lp_tile_b5g5r5x1_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B5G5R5A1_UNORM:
+      func = lp_tile_b5g5r5a1_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B4G4R4A4_UNORM:
+      func = lp_tile_b4g4r4a4_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B4G4R4X4_UNORM:
+      func = lp_tile_b4g4r4x4_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B5G6R5_UNORM:
+      func = lp_tile_b5g6r5_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10A2_UNORM:
+      func = lp_tile_r10g10b10a2_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B10G10R10A2_UNORM:
+      func = lp_tile_b10g10r10a2_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8_UNORM:
+      func = lp_tile_l8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8_UNORM:
+      func = lp_tile_a8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_I8_UNORM:
+      func = lp_tile_i8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_L4A4_UNORM:
+      func = lp_tile_l4a4_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8A8_UNORM:
+      func = lp_tile_l8a8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_L16_UNORM:
+      func = lp_tile_l16_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8_SRGB:
+      func = lp_tile_l8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8A8_SRGB:
+      func = lp_tile_l8a8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_SRGB:
+      func = lp_tile_r8g8b8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_SRGB:
+      func = lp_tile_r8g8b8a8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8B8G8R8_SRGB:
+      func = lp_tile_a8b8g8r8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8B8G8R8_SRGB:
+      func = lp_tile_x8b8g8r8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B8G8R8A8_SRGB:
+      func = lp_tile_b8g8r8a8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_B8G8R8X8_SRGB:
+      func = lp_tile_b8g8r8x8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8R8G8B8_SRGB:
+      func = lp_tile_a8r8g8b8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8R8G8B8_SRGB:
+      func = lp_tile_x8r8g8b8_srgb_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8SG8SB8UX8U_NORM:
+      func = lp_tile_r8sg8sb8ux8u_norm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10SG10SB10SA2U_NORM:
+      func = lp_tile_r10sg10sb10sa2u_norm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R5SG5SB6U_NORM:
+      func = lp_tile_r5sg5sb6u_norm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10A2_USCALED:
+      func = lp_tile_r10g10b10a2_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64_FLOAT:
+      func = lp_tile_r64_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64G64_FLOAT:
+      func = lp_tile_r64g64_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64G64B64_FLOAT:
+      func = lp_tile_r64g64b64_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64G64B64A64_FLOAT:
+      func = lp_tile_r64g64b64a64_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_FLOAT:
+      func = lp_tile_r32_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_FLOAT:
+      func = lp_tile_r32g32_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_FLOAT:
+      func = lp_tile_r32g32b32_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_FLOAT:
+      func = lp_tile_r32g32b32a32_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_UNORM:
+      func = lp_tile_r32_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_UNORM:
+      func = lp_tile_r32g32_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_UNORM:
+      func = lp_tile_r32g32b32_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_UNORM:
+      func = lp_tile_r32g32b32a32_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_USCALED:
+      func = lp_tile_r32_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_USCALED:
+      func = lp_tile_r32g32_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_USCALED:
+      func = lp_tile_r32g32b32_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_USCALED:
+      func = lp_tile_r32g32b32a32_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_SNORM:
+      func = lp_tile_r32_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_SNORM:
+      func = lp_tile_r32g32_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_SNORM:
+      func = lp_tile_r32g32b32_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_SNORM:
+      func = lp_tile_r32g32b32a32_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_SSCALED:
+      func = lp_tile_r32_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_SSCALED:
+      func = lp_tile_r32g32_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_SSCALED:
+      func = lp_tile_r32g32b32_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_SSCALED:
+      func = lp_tile_r32g32b32a32_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_FLOAT:
+      func = lp_tile_r16_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_FLOAT:
+      func = lp_tile_r16g16_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_FLOAT:
+      func = lp_tile_r16g16b16_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_FLOAT:
+      func = lp_tile_r16g16b16a16_float_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_UNORM:
+      func = lp_tile_r16_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_UNORM:
+      func = lp_tile_r16g16_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_UNORM:
+      func = lp_tile_r16g16b16_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_UNORM:
+      func = lp_tile_r16g16b16a16_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_USCALED:
+      func = lp_tile_r16_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_USCALED:
+      func = lp_tile_r16g16_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_USCALED:
+      func = lp_tile_r16g16b16_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_USCALED:
+      func = lp_tile_r16g16b16a16_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_SNORM:
+      func = lp_tile_r16_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_SNORM:
+      func = lp_tile_r16g16_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_SNORM:
+      func = lp_tile_r16g16b16_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_SNORM:
+      func = lp_tile_r16g16b16a16_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_SSCALED:
+      func = lp_tile_r16_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_SSCALED:
+      func = lp_tile_r16g16_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_SSCALED:
+      func = lp_tile_r16g16b16_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_SSCALED:
+      func = lp_tile_r16g16b16a16_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_UNORM:
+      func = lp_tile_r8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_UNORM:
+      func = lp_tile_r8g8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_UNORM:
+      func = lp_tile_r8g8b8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_UNORM:
+      func = lp_tile_r8g8b8a8_unorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_USCALED:
+      func = lp_tile_r8_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_USCALED:
+      func = lp_tile_r8g8_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_USCALED:
+      func = lp_tile_r8g8b8_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_USCALED:
+      func = lp_tile_r8g8b8a8_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_SNORM:
+      func = lp_tile_r8_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_SNORM:
+      func = lp_tile_r8g8_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_SNORM:
+      func = lp_tile_r8g8b8_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_SNORM:
+      func = lp_tile_r8g8b8a8_snorm_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_SSCALED:
+      func = lp_tile_r8_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_SSCALED:
+      func = lp_tile_r8g8_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_SSCALED:
+      func = lp_tile_r8g8b8_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_SSCALED:
+      func = lp_tile_r8g8b8a8_sscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10X2_USCALED:
+      func = lp_tile_r10g10b10x2_uscaled_swizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10X2_SNORM:
+      func = lp_tile_r10g10b10x2_snorm_swizzle_4ub;
+      break;
+   default:
+      debug_printf("%s: unsupported format %s\n", __FUNCTION__, util_format_name(format));
+      return;
+   }
+   func(dst, (const uint8_t *)src, src_stride, x, y);
+}
+
+static void
+lp_tile_none_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 1;
+   uint8_t *dstpix = (uint8_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint8_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0);
+            const uint32_t pixel1 = (((uint8_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b8g8r8a8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((b[i+0]) << 0) | ((g[i+0]) << 8) | ((r[i+0]) << 16) | ((a[i+0]) << 24);
+            const uint32_t pixel1 = ((b[i+1]) << 0) | ((g[i+1]) << 8) | ((r[i+1]) << 16) | ((a[i+1]) << 24);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b8g8r8x8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((b[i+0]) << 0) | ((g[i+0]) << 8) | ((r[i+0]) << 16);
+            const uint32_t pixel1 = ((b[i+1]) << 0) | ((g[i+1]) << 8) | ((r[i+1]) << 16);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_a8r8g8b8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((a[i+0]) << 0) | ((r[i+0]) << 8) | ((g[i+0]) << 16) | ((b[i+0]) << 24);
+            const uint32_t pixel1 = ((a[i+1]) << 0) | ((r[i+1]) << 8) | ((g[i+1]) << 16) | ((b[i+1]) << 24);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_x8r8g8b8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((r[i+0]) << 8) | ((g[i+0]) << 16) | ((b[i+0]) << 24);
+            const uint32_t pixel1 = ((r[i+1]) << 8) | ((g[i+1]) << 16) | ((b[i+1]) << 24);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_a8b8g8r8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((a[i+0]) << 0) | ((b[i+0]) << 8) | ((g[i+0]) << 16) | ((r[i+0]) << 24);
+            const uint32_t pixel1 = ((a[i+1]) << 0) | ((b[i+1]) << 8) | ((g[i+1]) << 16) | ((r[i+1]) << 24);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_x8b8g8r8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((b[i+0]) << 8) | ((g[i+0]) << 16) | ((r[i+0]) << 24);
+            const uint32_t pixel1 = ((b[i+1]) << 8) | ((g[i+1]) << 16) | ((r[i+1]) << 24);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r8g8b8x8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((r[i+0]) << 0) | ((g[i+0]) << 8) | ((b[i+0]) << 16);
+            const uint32_t pixel1 = ((r[i+1]) << 0) | ((g[i+1]) << 8) | ((b[i+1]) << 16);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b5g5r5x1_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(b[i+0] >> 3)) << 0) | (((uint16_t)(g[i+0] >> 3)) << 5) | (((uint16_t)(r[i+0] >> 3)) << 10);
+            const uint32_t pixel1 = (((uint16_t)(b[i+1] >> 3)) << 0) | (((uint16_t)(g[i+1] >> 3)) << 5) | (((uint16_t)(r[i+1] >> 3)) << 10);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b5g5r5a1_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(b[i+0] >> 3)) << 0) | (((uint16_t)(g[i+0] >> 3)) << 5) | (((uint16_t)(r[i+0] >> 3)) << 10) | (((uint16_t)(a[i+0] >> 7)) << 15);
+            const uint32_t pixel1 = (((uint16_t)(b[i+1] >> 3)) << 0) | (((uint16_t)(g[i+1] >> 3)) << 5) | (((uint16_t)(r[i+1] >> 3)) << 10) | (((uint16_t)(a[i+1] >> 7)) << 15);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b4g4r4a4_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(b[i+0] >> 4)) << 0) | (((uint16_t)(g[i+0] >> 4)) << 4) | (((uint16_t)(r[i+0] >> 4)) << 8) | (((uint16_t)(a[i+0] >> 4)) << 12);
+            const uint32_t pixel1 = (((uint16_t)(b[i+1] >> 4)) << 0) | (((uint16_t)(g[i+1] >> 4)) << 4) | (((uint16_t)(r[i+1] >> 4)) << 8) | (((uint16_t)(a[i+1] >> 4)) << 12);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b4g4r4x4_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(b[i+0] >> 4)) << 0) | (((uint16_t)(g[i+0] >> 4)) << 4) | (((uint16_t)(r[i+0] >> 4)) << 8);
+            const uint32_t pixel1 = (((uint16_t)(b[i+1] >> 4)) << 0) | (((uint16_t)(g[i+1] >> 4)) << 4) | (((uint16_t)(r[i+1] >> 4)) << 8);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b5g6r5_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(b[i+0] >> 3)) << 0) | (((uint16_t)(g[i+0] >> 2)) << 5) | (((uint16_t)(r[i+0] >> 3)) << 11);
+            const uint32_t pixel1 = (((uint16_t)(b[i+1] >> 3)) << 0) | (((uint16_t)(g[i+1] >> 2)) << 5) | (((uint16_t)(r[i+1] >> 3)) << 11);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r10g10b10a2_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint32_t)(((uint32_t)r[i+0]) * 0x3ff / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+0]) * 0x3ff / 0xff)) << 10) | (((uint32_t)(((uint32_t)b[i+0]) * 0x3ff / 0xff)) << 20) | (((uint32_t)(a[i+0] >> 6)) << 30);
+            const uint32_t pixel1 = (((uint32_t)(((uint32_t)r[i+1]) * 0x3ff / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+1]) * 0x3ff / 0xff)) << 10) | (((uint32_t)(((uint32_t)b[i+1]) * 0x3ff / 0xff)) << 20) | (((uint32_t)(a[i+1] >> 6)) << 30);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_b10g10r10a2_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint32_t)(((uint32_t)b[i+0]) * 0x3ff / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+0]) * 0x3ff / 0xff)) << 10) | (((uint32_t)(((uint32_t)r[i+0]) * 0x3ff / 0xff)) << 20) | (((uint32_t)(a[i+0] >> 6)) << 30);
+            const uint32_t pixel1 = (((uint32_t)(((uint32_t)b[i+1]) * 0x3ff / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+1]) * 0x3ff / 0xff)) << 10) | (((uint32_t)(((uint32_t)r[i+1]) * 0x3ff / 0xff)) << 20) | (((uint32_t)(a[i+1] >> 6)) << 30);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_l8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 1;
+   uint8_t *dstpix = (uint8_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((b[i+0]) << 0);
+            const uint32_t pixel1 = ((b[i+1]) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_a8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 1;
+   uint8_t *dstpix = (uint8_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((a[i+0]) << 0);
+            const uint32_t pixel1 = ((a[i+1]) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_i8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 1;
+   uint8_t *dstpix = (uint8_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((a[i+0]) << 0);
+            const uint32_t pixel1 = ((a[i+1]) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_l4a4_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 1;
+   uint8_t *dstpix = (uint8_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint8_t)(b[i+0] >> 4)) << 0) | (((uint8_t)(a[i+0] >> 4)) << 4);
+            const uint32_t pixel1 = (((uint8_t)(b[i+1] >> 4)) << 0) | (((uint8_t)(a[i+1] >> 4)) << 4);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_l8a8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((b[i+0]) << 0) | ((a[i+0]) << 8);
+            const uint32_t pixel1 = ((b[i+1]) << 0) | ((a[i+1]) << 8);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_l16_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(((uint32_t)b[i+0]) * 0xffff / 0xff)) << 0);
+            const uint32_t pixel1 = (((uint16_t)(((uint32_t)b[i+1]) * 0xffff / 0xff)) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_l8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_l8a8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 3);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 3);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_a8b8g8r8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 3);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_x8b8g8r8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         ++dst_pixel;
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_b8g8r8a8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 3);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_b8g8r8x8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+         ++dst_pixel;
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_a8r8g8b8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 3);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_x8r8g8b8_srgb_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         ++dst_pixel;
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8sg8sb8ux8u_norm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 0) >> 1);
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 1) >> 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+         ++dst_pixel;
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r10sg10sb10sa2u_norm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = 0;
+         pixel |= (uint32_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1ff / 0xff);
+         pixel |= ((uint32_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1ff / 0xff) << 10);
+         pixel |= ((uint32_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1ff / 0xff) << 20);
+         pixel |= ((uint32_t)(TILE_PIXEL(src, x, y, 3) >> 6) << 30);
+         *dst_pixel++ = pixel;
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r5sg5sb6u_norm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint16_t pixel = 0;
+         pixel |= (uint16_t)(TILE_PIXEL(src, x, y, 0) >> 4);
+         pixel |= ((uint16_t)(TILE_PIXEL(src, x, y, 1) >> 4) << 5);
+         pixel |= ((uint16_t)(TILE_PIXEL(src, x, y, 2) >> 2) << 10);
+         *dst_pixel++ = pixel;
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r10g10b10a2_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint32_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+0]) * 0x1 / 0xff)) << 10) | (((uint32_t)(((uint32_t)b[i+0]) * 0x1 / 0xff)) << 20) | (((uint32_t)(((uint32_t)a[i+0]) * 0x1 / 0xff)) << 30);
+            const uint32_t pixel1 = (((uint32_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+1]) * 0x1 / 0xff)) << 10) | (((uint32_t)(((uint32_t)b[i+1]) * 0x1 / 0xff)) << 20) | (((uint32_t)(((uint32_t)a[i+1]) * 0x1 / 0xff)) << 30);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r64_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      double *dst_pixel = (double *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r64g64_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      double *dst_pixel = (double *)(dst_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff));
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 1) * (1.0f/0xff));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r64g64b64_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      double *dst_pixel = (double *)(dst_row + x0*24);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff));
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 1) * (1.0f/0xff));
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 2) * (1.0f/0xff));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r64g64b64a64_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      double *dst_pixel = (double *)(dst_row + x0*32);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff));
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 1) * (1.0f/0xff));
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 2) * (1.0f/0xff));
+         *dst_pixel++ = (double)(TILE_PIXEL(src, x, y, 3) * (1.0f/0xff));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      float *dst_pixel = (float *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 0));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      float *dst_pixel = (float *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 0));
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 1));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      float *dst_pixel = (float *)(dst_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 0));
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 1));
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 2));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      float *dst_pixel = (float *)(dst_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 0));
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 1));
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 2));
+         *dst_pixel++ = ubyte_to_float(TILE_PIXEL(src, x, y, 3));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint32_t)(((uint64_t)r[i+0]) * 0xffffffff / 0xff)) << 0);
+            const uint32_t pixel1 = (((uint32_t)(((uint64_t)r[i+1]) * 0xffffffff / 0xff)) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r32g32_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0xffffffff / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0xffffffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0xffffffff / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0xffffffff / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0xffffffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0xffffffff / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0xffffffff / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0xffffffff / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 3)) * 0xffffffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint32_t)(((uint64_t)r[i+0]) * 0x1 / 0xff)) << 0);
+            const uint32_t pixel1 = (((uint32_t)(((uint64_t)r[i+1]) * 0x1 / 0xff)) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r32g32_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint32_t)(((uint64_t)TILE_PIXEL(src, x, y, 3)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x7fffffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x7fffffff / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x7fffffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x7fffffff / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x7fffffff / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0x7fffffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x7fffffff / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x7fffffff / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0x7fffffff / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 3)) * 0x7fffffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*12);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r32g32b32a32_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int32_t *dst_pixel = (int32_t *)(dst_row + x0*16);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+         *dst_pixel++ = (int32_t)(((uint64_t)TILE_PIXEL(src, x, y, 3)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff)));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff)));
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 1) * (1.0f/0xff)));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff)));
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 1) * (1.0f/0xff)));
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 2) * (1.0f/0xff)));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_float_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 0) * (1.0f/0xff)));
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 1) * (1.0f/0xff)));
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 2) * (1.0f/0xff)));
+         *dst_pixel++ = util_float_to_half((float)(TILE_PIXEL(src, x, y, 3) * (1.0f/0xff)));
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(((uint32_t)r[i+0]) * 0xffff / 0xff)) << 0);
+            const uint32_t pixel1 = (((uint16_t)(((uint32_t)r[i+1]) * 0xffff / 0xff)) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r16g16_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(((uint32_t)r[i+0]) * 0xffff / 0xff)) << 0) | (((uint16_t)(((uint32_t)g[i+0]) * 0xffff / 0xff)) << 16);
+            const uint32_t pixel1 = (((uint16_t)(((uint32_t)r[i+1]) * 0xffff / 0xff)) << 0) | (((uint16_t)(((uint32_t)g[i+1]) * 0xffff / 0xff)) << 16);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r16g16b16_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0xffff / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0xffff / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0xffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0xffff / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0xffff / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0xffff / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 3)) * 0xffff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0);
+            const uint32_t pixel1 = (((uint16_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r16g16_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint16_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0) | (((uint16_t)(((uint32_t)g[i+0]) * 0x1 / 0xff)) << 16);
+            const uint32_t pixel1 = (((uint16_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0) | (((uint16_t)(((uint32_t)g[i+1]) * 0x1 / 0xff)) << 16);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r16g16b16_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint16_t *dst_pixel = (uint16_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint16_t)(((uint32_t)TILE_PIXEL(src, x, y, 3)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x7fff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x7fff / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x7fff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x7fff / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x7fff / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x7fff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x7fff / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x7fff / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x7fff / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 3)) * 0x7fff / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*6);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r16g16b16a16_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int16_t *dst_pixel = (int16_t *)(dst_row + x0*8);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+         *dst_pixel++ = (int16_t)(((uint32_t)TILE_PIXEL(src, x, y, 3)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 1;
+   uint8_t *dstpix = (uint8_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((r[i+0]) << 0);
+            const uint32_t pixel1 = ((r[i+1]) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r8g8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((r[i+0]) << 0) | ((g[i+0]) << 8);
+            const uint32_t pixel1 = ((r[i+1]) << 0) | ((g[i+1]) << 8);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r8g8b8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 0);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 1);
+         *dst_pixel++ = TILE_PIXEL(src, x, y, 2);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_unorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = ((r[i+0]) << 0) | ((g[i+0]) << 8) | ((b[i+0]) << 16) | ((a[i+0]) << 24);
+            const uint32_t pixel1 = ((r[i+1]) << 0) | ((g[i+1]) << 8) | ((b[i+1]) << 16) | ((a[i+1]) << 24);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r8_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 1;
+   uint8_t *dstpix = (uint8_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint8_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0);
+            const uint32_t pixel1 = (((uint8_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r8g8_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 2;
+   uint16_t *dstpix = (uint16_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint8_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0) | (((uint8_t)(((uint32_t)g[i+0]) * 0x1 / 0xff)) << 8);
+            const uint32_t pixel1 = (((uint8_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0) | (((uint8_t)(((uint32_t)g[i+1]) * 0x1 / 0xff)) << 8);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r8g8b8_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint8_t *dst_pixel = (uint8_t *)(dst_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (uint8_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint8_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (uint8_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint8_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0) | (((uint8_t)(((uint32_t)g[i+0]) * 0x1 / 0xff)) << 8) | (((uint8_t)(((uint32_t)b[i+0]) * 0x1 / 0xff)) << 16) | (((uint8_t)(((uint32_t)a[i+0]) * 0x1 / 0xff)) << 24);
+            const uint32_t pixel1 = (((uint8_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0) | (((uint8_t)(((uint32_t)g[i+1]) * 0x1 / 0xff)) << 8) | (((uint8_t)(((uint32_t)b[i+1]) * 0x1 / 0xff)) << 16) | (((uint8_t)(((uint32_t)a[i+1]) * 0x1 / 0xff)) << 24);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r8_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 0) >> 1);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 0) >> 1);
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 1) >> 1);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 0) >> 1);
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 1) >> 1);
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 2) >> 1);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 0) >> 1);
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 1) >> 1);
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 2) >> 1);
+         *dst_pixel++ = (int8_t)(TILE_PIXEL(src, x, y, 3) >> 1);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*1);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*2);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*3);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r8g8b8a8_sscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      int8_t *dst_pixel = (int8_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1 / 0xff);
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1 / 0xff);
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1 / 0xff);
+         *dst_pixel++ = (int8_t)(((uint32_t)TILE_PIXEL(src, x, y, 3)) * 0x1 / 0xff);
+      }
+      dst_row += dst_stride;
+   }
+}
+
+static void
+lp_tile_r10g10b10x2_uscaled_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   const unsigned dstpix_stride = dst_stride / 4;
+   uint32_t *dstpix = (uint32_t *) dst;
+   unsigned int qx, qy, i;
+
+   for (qy = 0; qy < TILE_SIZE; qy += TILE_VECTOR_HEIGHT) {
+      const unsigned py = y0 + qy;
+      for (qx = 0; qx < TILE_SIZE; qx += TILE_VECTOR_WIDTH) {
+         const unsigned px = x0 + qx;
+         const uint8_t *r = src + 0 * TILE_C_STRIDE;
+         const uint8_t *g = src + 1 * TILE_C_STRIDE;
+         const uint8_t *b = src + 2 * TILE_C_STRIDE;
+         const uint8_t *a = src + 3 * TILE_C_STRIDE;
+         (void) r; (void) g; (void) b; (void) a; /* silence warnings */
+         for (i = 0; i < TILE_C_STRIDE; i += 2) {
+            const uint32_t pixel0 = (((uint32_t)(((uint32_t)r[i+0]) * 0x1 / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+0]) * 0x1 / 0xff)) << 10) | (((uint32_t)(((uint32_t)b[i+0]) * 0x1 / 0xff)) << 20);
+            const uint32_t pixel1 = (((uint32_t)(((uint32_t)r[i+1]) * 0x1 / 0xff)) << 0) | (((uint32_t)(((uint32_t)g[i+1]) * 0x1 / 0xff)) << 10) | (((uint32_t)(((uint32_t)b[i+1]) * 0x1 / 0xff)) << 20);
+            const unsigned offset = (py + tile_y_offset[i]) * dstpix_stride + (px + tile_x_offset[i]);
+            dstpix[offset + 0] = pixel0;
+            dstpix[offset + 1] = pixel1;
+         }
+         src += TILE_X_STRIDE;
+      }
+   }
+}
+
+static void
+lp_tile_r10g10b10x2_snorm_unswizzle_4ub(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0)
+{
+   unsigned x, y;
+   uint8_t *dst_row = dst + y0*dst_stride;
+   for (y = 0; y < TILE_SIZE; ++y) {
+      uint32_t *dst_pixel = (uint32_t *)(dst_row + x0*4);
+      for (x = 0; x < TILE_SIZE; ++x) {
+         uint32_t pixel = 0;
+         pixel |= (uint32_t)(((uint32_t)TILE_PIXEL(src, x, y, 0)) * 0x1ff / 0xff);
+         pixel |= ((uint32_t)(((uint32_t)TILE_PIXEL(src, x, y, 1)) * 0x1ff / 0xff) << 10);
+         pixel |= ((uint32_t)(((uint32_t)TILE_PIXEL(src, x, y, 2)) * 0x1ff / 0xff) << 20);
+         *dst_pixel++ = pixel;
+      }
+      dst_row += dst_stride;
+   }
+}
+
+void
+lp_tile_unswizzle_4ub(enum pipe_format format, const uint8_t *src, void *dst, unsigned dst_stride, unsigned x, unsigned y)
+{
+   void (*func)(const uint8_t *src, uint8_t *dst, unsigned dst_stride, unsigned x0, unsigned y0);
+#ifdef DEBUG
+   lp_tile_unswizzle_count += 1;
+#endif
+   switch(format) {
+   case PIPE_FORMAT_NONE:
+      func = lp_tile_none_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B8G8R8A8_UNORM:
+#ifdef PIPE_ARCH_SSE
+      func = util_cpu_caps.has_sse2 ? lp_tile_b8g8r8a8_unorm_unswizzle_4ub_sse2 : lp_tile_b8g8r8a8_unorm_unswizzle_4ub;
+#else
+      func = lp_tile_b8g8r8a8_unorm_unswizzle_4ub;
+#endif
+      break;
+   case PIPE_FORMAT_B8G8R8X8_UNORM:
+      func = lp_tile_b8g8r8x8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8R8G8B8_UNORM:
+      func = lp_tile_a8r8g8b8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8R8G8B8_UNORM:
+      func = lp_tile_x8r8g8b8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8B8G8R8_UNORM:
+      func = lp_tile_a8b8g8r8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8B8G8R8_UNORM:
+      func = lp_tile_x8b8g8r8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8X8_UNORM:
+      func = lp_tile_r8g8b8x8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B5G5R5X1_UNORM:
+      func = lp_tile_b5g5r5x1_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B5G5R5A1_UNORM:
+      func = lp_tile_b5g5r5a1_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B4G4R4A4_UNORM:
+      func = lp_tile_b4g4r4a4_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B4G4R4X4_UNORM:
+      func = lp_tile_b4g4r4x4_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B5G6R5_UNORM:
+      func = lp_tile_b5g6r5_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10A2_UNORM:
+      func = lp_tile_r10g10b10a2_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B10G10R10A2_UNORM:
+      func = lp_tile_b10g10r10a2_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8_UNORM:
+      func = lp_tile_l8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8_UNORM:
+      func = lp_tile_a8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_I8_UNORM:
+      func = lp_tile_i8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_L4A4_UNORM:
+      func = lp_tile_l4a4_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8A8_UNORM:
+      func = lp_tile_l8a8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_L16_UNORM:
+      func = lp_tile_l16_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8_SRGB:
+      func = lp_tile_l8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_L8A8_SRGB:
+      func = lp_tile_l8a8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_SRGB:
+      func = lp_tile_r8g8b8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_SRGB:
+      func = lp_tile_r8g8b8a8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8B8G8R8_SRGB:
+      func = lp_tile_a8b8g8r8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8B8G8R8_SRGB:
+      func = lp_tile_x8b8g8r8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B8G8R8A8_SRGB:
+      func = lp_tile_b8g8r8a8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_B8G8R8X8_SRGB:
+      func = lp_tile_b8g8r8x8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_A8R8G8B8_SRGB:
+      func = lp_tile_a8r8g8b8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_X8R8G8B8_SRGB:
+      func = lp_tile_x8r8g8b8_srgb_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8SG8SB8UX8U_NORM:
+      func = lp_tile_r8sg8sb8ux8u_norm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10SG10SB10SA2U_NORM:
+      func = lp_tile_r10sg10sb10sa2u_norm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R5SG5SB6U_NORM:
+      func = lp_tile_r5sg5sb6u_norm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10A2_USCALED:
+      func = lp_tile_r10g10b10a2_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64_FLOAT:
+      func = lp_tile_r64_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64G64_FLOAT:
+      func = lp_tile_r64g64_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64G64B64_FLOAT:
+      func = lp_tile_r64g64b64_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R64G64B64A64_FLOAT:
+      func = lp_tile_r64g64b64a64_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_FLOAT:
+      func = lp_tile_r32_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_FLOAT:
+      func = lp_tile_r32g32_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_FLOAT:
+      func = lp_tile_r32g32b32_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_FLOAT:
+      func = lp_tile_r32g32b32a32_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_UNORM:
+      func = lp_tile_r32_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_UNORM:
+      func = lp_tile_r32g32_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_UNORM:
+      func = lp_tile_r32g32b32_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_UNORM:
+      func = lp_tile_r32g32b32a32_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_USCALED:
+      func = lp_tile_r32_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_USCALED:
+      func = lp_tile_r32g32_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_USCALED:
+      func = lp_tile_r32g32b32_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_USCALED:
+      func = lp_tile_r32g32b32a32_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_SNORM:
+      func = lp_tile_r32_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_SNORM:
+      func = lp_tile_r32g32_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_SNORM:
+      func = lp_tile_r32g32b32_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_SNORM:
+      func = lp_tile_r32g32b32a32_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32_SSCALED:
+      func = lp_tile_r32_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32_SSCALED:
+      func = lp_tile_r32g32_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32_SSCALED:
+      func = lp_tile_r32g32b32_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R32G32B32A32_SSCALED:
+      func = lp_tile_r32g32b32a32_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_FLOAT:
+      func = lp_tile_r16_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_FLOAT:
+      func = lp_tile_r16g16_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_FLOAT:
+      func = lp_tile_r16g16b16_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_FLOAT:
+      func = lp_tile_r16g16b16a16_float_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_UNORM:
+      func = lp_tile_r16_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_UNORM:
+      func = lp_tile_r16g16_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_UNORM:
+      func = lp_tile_r16g16b16_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_UNORM:
+      func = lp_tile_r16g16b16a16_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_USCALED:
+      func = lp_tile_r16_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_USCALED:
+      func = lp_tile_r16g16_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_USCALED:
+      func = lp_tile_r16g16b16_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_USCALED:
+      func = lp_tile_r16g16b16a16_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_SNORM:
+      func = lp_tile_r16_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_SNORM:
+      func = lp_tile_r16g16_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_SNORM:
+      func = lp_tile_r16g16b16_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_SNORM:
+      func = lp_tile_r16g16b16a16_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16_SSCALED:
+      func = lp_tile_r16_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16_SSCALED:
+      func = lp_tile_r16g16_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16_SSCALED:
+      func = lp_tile_r16g16b16_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R16G16B16A16_SSCALED:
+      func = lp_tile_r16g16b16a16_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_UNORM:
+      func = lp_tile_r8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_UNORM:
+      func = lp_tile_r8g8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_UNORM:
+      func = lp_tile_r8g8b8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_UNORM:
+      func = lp_tile_r8g8b8a8_unorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_USCALED:
+      func = lp_tile_r8_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_USCALED:
+      func = lp_tile_r8g8_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_USCALED:
+      func = lp_tile_r8g8b8_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_USCALED:
+      func = lp_tile_r8g8b8a8_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_SNORM:
+      func = lp_tile_r8_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_SNORM:
+      func = lp_tile_r8g8_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_SNORM:
+      func = lp_tile_r8g8b8_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_SNORM:
+      func = lp_tile_r8g8b8a8_snorm_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8_SSCALED:
+      func = lp_tile_r8_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8_SSCALED:
+      func = lp_tile_r8g8_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8_SSCALED:
+      func = lp_tile_r8g8b8_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R8G8B8A8_SSCALED:
+      func = lp_tile_r8g8b8a8_sscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10X2_USCALED:
+      func = lp_tile_r10g10b10x2_uscaled_unswizzle_4ub;
+      break;
+   case PIPE_FORMAT_R10G10B10X2_SNORM:
+      func = lp_tile_r10g10b10x2_snorm_unswizzle_4ub;
+      break;
+   default:
+      debug_printf("%s: unsupported format %s\n", __FUNCTION__, util_format_name(format));
+      return;
+   }
+   func(src, (uint8_t *)dst, dst_stride, x, y);
+}
+
diff --git a/lib/libGL/gallium/drivers/rbug/Makefile b/lib/libGL/gallium/drivers/rbug/Makefile
new file mode 100644
index 000000000..b38ab1195
--- /dev/null
+++ b/lib/libGL/gallium/drivers/rbug/Makefile
@@ -0,0 +1,11 @@
+# $OpenBSD: Makefile,v 1.1.1.1 2012/07/13 12:43:34 mpi Exp $
+
+LIB = rbug
+
+SRCS = \
+	rbug_core.c \
+	rbug_context.c \
+	rbug_objects.c \
+	rbug_screen.c
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/drivers/softpipe/Makefile b/lib/libGL/gallium/drivers/softpipe/Makefile
new file mode 100644
index 000000000..7dd0edc8d
--- /dev/null
+++ b/lib/libGL/gallium/drivers/softpipe/Makefile
@@ -0,0 +1,37 @@
+# $OpenBSD: Makefile,v 1.1.1.1 2012/07/13 12:43:34 mpi Exp $
+
+LIB = softpipe
+
+SRCS = \
+	sp_fs_exec.c \
+	sp_fs_sse.c \
+	sp_clear.c \
+	sp_fence.c \
+	sp_flush.c \
+	sp_query.c \
+	sp_context.c \
+	sp_draw_arrays.c \
+	sp_prim_vbuf.c \
+	sp_quad_pipe.c \
+	sp_quad_stipple.c \
+	sp_quad_depth_test.c \
+	sp_quad_fs.c \
+	sp_quad_blend.c \
+	sp_screen.c \
+        sp_setup.c \
+	sp_state_blend.c \
+	sp_state_clip.c \
+	sp_state_derived.c \
+	sp_state_sampler.c \
+	sp_state_shader.c \
+	sp_state_so.c \
+	sp_state_rasterizer.c \
+	sp_state_surface.c \
+	sp_state_vertex.c \
+	sp_texture.c \
+	sp_tex_sample.c \
+	sp_tex_tile_cache.c \
+	sp_tile_cache.c \
+	sp_surface.c
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/drivers/trace/Makefile b/lib/libGL/gallium/drivers/trace/Makefile
new file mode 100644
index 000000000..14e36a840
--- /dev/null
+++ b/lib/libGL/gallium/drivers/trace/Makefile
@@ -0,0 +1,12 @@
+# $OpenBSD: Makefile,v 1.1.1.1 2012/07/13 12:43:34 mpi Exp $
+
+LIB = trace
+
+SRCS = \
+	tr_context.c \
+	tr_dump.c \
+	tr_dump_state.c \
+	tr_screen.c \
+	tr_texture.c
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/libgallium/Makefile b/lib/libGL/gallium/libgallium/Makefile
new file mode 100644
index 000000000..ca1023a3b
--- /dev/null
+++ b/lib/libGL/gallium/libgallium/Makefile
@@ -0,0 +1,221 @@
+# $OpenBSD: Makefile,v 1.1 2012/07/13 12:43:30 mpi Exp $
+
+LIB=	gallium
+
+GALLIUM=	${.CURDIR}/../../../../dist/Mesa/src/gallium
+
+CPPFLAGS+= \
+	-I${GALLIUM}/include \
+	-I${GALLIUM}/auxiliary \
+	-I${GALLIUM}/auxiliary/util
+
+C_SOURCES = \
+	cso_cache.c \
+	cso_context.c \
+	cso_hash.c \
+	draw_context.c \
+	draw_fs.c \
+	draw_gs.c \
+	draw_pipe.c \
+	draw_pipe_aaline.c \
+	draw_pipe_aapoint.c \
+	draw_pipe_clip.c \
+	draw_pipe_cull.c \
+	draw_pipe_flatshade.c \
+	draw_pipe_offset.c \
+	draw_pipe_pstipple.c \
+	draw_pipe_stipple.c \
+	draw_pipe_twoside.c \
+	draw_pipe_unfilled.c \
+	draw_pipe_util.c \
+	draw_pipe_validate.c \
+	draw_pipe_vbuf.c \
+	draw_pipe_wide_line.c \
+	draw_pipe_wide_point.c \
+	draw_pt.c \
+	draw_pt_emit.c \
+	draw_pt_fetch.c \
+	draw_pt_fetch_emit.c \
+	draw_pt_fetch_shade_emit.c \
+	draw_pt_fetch_shade_pipeline.c \
+	draw_pt_post_vs.c \
+	draw_pt_so_emit.c \
+	draw_pt_util.c \
+	draw_pt_vsplit.c \
+	draw_vertex.c \
+	draw_vs.c \
+	draw_vs_aos.c \
+	draw_vs_aos_io.c \
+	draw_vs_aos_machine.c \
+	draw_vs_exec.c \
+	draw_vs_ppc.c \
+	draw_vs_sse.c \
+	draw_vs_varient.c \
+	u_indices_gen.c \
+	u_unfilled_gen.c \
+	os_misc.c \
+	os_stream.c \
+	os_stream_log.c \
+	os_stream_null.c \
+	os_stream_stdc.c \
+	os_stream_str.c \
+	os_time.c \
+	pb_buffer_fenced.c \
+	pb_buffer_malloc.c \
+	pb_bufmgr_alt.c \
+	pb_bufmgr_cache.c \
+	pb_bufmgr_debug.c \
+	pb_bufmgr_mm.c \
+	pb_bufmgr_ondemand.c \
+	pb_bufmgr_pool.c \
+	pb_bufmgr_slab.c \
+	pb_validate.c \
+	rbug_connection.c \
+	rbug_context.c \
+	rbug_core.c \
+	rbug_demarshal.c \
+	rbug_texture.c \
+	rbug_shader.c \
+	rtasm_cpu.c \
+	rtasm_execmem.c \
+	rtasm_ppc.c \
+	rtasm_ppc_spe.c \
+	rtasm_x86sse.c \
+	tgsi_build.c \
+	tgsi_dump.c \
+	tgsi_exec.c \
+	tgsi_info.c \
+	tgsi_iterate.c \
+	tgsi_parse.c \
+	tgsi_ppc.c \
+	tgsi_sanity.c \
+	tgsi_scan.c \
+	tgsi_sse2.c \
+	tgsi_text.c \
+	tgsi_transform.c \
+	tgsi_ureg.c \
+	tgsi_util.c \
+	translate.c \
+	translate_cache.c \
+	translate_generic.c \
+	translate_sse.c \
+	u_debug.c \
+	u_debug_describe.c \
+	u_debug_refcnt.c \
+	u_debug_stack.c \
+	u_debug_symbol.c \
+	u_dump_defines.c \
+	u_dump_state.c \
+	u_bitmask.c \
+	u_blit.c \
+	u_blitter.c \
+	u_cache.c \
+	u_caps.c \
+	u_cpu_detect.c \
+	u_dl.c \
+	u_draw_quad.c \
+	u_format.c \
+	u_format_other.c \
+	u_format_s3tc.c \
+	u_format_srgb.c \
+	u_format_table.c \
+	u_format_tests.c \
+	u_format_yuv.c \
+	u_format_zs.c \
+	u_framebuffer.c \
+	u_gen_mipmap.c \
+	u_half.c \
+	u_handle_table.c \
+	u_hash.c \
+	u_hash_table.c \
+	u_index_modify.c \
+	u_keymap.c \
+	u_linear.c \
+	u_linkage.c \
+	u_network.c \
+	u_math.c \
+	u_mm.c \
+	u_rect.c \
+	u_ringbuffer.c \
+	u_sampler.c \
+	u_simple_shaders.c \
+	u_slab.c \
+	u_snprintf.c \
+	u_staging.c \
+	u_surface.c \
+	u_surfaces.c \
+	u_texture.c \
+	u_tile.c \
+	u_transfer.c \
+	u_resource.c \
+	u_upload_mgr.c
+
+GALLIVM_SOURCES = \
+        lp_bld_arit.c \
+        lp_bld_assert.c \
+        lp_bld_bitarit.c \
+        lp_bld_const.c \
+        lp_bld_conv.c \
+        lp_bld_debug.c \
+        lp_bld_flow.c \
+        lp_bld_format_aos.c \
+        lp_bld_format_soa.c \
+        lp_bld_format_yuv.c \
+        lp_bld_gather.c \
+        lp_bld_init.c \
+        lp_bld_intr.c \
+        lp_bld_logic.c \
+        lp_bld_pack.c \
+        lp_bld_printf.c \
+        lp_bld_quad.c \
+        lp_bld_sample.c \
+        lp_bld_sample_aos.c \
+        lp_bld_sample_soa.c \
+        lp_bld_struct.c \
+        lp_bld_swizzle.c \
+        lp_bld_tgsi_aos.c \
+        lp_bld_tgsi_info.c \
+        lp_bld_tgsi_soa.c \
+        lp_bld_type.c \
+        draw_llvm.c \
+        draw_llvm_sample.c \
+        draw_llvm_translate.c \
+        draw_vs_llvm.c \
+        draw_pt_fetch_shade_pipeline_llvm.c
+
+GALLIVM_CPP_SOURCES = \
+	lp_bld_misc.cpp
+
+SRCS=	${C_SOURCES}
+
+.ifdef XENOCARA_HAVE_LLVM
+SRCS+=	${GALLIVM_SOURCES} \
+	${GALLIVM_CPP_SOURCES}
+.endif
+
+all: lib${LIB}_pic.a
+
+obj: _xenocara_obj
+
+install:
+	@echo "Not installing lib${LIB}"
+
+clean: 
+	rm -f lib${LIB}_pic.a ${OBJS}
+
+cleandir:	clean
+
+.include <bsd.xorg.mk>
+
+.PATH: ${GALLIUM}/auxiliary/cso_cache
+.PATH: ${GALLIUM}/auxiliary/draw
+.PATH: ${GALLIUM}/auxiliary/indices
+.PATH: ${GALLIUM}/auxiliary/os
+.PATH: ${GALLIUM}/auxiliary/pipebuffer
+.PATH: ${GALLIUM}/auxiliary/rbug
+.PATH: ${GALLIUM}/auxiliary/rtasm
+.PATH: ${GALLIUM}/auxiliary/tgsi
+.PATH: ${GALLIUM}/auxiliary/translate
+.PATH: ${GALLIUM}/auxiliary/util
+.PATH: ${GALLIUM}/auxiliary/gallivm
+.PATH: ${.CURDIR}/generated
diff --git a/lib/libGL/gallium/libgallium/generated/Makefile b/lib/libGL/gallium/libgallium/generated/Makefile
new file mode 100644
index 000000000..bc047a40c
--- /dev/null
+++ b/lib/libGL/gallium/libgallium/generated/Makefile
@@ -0,0 +1,42 @@
+# $OpenBSD: Makefile,v 1.1.1.1 2012/07/13 12:43:30 mpi Exp $
+
+AUX=	${.CURDIR}/../../../../../dist/Mesa/src/gallium/auxiliary
+
+PYTHON=		python${PYTHON_VERSION}
+PYTHON_FLAGS=	-t -O -O
+
+GENERATED=	u_indices_gen.c u_unfilled_gen.c u_format_srgb.c \
+		u_format_table.c u_half.c
+
+all: ${GENERATED}
+
+obj:
+
+depend:
+
+install:
+
+clean distclean:
+	rm -f ${GENERATED}
+
+u_indices_gen.c: u_indices_gen.py
+	${PYTHON} ${PYTHON_FLAGS} ${AUX}/indices/u_indices_gen.py > $@
+
+u_unfilled_gen.c: u_unfilled_gen.py
+	${PYTHON} ${PYTHON_FLAGS} ${AUX}/indices/u_unfilled_gen.py > $@
+
+u_format_srgb.c: ${AUX}/util/u_format_srgb.py
+	${PYTHON} ${PYTHON_FLAGS} ${AUX}/util/u_format_srgb.py > $@
+
+u_format_table.c: u_format_table.py u_format_pack.py u_format_parse.py u_format.csv
+	${PYTHON} ${PYTHON_FLAGS} ${AUX}/util/u_format_table.py ${AUX}/util/u_format.csv > $@
+
+u_half.c: ${AUX}/util/u_half.py
+	${PYTHON} ${PYTHON_FLAGS} ${AUX}/util/u_half.py > $@
+
+.SUFFIXES: .py
+
+.PATH: ${AUX}/indices
+.PATH: ${AUX}/util
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/libgallium/generated/u_format_srgb.c b/lib/libGL/gallium/libgallium/generated/u_format_srgb.c
new file mode 100644
index 000000000..4ac8c7b48
--- /dev/null
+++ b/lib/libGL/gallium/libgallium/generated/u_format_srgb.c
@@ -0,0 +1,148 @@
+/* This file is autogenerated by u_format_srgb.py. Do not edit directly. */
+
+/**************************************************************************
+ *
+ * Copyright 2010 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * SRGB translation.
+ *
+ * @author Brian Paul <brianp@vmware.com>
+ * @author Michal Krol <michal@vmware.com>
+ * @author Jose Fonseca <jfonseca@vmware.com>
+ */
+
+#include "u_format_srgb.h"
+
+const float
+util_format_srgb_8unorm_to_linear_float_table[256] = {
+    0.0000000e+00, 3.0352698e-04, 6.0705397e-04, 9.1058095e-04,
+    1.2141079e-03, 1.5176349e-03, 1.8211619e-03, 2.1246889e-03,
+    2.4282159e-03, 2.7317429e-03, 3.0352698e-03, 3.3465358e-03,
+    3.6765073e-03, 4.0247170e-03, 4.3914420e-03, 4.7769535e-03,
+    5.1815167e-03, 5.6053916e-03, 6.0488330e-03, 6.5120908e-03,
+    6.9954102e-03, 7.4990320e-03, 8.0231930e-03, 8.5681256e-03,
+    9.1340587e-03, 9.7212173e-03, 1.0329823e-02, 1.0960094e-02,
+    1.1612245e-02, 1.2286488e-02, 1.2983032e-02, 1.3702083e-02,
+    1.4443844e-02, 1.5208514e-02, 1.5996293e-02, 1.6807376e-02,
+    1.7641954e-02, 1.8500220e-02, 1.9382361e-02, 2.0288563e-02,
+    2.1219010e-02, 2.2173885e-02, 2.3153366e-02, 2.4157632e-02,
+    2.5186860e-02, 2.6241222e-02, 2.7320892e-02, 2.8426040e-02,
+    2.9556834e-02, 3.0713444e-02, 3.1896033e-02, 3.3104767e-02,
+    3.4339807e-02, 3.5601315e-02, 3.6889450e-02, 3.8204372e-02,
+    3.9546235e-02, 4.0915197e-02, 4.2311411e-02, 4.3735029e-02,
+    4.5186204e-02, 4.6665086e-02, 4.8171824e-02, 4.9706566e-02,
+    5.1269458e-02, 5.2860647e-02, 5.4480276e-02, 5.6128490e-02,
+    5.7805430e-02, 5.9511238e-02, 6.1246054e-02, 6.3010018e-02,
+    6.4803267e-02, 6.6625939e-02, 6.8478170e-02, 7.0360096e-02,
+    7.2271851e-02, 7.4213568e-02, 7.6185381e-02, 7.8187422e-02,
+    8.0219820e-02, 8.2282707e-02, 8.4376212e-02, 8.6500462e-02,
+    8.8655586e-02, 9.0841711e-02, 9.3058963e-02, 9.5307467e-02,
+    9.7587347e-02, 9.9898728e-02, 1.0224173e-01, 1.0461648e-01,
+    1.0702310e-01, 1.0946171e-01, 1.1193243e-01, 1.1443537e-01,
+    1.1697067e-01, 1.1953843e-01, 1.2213877e-01, 1.2477182e-01,
+    1.2743768e-01, 1.3013648e-01, 1.3286832e-01, 1.3563333e-01,
+    1.3843162e-01, 1.4126329e-01, 1.4412847e-01, 1.4702727e-01,
+    1.4995979e-01, 1.5292615e-01, 1.5592646e-01, 1.5896084e-01,
+    1.6202938e-01, 1.6513219e-01, 1.6826940e-01, 1.7144110e-01,
+    1.7464740e-01, 1.7788842e-01, 1.8116424e-01, 1.8447499e-01,
+    1.8782077e-01, 1.9120168e-01, 1.9461783e-01, 1.9806932e-01,
+    2.0155625e-01, 2.0507874e-01, 2.0863687e-01, 2.1223076e-01,
+    2.1586050e-01, 2.1952620e-01, 2.2322796e-01, 2.2696587e-01,
+    2.3074005e-01, 2.3455058e-01, 2.3839757e-01, 2.4228112e-01,
+    2.4620133e-01, 2.5015828e-01, 2.5415209e-01, 2.5818285e-01,
+    2.6225066e-01, 2.6635560e-01, 2.7049779e-01, 2.7467731e-01,
+    2.7889426e-01, 2.8314874e-01, 2.8744084e-01, 2.9177065e-01,
+    2.9613827e-01, 3.0054379e-01, 3.0498731e-01, 3.0946892e-01,
+    3.1398871e-01, 3.1854678e-01, 3.2314321e-01, 3.2777810e-01,
+    3.3245154e-01, 3.3716362e-01, 3.4191442e-01, 3.4670406e-01,
+    3.5153260e-01, 3.5640014e-01, 3.6130678e-01, 3.6625260e-01,
+    3.7123768e-01, 3.7626212e-01, 3.8132601e-01, 3.8642943e-01,
+    3.9157248e-01, 3.9675523e-01, 4.0197778e-01, 4.0724021e-01,
+    4.1254261e-01, 4.1788507e-01, 4.2326767e-01, 4.2869050e-01,
+    4.3415364e-01, 4.3965717e-01, 4.4520119e-01, 4.5078578e-01,
+    4.5641102e-01, 4.6207700e-01, 4.6778380e-01, 4.7353150e-01,
+    4.7932018e-01, 4.8514994e-01, 4.9102085e-01, 4.9693300e-01,
+    5.0288646e-01, 5.0888132e-01, 5.1491767e-01, 5.2099557e-01,
+    5.2711513e-01, 5.3327640e-01, 5.3947949e-01, 5.4572446e-01,
+    5.5201140e-01, 5.5834039e-01, 5.6471151e-01, 5.7112483e-01,
+    5.7758044e-01, 5.8407842e-01, 5.9061884e-01, 5.9720179e-01,
+    6.0382734e-01, 6.1049557e-01, 6.1720656e-01, 6.2396039e-01,
+    6.3075714e-01, 6.3759687e-01, 6.4447968e-01, 6.5140564e-01,
+    6.5837482e-01, 6.6538730e-01, 6.7244316e-01, 6.7954247e-01,
+    6.8668531e-01, 6.9387176e-01, 7.0110189e-01, 7.0837578e-01,
+    7.1569350e-01, 7.2305513e-01, 7.3046074e-01, 7.3791041e-01,
+    7.4540421e-01, 7.5294222e-01, 7.6052450e-01, 7.6815115e-01,
+    7.7582222e-01, 7.8353779e-01, 7.9129794e-01, 7.9910274e-01,
+    8.0695226e-01, 8.1484657e-01, 8.2278575e-01, 8.3076988e-01,
+    8.3879901e-01, 8.4687323e-01, 8.5499261e-01, 8.6315721e-01,
+    8.7136712e-01, 8.7962240e-01, 8.8792312e-01, 8.9626935e-01,
+    9.0466117e-01, 9.1309865e-01, 9.2158186e-01, 9.3011086e-01,
+    9.3868573e-01, 9.4730654e-01, 9.5597335e-01, 9.6468625e-01,
+    9.7344529e-01, 9.8225055e-01, 9.9110210e-01, 1.0000000e+00,
+};
+
+const uint8_t
+util_format_srgb_to_linear_8unorm_table[256] = {
+      0,   0,   0,   0,   0,   0,   0,   1,   1,   1,   1,   1,   1,   1,   1,   1,
+      1,   1,   2,   2,   2,   2,   2,   2,   2,   2,   3,   3,   3,   3,   3,   3,
+      4,   4,   4,   4,   4,   5,   5,   5,   5,   6,   6,   6,   6,   7,   7,   7,
+      8,   8,   8,   8,   9,   9,   9,  10,  10,  10,  11,  11,  12,  12,  12,  13,
+     13,  13,  14,  14,  15,  15,  16,  16,  17,  17,  17,  18,  18,  19,  19,  20,
+     20,  21,  22,  22,  23,  23,  24,  24,  25,  25,  26,  27,  27,  28,  29,  29,
+     30,  30,  31,  32,  32,  33,  34,  35,  35,  36,  37,  37,  38,  39,  40,  41,
+     41,  42,  43,  44,  45,  45,  46,  47,  48,  49,  50,  51,  51,  52,  53,  54,
+     55,  56,  57,  58,  59,  60,  61,  62,  63,  64,  65,  66,  67,  68,  69,  70,
+     71,  72,  73,  74,  76,  77,  78,  79,  80,  81,  82,  84,  85,  86,  87,  88,
+     90,  91,  92,  93,  95,  96,  97,  99, 100, 101, 103, 104, 105, 107, 108, 109,
+    111, 112, 114, 115, 116, 118, 119, 121, 122, 124, 125, 127, 128, 130, 131, 133,
+    134, 136, 138, 139, 141, 142, 144, 146, 147, 149, 151, 152, 154, 156, 157, 159,
+    161, 163, 164, 166, 168, 170, 171, 173, 175, 177, 179, 181, 183, 184, 186, 188,
+    190, 192, 194, 196, 198, 200, 202, 204, 206, 208, 210, 212, 214, 216, 218, 220,
+    222, 224, 226, 229, 231, 233, 235, 237, 239, 242, 244, 246, 248, 250, 253, 255,
+};
+
+const uint8_t
+util_format_linear_to_srgb_8unorm_table[256] = {
+      0,  13,  22,  28,  34,  38,  42,  46,  50,  53,  56,  59,  61,  64,  66,  69,
+     71,  73,  75,  77,  79,  81,  83,  85,  86,  88,  90,  92,  93,  95,  96,  98,
+     99, 101, 102, 104, 105, 106, 108, 109, 110, 112, 113, 114, 115, 117, 118, 119,
+    120, 121, 122, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136,
+    137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 148, 149, 150, 151,
+    152, 153, 154, 155, 155, 156, 157, 158, 159, 159, 160, 161, 162, 163, 163, 164,
+    165, 166, 167, 167, 168, 169, 170, 170, 171, 172, 173, 173, 174, 175, 175, 176,
+    177, 178, 178, 179, 180, 180, 181, 182, 182, 183, 184, 185, 185, 186, 187, 187,
+    188, 189, 189, 190, 190, 191, 192, 192, 193, 194, 194, 195, 196, 196, 197, 197,
+    198, 199, 199, 200, 200, 201, 202, 202, 203, 203, 204, 205, 205, 206, 206, 207,
+    208, 208, 209, 209, 210, 210, 211, 212, 212, 213, 213, 214, 214, 215, 215, 216,
+    216, 217, 218, 218, 219, 219, 220, 220, 221, 221, 222, 222, 223, 223, 224, 224,
+    225, 226, 226, 227, 227, 228, 228, 229, 229, 230, 230, 231, 231, 232, 232, 233,
+    233, 234, 234, 235, 235, 236, 236, 237, 237, 238, 238, 238, 239, 239, 240, 240,
+    241, 241, 242, 242, 243, 243, 244, 244, 245, 245, 246, 246, 246, 247, 247, 248,
+    248, 249, 249, 250, 250, 251, 251, 251, 252, 252, 253, 253, 254, 254, 255, 255,
+};
+
diff --git a/lib/libGL/gallium/libgallium/generated/u_format_table.c b/lib/libGL/gallium/libgallium/generated/u_format_table.c
new file mode 100644
index 000000000..40dd09ebf
--- /dev/null
+++ b/lib/libGL/gallium/libgallium/generated/u_format_table.c
@@ -0,0 +1,17561 @@
+/* This file is autogenerated by u_format_table.py from u_format.csv. Do not edit directly. */
+
+/**************************************************************************
+ *
+ * Copyright 2010 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+#include "u_format.h"
+#include "u_format_s3tc.h"
+
+
+#include "pipe/p_compiler.h"
+#include "u_math.h"
+#include "u_half.h"
+#include "u_format.h"
+#include "u_format_other.h"
+#include "u_format_srgb.h"
+#include "u_format_yuv.h"
+#include "u_format_zs.h"
+
+union util_format_none {
+   uint8_t value;
+   struct {
+      uint8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_none_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_none_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)CLAMP(src[0], 0, 255);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_none_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_none_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_none_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b8g8r8a8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_b8g8r8a8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8a8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[2])) & 0xff;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[0])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[3])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8a8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_b8g8r8a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[2]) & 0xff;
+         value |= ((src[1]) & 0xff) << 8;
+         value |= ((src[0]) & 0xff) << 16;
+         value |= (src[3]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b8g8r8x8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+      uint8_t x;
+   } chan;
+};
+
+static INLINE void
+util_format_b8g8r8x8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8x8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[2])) & 0xff;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[0])) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8x8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_b8g8r8x8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8x8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[2]) & 0xff;
+         value |= ((src[1]) & 0xff) << 8;
+         value |= ((src[0]) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_a8r8g8b8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t a;
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_a8r8g8b8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8r8g8b8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[3])) & 0xff;
+         value |= ((float_to_ubyte(src[0])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[2])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_a8r8g8b8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_a8r8g8b8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8r8g8b8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[3]) & 0xff;
+         value |= ((src[0]) & 0xff) << 8;
+         value |= ((src[1]) & 0xff) << 16;
+         value |= (src[2]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_x8r8g8b8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t x;
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_x8r8g8b8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8r8g8b8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((float_to_ubyte(src[0])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[2])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_x8r8g8b8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_x8r8g8b8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8r8g8b8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((src[0]) & 0xff) << 8;
+         value |= ((src[1]) & 0xff) << 16;
+         value |= (src[2]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_a8b8g8r8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t a;
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_a8b8g8r8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8b8g8r8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[3])) & 0xff;
+         value |= ((float_to_ubyte(src[2])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[0])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_a8b8g8r8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_a8b8g8r8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8b8g8r8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[3]) & 0xff;
+         value |= ((src[2]) & 0xff) << 8;
+         value |= ((src[1]) & 0xff) << 16;
+         value |= (src[0]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_x8b8g8r8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t x;
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_x8b8g8r8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8b8g8r8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((float_to_ubyte(src[2])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[0])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_x8b8g8r8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_x8b8g8r8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8b8g8r8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((src[2]) & 0xff) << 8;
+         value |= ((src[1]) & 0xff) << 16;
+         value |= (src[0]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8x8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+      uint8_t x;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8x8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8x8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[0])) & 0xff;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[2])) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8x8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8x8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8x8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[0]) & 0xff;
+         value |= ((src[1]) & 0xff) << 8;
+         value |= ((src[2]) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b5g5r5x1_unorm {
+   uint16_t value;
+   struct {
+      unsigned b:5;
+      unsigned g:5;
+      unsigned r:5;
+      unsigned x:1;
+   } chan;
+};
+
+static INLINE void
+util_format_b5g5r5x1_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x1f;
+         r = (value >> 10) & 0x1f;
+         dst[0] = (float)(r * (1.0f/0x1f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1f)); /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b5g5r5x1_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(CLAMP(src[2], 0, 1) * 0x1f)) & 0x1f;
+         value |= (((uint16_t)(CLAMP(src[1], 0, 1) * 0x1f)) & 0x1f) << 5;
+         value |= (((uint16_t)(CLAMP(src[0], 0, 1) * 0x1f)) & 0x1f) << 10;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b5g5r5x1_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x1f;
+         r = (value >> 10) & 0x1f;
+         dst[0] = (float)(r * (1.0f/0x1f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1f)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_b5g5r5x1_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x1f;
+         r = (value >> 10) & 0x1f;
+         dst[0] = (uint8_t)(((uint32_t)r) * 0xff / 0x1f); /* r */
+         dst[1] = (uint8_t)(((uint32_t)g) * 0xff / 0x1f); /* g */
+         dst[2] = (uint8_t)(((uint32_t)b) * 0xff / 0x1f); /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b5g5r5x1_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(src[2] >> 3)) & 0x1f;
+         value |= (((uint16_t)(src[1] >> 3)) & 0x1f) << 5;
+         value |= (((uint16_t)(src[0] >> 3)) & 0x1f) << 10;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b5g5r5a1_unorm {
+   uint16_t value;
+   struct {
+      unsigned b:5;
+      unsigned g:5;
+      unsigned r:5;
+      unsigned a:1;
+   } chan;
+};
+
+static INLINE void
+util_format_b5g5r5a1_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x1f;
+         r = (value >> 10) & 0x1f;
+         a = value >> 15;
+         dst[0] = (float)(r * (1.0f/0x1f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1f)); /* b */
+         dst[3] = (float)(a * (1.0f/0x1)); /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b5g5r5a1_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(CLAMP(src[2], 0, 1) * 0x1f)) & 0x1f;
+         value |= (((uint16_t)(CLAMP(src[1], 0, 1) * 0x1f)) & 0x1f) << 5;
+         value |= (((uint16_t)(CLAMP(src[0], 0, 1) * 0x1f)) & 0x1f) << 10;
+         value |= ((uint16_t)(CLAMP(src[3], 0, 1) * 0x1)) << 15;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b5g5r5a1_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x1f;
+         r = (value >> 10) & 0x1f;
+         a = value >> 15;
+         dst[0] = (float)(r * (1.0f/0x1f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1f)); /* b */
+         dst[3] = (float)(a * (1.0f/0x1)); /* a */
+}
+
+static INLINE void
+util_format_b5g5r5a1_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x1f;
+         r = (value >> 10) & 0x1f;
+         a = value >> 15;
+         dst[0] = (uint8_t)(((uint32_t)r) * 0xff / 0x1f); /* r */
+         dst[1] = (uint8_t)(((uint32_t)g) * 0xff / 0x1f); /* g */
+         dst[2] = (uint8_t)(((uint32_t)b) * 0xff / 0x1f); /* b */
+         dst[3] = (uint8_t)(((uint32_t)a) * 0xff / 0x1); /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b5g5r5a1_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(src[2] >> 3)) & 0x1f;
+         value |= (((uint16_t)(src[1] >> 3)) & 0x1f) << 5;
+         value |= (((uint16_t)(src[0] >> 3)) & 0x1f) << 10;
+         value |= ((uint16_t)(src[3] >> 7)) << 15;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b4g4r4a4_unorm {
+   uint16_t value;
+   struct {
+      unsigned b:4;
+      unsigned g:4;
+      unsigned r:4;
+      unsigned a:4;
+   } chan;
+};
+
+static INLINE void
+util_format_b4g4r4a4_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0xf;
+         g = (value >> 4) & 0xf;
+         r = (value >> 8) & 0xf;
+         a = value >> 12;
+         dst[0] = (float)(r * (1.0f/0xf)); /* r */
+         dst[1] = (float)(g * (1.0f/0xf)); /* g */
+         dst[2] = (float)(b * (1.0f/0xf)); /* b */
+         dst[3] = (float)(a * (1.0f/0xf)); /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b4g4r4a4_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(CLAMP(src[2], 0, 1) * 0xf)) & 0xf;
+         value |= (((uint16_t)(CLAMP(src[1], 0, 1) * 0xf)) & 0xf) << 4;
+         value |= (((uint16_t)(CLAMP(src[0], 0, 1) * 0xf)) & 0xf) << 8;
+         value |= ((uint16_t)(CLAMP(src[3], 0, 1) * 0xf)) << 12;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b4g4r4a4_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0xf;
+         g = (value >> 4) & 0xf;
+         r = (value >> 8) & 0xf;
+         a = value >> 12;
+         dst[0] = (float)(r * (1.0f/0xf)); /* r */
+         dst[1] = (float)(g * (1.0f/0xf)); /* g */
+         dst[2] = (float)(b * (1.0f/0xf)); /* b */
+         dst[3] = (float)(a * (1.0f/0xf)); /* a */
+}
+
+static INLINE void
+util_format_b4g4r4a4_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0xf;
+         g = (value >> 4) & 0xf;
+         r = (value >> 8) & 0xf;
+         a = value >> 12;
+         dst[0] = (uint8_t)(((uint32_t)r) * 0xff / 0xf); /* r */
+         dst[1] = (uint8_t)(((uint32_t)g) * 0xff / 0xf); /* g */
+         dst[2] = (uint8_t)(((uint32_t)b) * 0xff / 0xf); /* b */
+         dst[3] = (uint8_t)(((uint32_t)a) * 0xff / 0xf); /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b4g4r4a4_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(src[2] >> 4)) & 0xf;
+         value |= (((uint16_t)(src[1] >> 4)) & 0xf) << 4;
+         value |= (((uint16_t)(src[0] >> 4)) & 0xf) << 8;
+         value |= ((uint16_t)(src[3] >> 4)) << 12;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b4g4r4x4_unorm {
+   uint16_t value;
+   struct {
+      unsigned b:4;
+      unsigned g:4;
+      unsigned r:4;
+      unsigned x:4;
+   } chan;
+};
+
+static INLINE void
+util_format_b4g4r4x4_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0xf;
+         g = (value >> 4) & 0xf;
+         r = (value >> 8) & 0xf;
+         dst[0] = (float)(r * (1.0f/0xf)); /* r */
+         dst[1] = (float)(g * (1.0f/0xf)); /* g */
+         dst[2] = (float)(b * (1.0f/0xf)); /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b4g4r4x4_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(CLAMP(src[2], 0, 1) * 0xf)) & 0xf;
+         value |= (((uint16_t)(CLAMP(src[1], 0, 1) * 0xf)) & 0xf) << 4;
+         value |= (((uint16_t)(CLAMP(src[0], 0, 1) * 0xf)) & 0xf) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b4g4r4x4_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0xf;
+         g = (value >> 4) & 0xf;
+         r = (value >> 8) & 0xf;
+         dst[0] = (float)(r * (1.0f/0xf)); /* r */
+         dst[1] = (float)(g * (1.0f/0xf)); /* g */
+         dst[2] = (float)(b * (1.0f/0xf)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_b4g4r4x4_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0xf;
+         g = (value >> 4) & 0xf;
+         r = (value >> 8) & 0xf;
+         dst[0] = (uint8_t)(((uint32_t)r) * 0xff / 0xf); /* r */
+         dst[1] = (uint8_t)(((uint32_t)g) * 0xff / 0xf); /* g */
+         dst[2] = (uint8_t)(((uint32_t)b) * 0xff / 0xf); /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b4g4r4x4_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(src[2] >> 4)) & 0xf;
+         value |= (((uint16_t)(src[1] >> 4)) & 0xf) << 4;
+         value |= (((uint16_t)(src[0] >> 4)) & 0xf) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b5g6r5_unorm {
+   uint16_t value;
+   struct {
+      unsigned b:5;
+      unsigned g:6;
+      unsigned r:5;
+   } chan;
+};
+
+static INLINE void
+util_format_b5g6r5_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x3f;
+         r = value >> 11;
+         dst[0] = (float)(r * (1.0f/0x1f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x3f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1f)); /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b5g6r5_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(CLAMP(src[2], 0, 1) * 0x1f)) & 0x1f;
+         value |= (((uint16_t)(CLAMP(src[1], 0, 1) * 0x3f)) & 0x3f) << 5;
+         value |= ((uint16_t)(CLAMP(src[0], 0, 1) * 0x1f)) << 11;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b5g6r5_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x3f;
+         r = value >> 11;
+         dst[0] = (float)(r * (1.0f/0x1f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x3f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1f)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_b5g6r5_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t b;
+         uint16_t g;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         b = (value) & 0x1f;
+         g = (value >> 5) & 0x3f;
+         r = value >> 11;
+         dst[0] = (uint8_t)(((uint32_t)r) * 0xff / 0x1f); /* r */
+         dst[1] = (uint8_t)(((uint32_t)g) * 0xff / 0x3f); /* g */
+         dst[2] = (uint8_t)(((uint32_t)b) * 0xff / 0x1f); /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b5g6r5_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint16_t)(src[2] >> 3)) & 0x1f;
+         value |= (((uint16_t)(src[1] >> 2)) & 0x3f) << 5;
+         value |= ((uint16_t)(src[0] >> 3)) << 11;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r10g10b10a2_unorm {
+   uint32_t value;
+   struct {
+      unsigned r:10;
+      unsigned g:10;
+      unsigned b:10;
+      unsigned a:2;
+   } chan;
+};
+
+static INLINE void
+util_format_r10g10b10a2_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (float)(r * (1.0f/0x3ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x3ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x3ff)); /* b */
+         dst[3] = (float)(a * (1.0f/0x3)); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10a2_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)(CLAMP(src[0], 0, 1) * 0x3ff)) & 0x3ff;
+         value |= (((uint32_t)(CLAMP(src[1], 0, 1) * 0x3ff)) & 0x3ff) << 10;
+         value |= (((uint32_t)(CLAMP(src[2], 0, 1) * 0x3ff)) & 0x3ff) << 20;
+         value |= ((uint32_t)(CLAMP(src[3], 0, 1) * 0x3)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10a2_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (float)(r * (1.0f/0x3ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x3ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x3ff)); /* b */
+         dst[3] = (float)(a * (1.0f/0x3)); /* a */
+}
+
+static INLINE void
+util_format_r10g10b10a2_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (uint8_t)(r >> 2); /* r */
+         dst[1] = (uint8_t)(g >> 2); /* g */
+         dst[2] = (uint8_t)(b >> 2); /* b */
+         dst[3] = (uint8_t)(((uint32_t)a) * 0xff / 0x3); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10a2_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)(((uint32_t)src[0]) * 0x3ff / 0xff)) & 0x3ff;
+         value |= (((uint32_t)(((uint32_t)src[1]) * 0x3ff / 0xff)) & 0x3ff) << 10;
+         value |= (((uint32_t)(((uint32_t)src[2]) * 0x3ff / 0xff)) & 0x3ff) << 20;
+         value |= ((uint32_t)(src[3] >> 6)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b10g10r10a2_unorm {
+   uint32_t value;
+   struct {
+      unsigned b:10;
+      unsigned g:10;
+      unsigned r:10;
+      unsigned a:2;
+   } chan;
+};
+
+static INLINE void
+util_format_b10g10r10a2_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         r = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (float)(r * (1.0f/0x3ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x3ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x3ff)); /* b */
+         dst[3] = (float)(a * (1.0f/0x3)); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b10g10r10a2_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)(CLAMP(src[2], 0, 1) * 0x3ff)) & 0x3ff;
+         value |= (((uint32_t)(CLAMP(src[1], 0, 1) * 0x3ff)) & 0x3ff) << 10;
+         value |= (((uint32_t)(CLAMP(src[0], 0, 1) * 0x3ff)) & 0x3ff) << 20;
+         value |= ((uint32_t)(CLAMP(src[3], 0, 1) * 0x3)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b10g10r10a2_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         r = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (float)(r * (1.0f/0x3ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x3ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x3ff)); /* b */
+         dst[3] = (float)(a * (1.0f/0x3)); /* a */
+}
+
+static INLINE void
+util_format_b10g10r10a2_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         r = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (uint8_t)(r >> 2); /* r */
+         dst[1] = (uint8_t)(g >> 2); /* g */
+         dst[2] = (uint8_t)(b >> 2); /* b */
+         dst[3] = (uint8_t)(((uint32_t)a) * 0xff / 0x3); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b10g10r10a2_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)(((uint32_t)src[2]) * 0x3ff / 0xff)) & 0x3ff;
+         value |= (((uint32_t)(((uint32_t)src[1]) * 0x3ff / 0xff)) & 0x3ff) << 10;
+         value |= (((uint32_t)(((uint32_t)src[0]) * 0x3ff / 0xff)) & 0x3ff) << 20;
+         value |= ((uint32_t)(src[3] >> 6)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_l8_unorm {
+   uint8_t value;
+   struct {
+      uint8_t rgb;
+   } chan;
+};
+
+static INLINE void
+util_format_l8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         rgb = value;
+         dst[0] = ubyte_to_float(rgb); /* r */
+         dst[1] = ubyte_to_float(rgb); /* g */
+         dst[2] = ubyte_to_float(rgb); /* b */
+         dst[3] = 1; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= float_to_ubyte(src[2]);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_l8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         rgb = value;
+         dst[0] = ubyte_to_float(rgb); /* r */
+         dst[1] = ubyte_to_float(rgb); /* g */
+         dst[2] = ubyte_to_float(rgb); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_l8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         rgb = value;
+         dst[0] = rgb; /* r */
+         dst[1] = rgb; /* g */
+         dst[2] = rgb; /* b */
+         dst[3] = 255; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= src[2];
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_a8_unorm {
+   uint8_t value;
+   struct {
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_a8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t a;
+         a = value;
+         dst[0] = 0; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= float_to_ubyte(src[3]);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_a8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t a;
+         a = value;
+         dst[0] = 0; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t a;
+         a = value;
+         dst[0] = 0; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = a; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= src[3];
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_i8_unorm {
+   uint8_t value;
+   struct {
+      uint8_t rgba;
+   } chan;
+};
+
+static INLINE void
+util_format_i8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgba;
+         rgba = value;
+         dst[0] = ubyte_to_float(rgba); /* r */
+         dst[1] = ubyte_to_float(rgba); /* g */
+         dst[2] = ubyte_to_float(rgba); /* b */
+         dst[3] = ubyte_to_float(rgba); /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_i8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= float_to_ubyte(src[3]);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_i8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgba;
+         rgba = value;
+         dst[0] = ubyte_to_float(rgba); /* r */
+         dst[1] = ubyte_to_float(rgba); /* g */
+         dst[2] = ubyte_to_float(rgba); /* b */
+         dst[3] = ubyte_to_float(rgba); /* a */
+}
+
+static INLINE void
+util_format_i8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgba;
+         rgba = value;
+         dst[0] = rgba; /* r */
+         dst[1] = rgba; /* g */
+         dst[2] = rgba; /* b */
+         dst[3] = rgba; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_i8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= src[3];
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_l4a4_unorm {
+   uint8_t value;
+   struct {
+      unsigned rgb:4;
+      unsigned a:4;
+   } chan;
+};
+
+static INLINE void
+util_format_l4a4_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         uint8_t a;
+         rgb = (value) & 0xf;
+         a = value >> 4;
+         dst[0] = (float)(rgb * (1.0f/0xf)); /* r */
+         dst[1] = (float)(rgb * (1.0f/0xf)); /* g */
+         dst[2] = (float)(rgb * (1.0f/0xf)); /* b */
+         dst[3] = (float)(a * (1.0f/0xf)); /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l4a4_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= ((uint8_t)(CLAMP(src[2], 0, 1) * 0xf)) & 0xf;
+         value |= ((uint8_t)(CLAMP(src[3], 0, 1) * 0xf)) << 4;
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_l4a4_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         uint8_t a;
+         rgb = (value) & 0xf;
+         a = value >> 4;
+         dst[0] = (float)(rgb * (1.0f/0xf)); /* r */
+         dst[1] = (float)(rgb * (1.0f/0xf)); /* g */
+         dst[2] = (float)(rgb * (1.0f/0xf)); /* b */
+         dst[3] = (float)(a * (1.0f/0xf)); /* a */
+}
+
+static INLINE void
+util_format_l4a4_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         uint8_t a;
+         rgb = (value) & 0xf;
+         a = value >> 4;
+         dst[0] = (uint8_t)(((uint32_t)rgb) * 0xff / 0xf); /* r */
+         dst[1] = (uint8_t)(((uint32_t)rgb) * 0xff / 0xf); /* g */
+         dst[2] = (uint8_t)(((uint32_t)rgb) * 0xff / 0xf); /* b */
+         dst[3] = (uint8_t)(((uint32_t)a) * 0xff / 0xf); /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l4a4_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= ((uint8_t)(src[2] >> 4)) & 0xf;
+         value |= ((uint8_t)(src[3] >> 4)) << 4;
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_l8a8_unorm {
+   uint16_t value;
+   struct {
+      uint8_t rgb;
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_l8a8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = (value) & 0xff;
+         a = value >> 8;
+         dst[0] = ubyte_to_float(rgb); /* r */
+         dst[1] = ubyte_to_float(rgb); /* g */
+         dst[2] = ubyte_to_float(rgb); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8a8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (float_to_ubyte(src[2])) & 0xff;
+         value |= (float_to_ubyte(src[3])) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_l8a8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = (value) & 0xff;
+         a = value >> 8;
+         dst[0] = ubyte_to_float(rgb); /* r */
+         dst[1] = ubyte_to_float(rgb); /* g */
+         dst[2] = ubyte_to_float(rgb); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_l8a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = (value) & 0xff;
+         a = value >> 8;
+         dst[0] = rgb; /* r */
+         dst[1] = rgb; /* g */
+         dst[2] = rgb; /* b */
+         dst[3] = a; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (src[2]) & 0xff;
+         value |= (src[3]) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_l16_unorm {
+   uint16_t value;
+   struct {
+      uint16_t rgb;
+   } chan;
+};
+
+static INLINE void
+util_format_l16_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = value;
+         dst[0] = (float)(rgb * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(rgb * (1.0f/0xffff)); /* g */
+         dst[2] = (float)(rgb * (1.0f/0xffff)); /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l16_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(CLAMP(src[2], 0, 1) * 0xffff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_l16_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = value;
+         dst[0] = (float)(rgb * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(rgb * (1.0f/0xffff)); /* g */
+         dst[2] = (float)(rgb * (1.0f/0xffff)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_l16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = value;
+         dst[0] = (uint8_t)(rgb >> 8); /* r */
+         dst[1] = (uint8_t)(rgb >> 8); /* g */
+         dst[2] = (uint8_t)(rgb >> 8); /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((uint32_t)src[2]) * 0xffff / 0xff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_l8_srgb {
+   uint8_t value;
+   struct {
+      uint8_t rgb;
+   } chan;
+};
+
+static INLINE void
+util_format_l8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         rgb = value;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(rgb); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(rgb); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(rgb); /* b */
+         dst[3] = 1; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= util_format_linear_float_to_srgb_8unorm(src[2]);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_l8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         rgb = value;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(rgb); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(rgb); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(rgb); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_l8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t rgb;
+         rgb = value;
+         dst[0] = util_format_srgb_to_linear_8unorm(rgb); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(rgb); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(rgb); /* b */
+         dst[3] = 255; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= util_format_linear_to_srgb_8unorm(src[2]);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_l8a8_srgb {
+   uint16_t value;
+   struct {
+      uint8_t rgb;
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_l8a8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = (value) & 0xff;
+         a = value >> 8;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(rgb); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(rgb); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(rgb); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8a8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[2])) & 0xff;
+         value |= (float_to_ubyte(src[3])) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_l8a8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = (value) & 0xff;
+         a = value >> 8;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(rgb); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(rgb); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(rgb); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_l8a8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t rgb;
+         uint16_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         rgb = (value) & 0xff;
+         a = value >> 8;
+         dst[0] = util_format_srgb_to_linear_8unorm(rgb); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(rgb); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(rgb); /* b */
+         dst[3] = a; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_l8a8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (util_format_linear_to_srgb_8unorm(src[2])) & 0xff;
+         value |= (src[3]) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8_srgb {
+   struct {
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_srgb pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_format_srgb_8unorm_to_linear_float(pixel.chan.r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(pixel.chan.g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(pixel.chan.b); /* b */
+         dst[3] = 1; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_srgb pixel;
+         pixel.chan.r = util_format_linear_float_to_srgb_8unorm(src[0]);
+         pixel.chan.g = util_format_linear_float_to_srgb_8unorm(src[1]);
+         pixel.chan.b = util_format_linear_float_to_srgb_8unorm(src[2]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r8g8b8_srgb pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_format_srgb_8unorm_to_linear_float(pixel.chan.r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(pixel.chan.g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(pixel.chan.b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_srgb pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_format_srgb_to_linear_8unorm(pixel.chan.r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(pixel.chan.g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(pixel.chan.b); /* b */
+         dst[3] = 255; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_srgb pixel;
+         pixel.chan.r = util_format_linear_to_srgb_8unorm(src[0]);
+         pixel.chan.g = util_format_linear_to_srgb_8unorm(src[1]);
+         pixel.chan.b = util_format_linear_to_srgb_8unorm(src[2]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8a8_srgb {
+   uint32_t value;
+   struct {
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8a8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[0])) & 0xff;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[1])) & 0xff) << 8;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[2])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[3])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_r8g8b8a8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = util_format_srgb_to_linear_8unorm(r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(b); /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (util_format_linear_to_srgb_8unorm(src[0])) & 0xff;
+         value |= ((util_format_linear_to_srgb_8unorm(src[1])) & 0xff) << 8;
+         value |= ((util_format_linear_to_srgb_8unorm(src[2])) & 0xff) << 16;
+         value |= (src[3]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_a8b8g8r8_srgb {
+   uint32_t value;
+   struct {
+      uint8_t a;
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_a8b8g8r8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8b8g8r8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[3])) & 0xff;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[2])) & 0xff) << 8;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[0])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_a8b8g8r8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_a8b8g8r8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = util_format_srgb_to_linear_8unorm(r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(b); /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8b8g8r8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[3]) & 0xff;
+         value |= ((util_format_linear_to_srgb_8unorm(src[2])) & 0xff) << 8;
+         value |= ((util_format_linear_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_to_srgb_8unorm(src[0])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_x8b8g8r8_srgb {
+   uint32_t value;
+   struct {
+      uint8_t x;
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_x8b8g8r8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8b8g8r8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[2])) & 0xff) << 8;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[0])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_x8b8g8r8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_x8b8g8r8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         r = value >> 24;
+         dst[0] = util_format_srgb_to_linear_8unorm(r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(b); /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8b8g8r8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((util_format_linear_to_srgb_8unorm(src[2])) & 0xff) << 8;
+         value |= ((util_format_linear_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_to_srgb_8unorm(src[0])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b8g8r8a8_srgb {
+   uint32_t value;
+   struct {
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_b8g8r8a8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8a8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[2])) & 0xff;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[1])) & 0xff) << 8;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[0])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[3])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8a8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_b8g8r8a8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = util_format_srgb_to_linear_8unorm(r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(b); /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8a8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (util_format_linear_to_srgb_8unorm(src[2])) & 0xff;
+         value |= ((util_format_linear_to_srgb_8unorm(src[1])) & 0xff) << 8;
+         value |= ((util_format_linear_to_srgb_8unorm(src[0])) & 0xff) << 16;
+         value |= (src[3]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_b8g8r8x8_srgb {
+   uint32_t value;
+   struct {
+      uint8_t b;
+      uint8_t g;
+      uint8_t r;
+      uint8_t x;
+   } chan;
+};
+
+static INLINE void
+util_format_b8g8r8x8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8x8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[2])) & 0xff;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[1])) & 0xff) << 8;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[0])) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8x8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_b8g8r8x8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t b;
+         uint32_t g;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         b = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         r = (value >> 16) & 0xff;
+         dst[0] = util_format_srgb_to_linear_8unorm(r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(b); /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_b8g8r8x8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (util_format_linear_to_srgb_8unorm(src[2])) & 0xff;
+         value |= ((util_format_linear_to_srgb_8unorm(src[1])) & 0xff) << 8;
+         value |= ((util_format_linear_to_srgb_8unorm(src[0])) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_a8r8g8b8_srgb {
+   uint32_t value;
+   struct {
+      uint8_t a;
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_a8r8g8b8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8r8g8b8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[3])) & 0xff;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[0])) & 0xff) << 8;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[2])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_a8r8g8b8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_a8r8g8b8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t a;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         a = (value) & 0xff;
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = util_format_srgb_to_linear_8unorm(r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(b); /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_a8r8g8b8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[3]) & 0xff;
+         value |= ((util_format_linear_to_srgb_8unorm(src[0])) & 0xff) << 8;
+         value |= ((util_format_linear_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_to_srgb_8unorm(src[2])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_x8r8g8b8_srgb {
+   uint32_t value;
+   struct {
+      uint8_t x;
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_x8r8g8b8_srgb_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8r8g8b8_srgb_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[0])) & 0xff) << 8;
+         value |= ((util_format_linear_float_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_float_to_srgb_8unorm(src[2])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_x8r8g8b8_srgb_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = util_format_srgb_8unorm_to_linear_float(r); /* r */
+         dst[1] = util_format_srgb_8unorm_to_linear_float(g); /* g */
+         dst[2] = util_format_srgb_8unorm_to_linear_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_x8r8g8b8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value >> 8) & 0xff;
+         g = (value >> 16) & 0xff;
+         b = value >> 24;
+         dst[0] = util_format_srgb_to_linear_8unorm(r); /* r */
+         dst[1] = util_format_srgb_to_linear_8unorm(g); /* g */
+         dst[2] = util_format_srgb_to_linear_8unorm(b); /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_x8r8g8b8_srgb_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((util_format_linear_to_srgb_8unorm(src[0])) & 0xff) << 8;
+         value |= ((util_format_linear_to_srgb_8unorm(src[1])) & 0xff) << 16;
+         value |= (util_format_linear_to_srgb_8unorm(src[2])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8sg8sb8ux8u_norm {
+   uint32_t value;
+   struct {
+      int8_t r;
+      int8_t g;
+      uint8_t b;
+      uint8_t x;
+   } chan;
+};
+
+static INLINE void
+util_format_r8sg8sb8ux8u_norm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = (value >> 16) & 0xff;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7f)); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8sg8sb8ux8u_norm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int8_t)(CLAMP(src[0], -1, 1) * 0x7f)) & 0xff) ;
+         value |= (uint32_t)((((int8_t)(CLAMP(src[1], -1, 1) * 0x7f)) & 0xff) << 8) ;
+         value |= ((float_to_ubyte(src[2])) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8sg8sb8ux8u_norm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = (value >> 16) & 0xff;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7f)); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8sg8sb8ux8u_norm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = (value >> 16) & 0xff;
+         dst[0] = (uint8_t)(((uint32_t)MAX2(r, 0)) * 0xff / 0x7f); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MAX2(g, 0)) * 0xff / 0x7f); /* g */
+         dst[2] = b; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8sg8sb8ux8u_norm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int8_t)(src[0] >> 1)) & 0xff) ;
+         value |= (uint32_t)((((int8_t)(src[1] >> 1)) & 0xff) << 8) ;
+         value |= ((src[2]) & 0xff) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r10sg10sb10sa2u_norm {
+   uint32_t value;
+   struct {
+      int r:10;
+      int g:10;
+      int b:10;
+      unsigned a:2;
+   } chan;
+};
+
+static INLINE void
+util_format_r10sg10sb10sa2u_norm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 22) ) >> 22;
+         g = ((int32_t)(value << 12) ) >> 22;
+         b = ((int32_t)(value << 2) ) >> 22;
+         a = value >> 30;
+         dst[0] = (float)(r * (1.0f/0x1ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1ff)); /* b */
+         dst[3] = (float)(a * (1.0f/0x3)); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10sg10sb10sa2u_norm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((uint32_t)(CLAMP(src[0], -1, 1) * 0x1ff)) & 0x3ff) ;
+         value |= (uint32_t)((((uint32_t)(CLAMP(src[1], -1, 1) * 0x1ff)) & 0x3ff) << 10) ;
+         value |= (uint32_t)((((uint32_t)(CLAMP(src[2], -1, 1) * 0x1ff)) & 0x3ff) << 20) ;
+         value |= ((uint32_t)(CLAMP(src[3], 0, 1) * 0x3)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r10sg10sb10sa2u_norm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 22) ) >> 22;
+         g = ((int32_t)(value << 12) ) >> 22;
+         b = ((int32_t)(value << 2) ) >> 22;
+         a = value >> 30;
+         dst[0] = (float)(r * (1.0f/0x1ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1ff)); /* b */
+         dst[3] = (float)(a * (1.0f/0x3)); /* a */
+}
+
+static INLINE void
+util_format_r10sg10sb10sa2u_norm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 22) ) >> 22;
+         g = ((int32_t)(value << 12) ) >> 22;
+         b = ((int32_t)(value << 2) ) >> 22;
+         a = value >> 30;
+         dst[0] = (uint8_t)(MAX2(r, 0) >> 1); /* r */
+         dst[1] = (uint8_t)(MAX2(g, 0) >> 1); /* g */
+         dst[2] = (uint8_t)(MAX2(b, 0) >> 1); /* b */
+         dst[3] = (uint8_t)(((uint32_t)a) * 0xff / 0x3); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10sg10sb10sa2u_norm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((uint32_t)(((uint32_t)src[0]) * 0x1ff / 0xff)) & 0x3ff) ;
+         value |= (uint32_t)((((uint32_t)(((uint32_t)src[1]) * 0x1ff / 0xff)) & 0x3ff) << 10) ;
+         value |= (uint32_t)((((uint32_t)(((uint32_t)src[2]) * 0x1ff / 0xff)) & 0x3ff) << 20) ;
+         value |= ((uint32_t)(src[3] >> 6)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r5sg5sb6u_norm {
+   uint16_t value;
+   struct {
+      int r:5;
+      int g:5;
+      unsigned b:6;
+   } chan;
+};
+
+static INLINE void
+util_format_r5sg5sb6u_norm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+         uint16_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 11) ) >> 11;
+         g = ((int16_t)(value << 6) ) >> 11;
+         b = value >> 10;
+         dst[0] = (float)(r * (1.0f/0xf)); /* r */
+         dst[1] = (float)(g * (1.0f/0xf)); /* g */
+         dst[2] = (float)(b * (1.0f/0x3f)); /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r5sg5sb6u_norm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((uint16_t)(CLAMP(src[0], -1, 1) * 0xf)) & 0x1f) ;
+         value |= (uint16_t)((((uint16_t)(CLAMP(src[1], -1, 1) * 0xf)) & 0x1f) << 5) ;
+         value |= ((uint16_t)(CLAMP(src[2], 0, 1) * 0x3f)) << 10;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r5sg5sb6u_norm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+         uint16_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 11) ) >> 11;
+         g = ((int16_t)(value << 6) ) >> 11;
+         b = value >> 10;
+         dst[0] = (float)(r * (1.0f/0xf)); /* r */
+         dst[1] = (float)(g * (1.0f/0xf)); /* g */
+         dst[2] = (float)(b * (1.0f/0x3f)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r5sg5sb6u_norm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+         uint16_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 11) ) >> 11;
+         g = ((int16_t)(value << 6) ) >> 11;
+         b = value >> 10;
+         dst[0] = (uint8_t)(((uint32_t)MAX2(r, 0)) * 0xff / 0xf); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MAX2(g, 0)) * 0xff / 0xf); /* g */
+         dst[2] = (uint8_t)(((uint32_t)b) * 0xff / 0x3f); /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r5sg5sb6u_norm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((uint16_t)(src[0] >> 4)) & 0x1f) ;
+         value |= (uint16_t)((((uint16_t)(src[1] >> 4)) & 0x1f) << 5) ;
+         value |= ((uint16_t)(src[2] >> 2)) << 10;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r10g10b10a2_uscaled {
+   uint32_t value;
+   struct {
+      unsigned r:10;
+      unsigned g:10;
+      unsigned b:10;
+      unsigned a:2;
+   } chan;
+};
+
+static INLINE void
+util_format_r10g10b10a2_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = (float)a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10a2_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)CLAMP(src[0], 0, 1023)) & 0x3ff;
+         value |= (((uint32_t)CLAMP(src[1], 0, 1023)) & 0x3ff) << 10;
+         value |= (((uint32_t)CLAMP(src[2], 0, 1023)) & 0x3ff) << 20;
+         value |= ((uint32_t)CLAMP(src[3], 0, 3)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10a2_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = (float)a; /* a */
+}
+
+static INLINE void
+util_format_r10g10b10a2_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         a = value >> 30;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MIN2(b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = (uint8_t)(((uint32_t)MIN2(a, 1)) * 0xff / 0x1); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10a2_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0x3ff;
+         value |= (((uint32_t)(((uint32_t)src[1]) * 0x1 / 0xff)) & 0x3ff) << 10;
+         value |= (((uint32_t)(((uint32_t)src[2]) * 0x1 / 0xff)) & 0x3ff) << 20;
+         value |= ((uint32_t)(((uint32_t)src[3]) * 0x1 / 0xff)) << 30;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r11g11b10_float_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_r11g11b10_float_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_r11g11b10_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+}
+
+static INLINE void
+util_format_r11g11b10_float_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_r11g11b10_float_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+}
+
+static INLINE void
+util_format_rgtc1_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_snorm_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_snorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+}
+
+static INLINE void
+util_format_rgtc1_snorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc1_snorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_unorm_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+}
+
+static INLINE void
+util_format_rgtc2_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_snorm_unpack_rgba_float(float *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_snorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride, const float *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+}
+
+static INLINE void
+util_format_rgtc2_snorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+static INLINE void
+util_format_rgtc2_snorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride, const uint8_t *src_row, unsigned src_stride, unsigned width, unsigned height)
+{
+}
+
+union util_format_r64_float {
+   uint64_t value;
+   struct {
+      double r;
+   } chan;
+};
+
+static INLINE void
+util_format_r64_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64_float pixel;
+         pixel.chan.r = (double)src[0];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r64_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r64_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(CLAMP(pixel.chan.r, 0, 1) * 0xff); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64_float pixel;
+         pixel.chan.r = (double)(src[0] * (1.0f/0xff));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r64g64_float {
+   struct {
+      double r;
+      double g;
+   } chan;
+};
+
+static INLINE void
+util_format_r64g64_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64_float pixel;
+         pixel.chan.r = (double)src[0];
+         pixel.chan.g = (double)src[1];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r64g64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r64g64_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(CLAMP(pixel.chan.r, 0, 1) * 0xff); /* r */
+         dst[1] = (uint8_t)(CLAMP(pixel.chan.g, 0, 1) * 0xff); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64_float pixel;
+         pixel.chan.r = (double)(src[0] * (1.0f/0xff));
+         pixel.chan.g = (double)(src[1] * (1.0f/0xff));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r64g64b64_float {
+   struct {
+      double r;
+      double g;
+      double b;
+   } chan;
+};
+
+static INLINE void
+util_format_r64g64b64_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 24;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64b64_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64_float pixel;
+         pixel.chan.r = (double)src[0];
+         pixel.chan.g = (double)src[1];
+         pixel.chan.b = (double)src[2];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 24;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64b64_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r64g64b64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r64g64b64_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(CLAMP(pixel.chan.r, 0, 1) * 0xff); /* r */
+         dst[1] = (uint8_t)(CLAMP(pixel.chan.g, 0, 1) * 0xff); /* g */
+         dst[2] = (uint8_t)(CLAMP(pixel.chan.b, 0, 1) * 0xff); /* b */
+         dst[3] = 255; /* a */
+         src += 24;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64b64_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64_float pixel;
+         pixel.chan.r = (double)(src[0] * (1.0f/0xff));
+         pixel.chan.g = (double)(src[1] * (1.0f/0xff));
+         pixel.chan.b = (double)(src[2] * (1.0f/0xff));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 24;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r64g64b64a64_float {
+   struct {
+      double r;
+      double g;
+      double b;
+      double a;
+   } chan;
+};
+
+static INLINE void
+util_format_r64g64b64a64_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64a64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+         src += 32;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64b64a64_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64a64_float pixel;
+         pixel.chan.r = (double)src[0];
+         pixel.chan.g = (double)src[1];
+         pixel.chan.b = (double)src[2];
+         pixel.chan.a = (double)src[3];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 32;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64b64a64_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r64g64b64a64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+}
+
+static INLINE void
+util_format_r64g64b64a64_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64a64_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(CLAMP(pixel.chan.r, 0, 1) * 0xff); /* r */
+         dst[1] = (uint8_t)(CLAMP(pixel.chan.g, 0, 1) * 0xff); /* g */
+         dst[2] = (uint8_t)(CLAMP(pixel.chan.b, 0, 1) * 0xff); /* b */
+         dst[3] = (uint8_t)(CLAMP(pixel.chan.a, 0, 1) * 0xff); /* a */
+         src += 32;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r64g64b64a64_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r64g64b64a64_float pixel;
+         pixel.chan.r = (double)(src[0] * (1.0f/0xff));
+         pixel.chan.g = (double)(src[1] * (1.0f/0xff));
+         pixel.chan.b = (double)(src[2] * (1.0f/0xff));
+         pixel.chan.a = (double)(src[3] * (1.0f/0xff));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 32;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32_float {
+   uint32_t value;
+   struct {
+      float r;
+   } chan;
+};
+
+static INLINE void
+util_format_r32_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_float pixel;
+         pixel.chan.r = src[0];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(pixel.chan.r); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_float pixel;
+         pixel.chan.r = ubyte_to_float(src[0]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32_float {
+   uint64_t value;
+   struct {
+      float r;
+      float g;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_float pixel;
+         pixel.chan.r = src[0];
+         pixel.chan.g = src[1];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(pixel.chan.r); /* r */
+         dst[1] = float_to_ubyte(pixel.chan.g); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_float pixel;
+         pixel.chan.r = ubyte_to_float(src[0]);
+         pixel.chan.g = ubyte_to_float(src[1]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32_float {
+   struct {
+      float r;
+      float g;
+      float b;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = pixel.chan.g; /* g */
+         dst[2] = pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_float pixel;
+         pixel.chan.r = src[0];
+         pixel.chan.g = src[1];
+         pixel.chan.b = src[2];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = pixel.chan.g; /* g */
+         dst[2] = pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(pixel.chan.r); /* r */
+         dst[1] = float_to_ubyte(pixel.chan.g); /* g */
+         dst[2] = float_to_ubyte(pixel.chan.b); /* b */
+         dst[3] = 255; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_float pixel;
+         pixel.chan.r = ubyte_to_float(src[0]);
+         pixel.chan.g = ubyte_to_float(src[1]);
+         pixel.chan.b = ubyte_to_float(src[2]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32a32_float {
+   struct {
+      float r;
+      float g;
+      float b;
+      float a;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32a32_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = pixel.chan.g; /* g */
+         dst[2] = pixel.chan.b; /* b */
+         dst[3] = pixel.chan.a; /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_float pixel;
+         pixel.chan.r = src[0];
+         pixel.chan.g = src[1];
+         pixel.chan.b = src[2];
+         pixel.chan.a = src[3];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32a32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = pixel.chan.g; /* g */
+         dst[2] = pixel.chan.b; /* b */
+         dst[3] = pixel.chan.a; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32a32_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(pixel.chan.r); /* r */
+         dst[1] = float_to_ubyte(pixel.chan.g); /* g */
+         dst[2] = float_to_ubyte(pixel.chan.b); /* b */
+         dst[3] = float_to_ubyte(pixel.chan.a); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_float pixel;
+         pixel.chan.r = ubyte_to_float(src[0]);
+         pixel.chan.g = ubyte_to_float(src[1]);
+         pixel.chan.b = ubyte_to_float(src[2]);
+         pixel.chan.a = ubyte_to_float(src[3]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32_unorm {
+   uint32_t value;
+   struct {
+      uint32_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r32_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = value;
+         dst[0] = (float)(r * (1.0/0xffffffff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(CLAMP(src[0], 0, 1) * (double)0xffffffff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = value;
+         dst[0] = (float)(r * (1.0/0xffffffff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = value;
+         dst[0] = (uint8_t)(r >> 24); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((uint64_t)src[0]) * 0xffffffff / 0xff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32_unorm {
+   uint64_t value;
+   struct {
+      uint32_t r;
+      uint32_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0xffffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0xffffffff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_unorm pixel;
+         pixel.chan.r = (uint32_t)(CLAMP(src[0], 0, 1) * (double)0xffffffff);
+         pixel.chan.g = (uint32_t)(CLAMP(src[1], 0, 1) * (double)0xffffffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0xffffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0xffffffff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(pixel.chan.r >> 24); /* r */
+         dst[1] = (uint8_t)(pixel.chan.g >> 24); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_unorm pixel;
+         pixel.chan.r = (uint32_t)(((uint64_t)src[0]) * 0xffffffff / 0xff);
+         pixel.chan.g = (uint32_t)(((uint64_t)src[1]) * 0xffffffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32_unorm {
+   struct {
+      uint32_t r;
+      uint32_t g;
+      uint32_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0xffffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0xffffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0xffffffff)); /* b */
+         dst[3] = 1; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_unorm pixel;
+         pixel.chan.r = (uint32_t)(CLAMP(src[0], 0, 1) * (double)0xffffffff);
+         pixel.chan.g = (uint32_t)(CLAMP(src[1], 0, 1) * (double)0xffffffff);
+         pixel.chan.b = (uint32_t)(CLAMP(src[2], 0, 1) * (double)0xffffffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0xffffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0xffffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0xffffffff)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(pixel.chan.r >> 24); /* r */
+         dst[1] = (uint8_t)(pixel.chan.g >> 24); /* g */
+         dst[2] = (uint8_t)(pixel.chan.b >> 24); /* b */
+         dst[3] = 255; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_unorm pixel;
+         pixel.chan.r = (uint32_t)(((uint64_t)src[0]) * 0xffffffff / 0xff);
+         pixel.chan.g = (uint32_t)(((uint64_t)src[1]) * 0xffffffff / 0xff);
+         pixel.chan.b = (uint32_t)(((uint64_t)src[2]) * 0xffffffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32a32_unorm {
+   struct {
+      uint32_t r;
+      uint32_t g;
+      uint32_t b;
+      uint32_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32a32_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0xffffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0xffffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0xffffffff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0/0xffffffff)); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_unorm pixel;
+         pixel.chan.r = (uint32_t)(CLAMP(src[0], 0, 1) * (double)0xffffffff);
+         pixel.chan.g = (uint32_t)(CLAMP(src[1], 0, 1) * (double)0xffffffff);
+         pixel.chan.b = (uint32_t)(CLAMP(src[2], 0, 1) * (double)0xffffffff);
+         pixel.chan.a = (uint32_t)(CLAMP(src[3], 0, 1) * (double)0xffffffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32a32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0xffffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0xffffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0xffffffff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0/0xffffffff)); /* a */
+}
+
+static INLINE void
+util_format_r32g32b32a32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(pixel.chan.r >> 24); /* r */
+         dst[1] = (uint8_t)(pixel.chan.g >> 24); /* g */
+         dst[2] = (uint8_t)(pixel.chan.b >> 24); /* b */
+         dst[3] = (uint8_t)(pixel.chan.a >> 24); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_unorm pixel;
+         pixel.chan.r = (uint32_t)(((uint64_t)src[0]) * 0xffffffff / 0xff);
+         pixel.chan.g = (uint32_t)(((uint64_t)src[1]) * 0xffffffff / 0xff);
+         pixel.chan.b = (uint32_t)(((uint64_t)src[2]) * 0xffffffff / 0xff);
+         pixel.chan.a = (uint32_t)(((uint64_t)src[3]) * 0xffffffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32_uscaled {
+   uint32_t value;
+   struct {
+      uint32_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r32_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)CLAMP(src[0], 0, 4294967295);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = value;
+         dst[0] = (uint8_t)(((uint64_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((uint64_t)src[0]) * 0x1 / 0xff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32_uscaled {
+   uint64_t value;
+   struct {
+      uint32_t r;
+      uint32_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_uscaled pixel;
+         pixel.chan.r = (uint32_t)CLAMP(src[0], 0, 4294967295);
+         pixel.chan.g = (uint32_t)CLAMP(src[1], 0, 4294967295);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint64_t)MIN2(pixel.chan.r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint64_t)MIN2(pixel.chan.g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_uscaled pixel;
+         pixel.chan.r = (uint32_t)(((uint64_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (uint32_t)(((uint64_t)src[1]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32_uscaled {
+   struct {
+      uint32_t r;
+      uint32_t g;
+      uint32_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_uscaled pixel;
+         pixel.chan.r = (uint32_t)CLAMP(src[0], 0, 4294967295);
+         pixel.chan.g = (uint32_t)CLAMP(src[1], 0, 4294967295);
+         pixel.chan.b = (uint32_t)CLAMP(src[2], 0, 4294967295);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint64_t)MIN2(pixel.chan.r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint64_t)MIN2(pixel.chan.g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint64_t)MIN2(pixel.chan.b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = 255; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_uscaled pixel;
+         pixel.chan.r = (uint32_t)(((uint64_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (uint32_t)(((uint64_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (uint32_t)(((uint64_t)src[2]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32a32_uscaled {
+   struct {
+      uint32_t r;
+      uint32_t g;
+      uint32_t b;
+      uint32_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32a32_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_uscaled pixel;
+         pixel.chan.r = (uint32_t)CLAMP(src[0], 0, 4294967295);
+         pixel.chan.g = (uint32_t)CLAMP(src[1], 0, 4294967295);
+         pixel.chan.b = (uint32_t)CLAMP(src[2], 0, 4294967295);
+         pixel.chan.a = (uint32_t)CLAMP(src[3], 0, 4294967295);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32a32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32a32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint64_t)MIN2(pixel.chan.r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint64_t)MIN2(pixel.chan.g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint64_t)MIN2(pixel.chan.b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = (uint8_t)(((uint64_t)MIN2(pixel.chan.a, 1)) * 0xff / 0x1); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_uscaled pixel;
+         pixel.chan.r = (uint32_t)(((uint64_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (uint32_t)(((uint64_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (uint32_t)(((uint64_t)src[2]) * 0x1 / 0xff);
+         pixel.chan.a = (uint32_t)(((uint64_t)src[3]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32_snorm {
+   uint32_t value;
+   struct {
+      int32_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r32_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (int32_t)(value) ;
+         dst[0] = (float)(r * (1.0/0x7fffffff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)((int32_t)(CLAMP(src[0], -1, 1) * (double)0x7fffffff)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (int32_t)(value) ;
+         dst[0] = (float)(r * (1.0/0x7fffffff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (int32_t)(value) ;
+         dst[0] = (uint8_t)(MAX2(r, 0) >> 23); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)((int32_t)(((uint64_t)src[0]) * 0x7fffffff / 0xff)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32_snorm {
+   uint64_t value;
+   struct {
+      int32_t r;
+      int32_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x7fffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x7fffffff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_snorm pixel;
+         pixel.chan.r = (int32_t)(CLAMP(src[0], -1, 1) * (double)0x7fffffff);
+         pixel.chan.g = (int32_t)(CLAMP(src[1], -1, 1) * (double)0x7fffffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x7fffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x7fffffff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(MAX2(pixel.chan.r, 0) >> 23); /* r */
+         dst[1] = (uint8_t)(MAX2(pixel.chan.g, 0) >> 23); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_snorm pixel;
+         pixel.chan.r = (int32_t)(((uint64_t)src[0]) * 0x7fffffff / 0xff);
+         pixel.chan.g = (int32_t)(((uint64_t)src[1]) * 0x7fffffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32_snorm {
+   struct {
+      int32_t r;
+      int32_t g;
+      int32_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x7fffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x7fffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x7fffffff)); /* b */
+         dst[3] = 1; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_snorm pixel;
+         pixel.chan.r = (int32_t)(CLAMP(src[0], -1, 1) * (double)0x7fffffff);
+         pixel.chan.g = (int32_t)(CLAMP(src[1], -1, 1) * (double)0x7fffffff);
+         pixel.chan.b = (int32_t)(CLAMP(src[2], -1, 1) * (double)0x7fffffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x7fffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x7fffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x7fffffff)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(MAX2(pixel.chan.r, 0) >> 23); /* r */
+         dst[1] = (uint8_t)(MAX2(pixel.chan.g, 0) >> 23); /* g */
+         dst[2] = (uint8_t)(MAX2(pixel.chan.b, 0) >> 23); /* b */
+         dst[3] = 255; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_snorm pixel;
+         pixel.chan.r = (int32_t)(((uint64_t)src[0]) * 0x7fffffff / 0xff);
+         pixel.chan.g = (int32_t)(((uint64_t)src[1]) * 0x7fffffff / 0xff);
+         pixel.chan.b = (int32_t)(((uint64_t)src[2]) * 0x7fffffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32a32_snorm {
+   struct {
+      int32_t r;
+      int32_t g;
+      int32_t b;
+      int32_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32a32_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x7fffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x7fffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x7fffffff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0/0x7fffffff)); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_snorm pixel;
+         pixel.chan.r = (int32_t)(CLAMP(src[0], -1, 1) * (double)0x7fffffff);
+         pixel.chan.g = (int32_t)(CLAMP(src[1], -1, 1) * (double)0x7fffffff);
+         pixel.chan.b = (int32_t)(CLAMP(src[2], -1, 1) * (double)0x7fffffff);
+         pixel.chan.a = (int32_t)(CLAMP(src[3], -1, 1) * (double)0x7fffffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32a32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x7fffffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x7fffffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x7fffffff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0/0x7fffffff)); /* a */
+}
+
+static INLINE void
+util_format_r32g32b32a32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(MAX2(pixel.chan.r, 0) >> 23); /* r */
+         dst[1] = (uint8_t)(MAX2(pixel.chan.g, 0) >> 23); /* g */
+         dst[2] = (uint8_t)(MAX2(pixel.chan.b, 0) >> 23); /* b */
+         dst[3] = (uint8_t)(MAX2(pixel.chan.a, 0) >> 23); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_snorm pixel;
+         pixel.chan.r = (int32_t)(((uint64_t)src[0]) * 0x7fffffff / 0xff);
+         pixel.chan.g = (int32_t)(((uint64_t)src[1]) * 0x7fffffff / 0xff);
+         pixel.chan.b = (int32_t)(((uint64_t)src[2]) * 0x7fffffff / 0xff);
+         pixel.chan.a = (int32_t)(((uint64_t)src[3]) * 0x7fffffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32_sscaled {
+   uint32_t value;
+   struct {
+      int32_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r32_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (int32_t)(value) ;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)((int32_t)CLAMP(src[0], -2147483648, 2147483647)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (int32_t)(value) ;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (int32_t)(value) ;
+         dst[0] = (uint8_t)(((uint64_t)CLAMP(r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)((int32_t)(((uint64_t)src[0]) * 0x1 / 0xff)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32_sscaled {
+   uint64_t value;
+   struct {
+      int32_t r;
+      int32_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_sscaled pixel;
+         pixel.chan.r = (int32_t)CLAMP(src[0], -2147483648, 2147483647);
+         pixel.chan.g = (int32_t)CLAMP(src[1], -2147483648, 2147483647);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_sscaled pixel;
+         pixel.chan.r = (int32_t)(((uint64_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (int32_t)(((uint64_t)src[1]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32_sscaled {
+   struct {
+      int32_t r;
+      int32_t g;
+      int32_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_sscaled pixel;
+         pixel.chan.r = (int32_t)CLAMP(src[0], -2147483648, 2147483647);
+         pixel.chan.g = (int32_t)CLAMP(src[1], -2147483648, 2147483647);
+         pixel.chan.b = (int32_t)CLAMP(src[2], -2147483648, 2147483647);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.b, 0, 1)) * 0xff / 0x1); /* b */
+         dst[3] = 255; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_sscaled pixel;
+         pixel.chan.r = (int32_t)(((uint64_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (int32_t)(((uint64_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (int32_t)(((uint64_t)src[2]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32a32_sscaled {
+   struct {
+      int32_t r;
+      int32_t g;
+      int32_t b;
+      int32_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32a32_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_sscaled pixel;
+         pixel.chan.r = (int32_t)CLAMP(src[0], -2147483648, 2147483647);
+         pixel.chan.g = (int32_t)CLAMP(src[1], -2147483648, 2147483647);
+         pixel.chan.b = (int32_t)CLAMP(src[2], -2147483648, 2147483647);
+         pixel.chan.a = (int32_t)CLAMP(src[3], -2147483648, 2147483647);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32a32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32a32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.b, 0, 1)) * 0xff / 0x1); /* b */
+         dst[3] = (uint8_t)(((uint64_t)CLAMP(pixel.chan.a, 0, 1)) * 0xff / 0x1); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_sscaled pixel;
+         pixel.chan.r = (int32_t)(((uint64_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (int32_t)(((uint64_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (int32_t)(((uint64_t)src[2]) * 0x1 / 0xff);
+         pixel.chan.a = (int32_t)(((uint64_t)src[3]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16_float {
+   uint16_t value;
+   struct {
+      uint16_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r16_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16_float pixel;
+         pixel.chan.r = util_float_to_half(src[0]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(util_half_to_float(pixel.chan.r)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16_float pixel;
+         pixel.chan.r = util_float_to_half((float)(src[0] * (1.0f/0xff)));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16_float {
+   uint32_t value;
+   struct {
+      uint16_t r;
+      uint16_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = util_half_to_float(pixel.chan.g); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16_float pixel;
+         pixel.chan.r = util_float_to_half(src[0]);
+         pixel.chan.g = util_float_to_half(src[1]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = util_half_to_float(pixel.chan.g); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(util_half_to_float(pixel.chan.r)); /* r */
+         dst[1] = float_to_ubyte(util_half_to_float(pixel.chan.g)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16_float pixel;
+         pixel.chan.r = util_float_to_half((float)(src[0] * (1.0f/0xff)));
+         pixel.chan.g = util_float_to_half((float)(src[1] * (1.0f/0xff)));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16_float {
+   struct {
+      uint16_t r;
+      uint16_t g;
+      uint16_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = util_half_to_float(pixel.chan.g); /* g */
+         dst[2] = util_half_to_float(pixel.chan.b); /* b */
+         dst[3] = 1; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_float pixel;
+         pixel.chan.r = util_float_to_half(src[0]);
+         pixel.chan.g = util_float_to_half(src[1]);
+         pixel.chan.b = util_float_to_half(src[2]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = util_half_to_float(pixel.chan.g); /* g */
+         dst[2] = util_half_to_float(pixel.chan.b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16b16_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(util_half_to_float(pixel.chan.r)); /* r */
+         dst[1] = float_to_ubyte(util_half_to_float(pixel.chan.g)); /* g */
+         dst[2] = float_to_ubyte(util_half_to_float(pixel.chan.b)); /* b */
+         dst[3] = 255; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_float pixel;
+         pixel.chan.r = util_float_to_half((float)(src[0] * (1.0f/0xff)));
+         pixel.chan.g = util_float_to_half((float)(src[1] * (1.0f/0xff)));
+         pixel.chan.b = util_float_to_half((float)(src[2] * (1.0f/0xff)));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16a16_float {
+   uint64_t value;
+   struct {
+      uint16_t r;
+      uint16_t g;
+      uint16_t b;
+      uint16_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16a16_float_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = util_half_to_float(pixel.chan.g); /* g */
+         dst[2] = util_half_to_float(pixel.chan.b); /* b */
+         dst[3] = util_half_to_float(pixel.chan.a); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_float_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_float pixel;
+         pixel.chan.r = util_float_to_half(src[0]);
+         pixel.chan.g = util_float_to_half(src[1]);
+         pixel.chan.b = util_float_to_half(src[2]);
+         pixel.chan.a = util_float_to_half(src[3]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_float_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16a16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = util_half_to_float(pixel.chan.r); /* r */
+         dst[1] = util_half_to_float(pixel.chan.g); /* g */
+         dst[2] = util_half_to_float(pixel.chan.b); /* b */
+         dst[3] = util_half_to_float(pixel.chan.a); /* a */
+}
+
+static INLINE void
+util_format_r16g16b16a16_float_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_float pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = float_to_ubyte(util_half_to_float(pixel.chan.r)); /* r */
+         dst[1] = float_to_ubyte(util_half_to_float(pixel.chan.g)); /* g */
+         dst[2] = float_to_ubyte(util_half_to_float(pixel.chan.b)); /* b */
+         dst[3] = float_to_ubyte(util_half_to_float(pixel.chan.a)); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_float_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_float pixel;
+         pixel.chan.r = util_float_to_half((float)(src[0] * (1.0f/0xff)));
+         pixel.chan.g = util_float_to_half((float)(src[1] * (1.0f/0xff)));
+         pixel.chan.b = util_float_to_half((float)(src[2] * (1.0f/0xff)));
+         pixel.chan.a = util_float_to_half((float)(src[3] * (1.0f/0xff)));
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16_unorm {
+   uint16_t value;
+   struct {
+      uint16_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r16_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = value;
+         dst[0] = (float)(r * (1.0f/0xffff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(CLAMP(src[0], 0, 1) * 0xffff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = value;
+         dst[0] = (float)(r * (1.0f/0xffff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = value;
+         dst[0] = (uint8_t)(r >> 8); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((uint32_t)src[0]) * 0xffff / 0xff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16_unorm {
+   uint32_t value;
+   struct {
+      uint16_t r;
+      uint16_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xffff;
+         g = value >> 16;
+         dst[0] = (float)(r * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(g * (1.0f/0xffff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint16_t)(CLAMP(src[0], 0, 1) * 0xffff)) & 0xffff;
+         value |= ((uint16_t)(CLAMP(src[1], 0, 1) * 0xffff)) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xffff;
+         g = value >> 16;
+         dst[0] = (float)(r * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(g * (1.0f/0xffff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xffff;
+         g = value >> 16;
+         dst[0] = (uint8_t)(r >> 8); /* r */
+         dst[1] = (uint8_t)(g >> 8); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint16_t)(((uint32_t)src[0]) * 0xffff / 0xff)) & 0xffff;
+         value |= ((uint16_t)(((uint32_t)src[1]) * 0xffff / 0xff)) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16_unorm {
+   struct {
+      uint16_t r;
+      uint16_t g;
+      uint16_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0xffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0xffff)); /* b */
+         dst[3] = 1; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_unorm pixel;
+         pixel.chan.r = (uint16_t)(CLAMP(src[0], 0, 1) * 0xffff);
+         pixel.chan.g = (uint16_t)(CLAMP(src[1], 0, 1) * 0xffff);
+         pixel.chan.b = (uint16_t)(CLAMP(src[2], 0, 1) * 0xffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0xffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0xffff)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16b16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(pixel.chan.r >> 8); /* r */
+         dst[1] = (uint8_t)(pixel.chan.g >> 8); /* g */
+         dst[2] = (uint8_t)(pixel.chan.b >> 8); /* b */
+         dst[3] = 255; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_unorm pixel;
+         pixel.chan.r = (uint16_t)(((uint32_t)src[0]) * 0xffff / 0xff);
+         pixel.chan.g = (uint16_t)(((uint32_t)src[1]) * 0xffff / 0xff);
+         pixel.chan.b = (uint16_t)(((uint32_t)src[2]) * 0xffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16a16_unorm {
+   uint64_t value;
+   struct {
+      uint16_t r;
+      uint16_t g;
+      uint16_t b;
+      uint16_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16a16_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0xffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0xffff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0f/0xffff)); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_unorm pixel;
+         pixel.chan.r = (uint16_t)(CLAMP(src[0], 0, 1) * 0xffff);
+         pixel.chan.g = (uint16_t)(CLAMP(src[1], 0, 1) * 0xffff);
+         pixel.chan.b = (uint16_t)(CLAMP(src[2], 0, 1) * 0xffff);
+         pixel.chan.a = (uint16_t)(CLAMP(src[3], 0, 1) * 0xffff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16a16_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0xffff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0xffff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0xffff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0f/0xffff)); /* a */
+}
+
+static INLINE void
+util_format_r16g16b16a16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(pixel.chan.r >> 8); /* r */
+         dst[1] = (uint8_t)(pixel.chan.g >> 8); /* g */
+         dst[2] = (uint8_t)(pixel.chan.b >> 8); /* b */
+         dst[3] = (uint8_t)(pixel.chan.a >> 8); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_unorm pixel;
+         pixel.chan.r = (uint16_t)(((uint32_t)src[0]) * 0xffff / 0xff);
+         pixel.chan.g = (uint16_t)(((uint32_t)src[1]) * 0xffff / 0xff);
+         pixel.chan.b = (uint16_t)(((uint32_t)src[2]) * 0xffff / 0xff);
+         pixel.chan.a = (uint16_t)(((uint32_t)src[3]) * 0xffff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16_uscaled {
+   uint16_t value;
+   struct {
+      uint16_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r16_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)CLAMP(src[0], 0, 65535);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = value;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16_uscaled {
+   uint32_t value;
+   struct {
+      uint16_t r;
+      uint16_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xffff;
+         g = value >> 16;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint16_t)CLAMP(src[0], 0, 65535)) & 0xffff;
+         value |= ((uint16_t)CLAMP(src[1], 0, 65535)) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xffff;
+         g = value >> 16;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xffff;
+         g = value >> 16;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint16_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0xffff;
+         value |= ((uint16_t)(((uint32_t)src[1]) * 0x1 / 0xff)) << 16;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16_uscaled {
+   struct {
+      uint16_t r;
+      uint16_t g;
+      uint16_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_uscaled pixel;
+         pixel.chan.r = (uint16_t)CLAMP(src[0], 0, 65535);
+         pixel.chan.g = (uint16_t)CLAMP(src[1], 0, 65535);
+         pixel.chan.b = (uint16_t)CLAMP(src[2], 0, 65535);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16b16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint32_t)MIN2(pixel.chan.r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(pixel.chan.g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MIN2(pixel.chan.b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = 255; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_uscaled pixel;
+         pixel.chan.r = (uint16_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (uint16_t)(((uint32_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (uint16_t)(((uint32_t)src[2]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16a16_uscaled {
+   uint64_t value;
+   struct {
+      uint16_t r;
+      uint16_t g;
+      uint16_t b;
+      uint16_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16a16_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_uscaled pixel;
+         pixel.chan.r = (uint16_t)CLAMP(src[0], 0, 65535);
+         pixel.chan.g = (uint16_t)CLAMP(src[1], 0, 65535);
+         pixel.chan.b = (uint16_t)CLAMP(src[2], 0, 65535);
+         pixel.chan.a = (uint16_t)CLAMP(src[3], 0, 65535);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16a16_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+}
+
+static INLINE void
+util_format_r16g16b16a16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint32_t)MIN2(pixel.chan.r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(pixel.chan.g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MIN2(pixel.chan.b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = (uint8_t)(((uint32_t)MIN2(pixel.chan.a, 1)) * 0xff / 0x1); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_uscaled pixel;
+         pixel.chan.r = (uint16_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (uint16_t)(((uint32_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (uint16_t)(((uint32_t)src[2]) * 0x1 / 0xff);
+         pixel.chan.a = (uint16_t)(((uint32_t)src[3]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16_snorm {
+   uint16_t value;
+   struct {
+      int16_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r16_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (int16_t)(value) ;
+         dst[0] = (float)(r * (1.0f/0x7fff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)((int16_t)(CLAMP(src[0], -1, 1) * 0x7fff)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (int16_t)(value) ;
+         dst[0] = (float)(r * (1.0f/0x7fff)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (int16_t)(value) ;
+         dst[0] = (uint8_t)(MAX2(r, 0) >> 7); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)((int16_t)(((uint32_t)src[0]) * 0x7fff / 0xff)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16_snorm {
+   uint32_t value;
+   struct {
+      int16_t r;
+      int16_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 16) ) >> 16;
+         g = ((int32_t)(value) ) >> 16;
+         dst[0] = (float)(r * (1.0f/0x7fff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7fff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int16_t)(CLAMP(src[0], -1, 1) * 0x7fff)) & 0xffff) ;
+         value |= (uint32_t)(((int16_t)(CLAMP(src[1], -1, 1) * 0x7fff)) << 16) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 16) ) >> 16;
+         g = ((int32_t)(value) ) >> 16;
+         dst[0] = (float)(r * (1.0f/0x7fff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7fff)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 16) ) >> 16;
+         g = ((int32_t)(value) ) >> 16;
+         dst[0] = (uint8_t)(MAX2(r, 0) >> 7); /* r */
+         dst[1] = (uint8_t)(MAX2(g, 0) >> 7); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int16_t)(((uint32_t)src[0]) * 0x7fff / 0xff)) & 0xffff) ;
+         value |= (uint32_t)(((int16_t)(((uint32_t)src[1]) * 0x7fff / 0xff)) << 16) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16_snorm {
+   struct {
+      int16_t r;
+      int16_t g;
+      int16_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0x7fff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0x7fff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0x7fff)); /* b */
+         dst[3] = 1; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_snorm pixel;
+         pixel.chan.r = (int16_t)(CLAMP(src[0], -1, 1) * 0x7fff);
+         pixel.chan.g = (int16_t)(CLAMP(src[1], -1, 1) * 0x7fff);
+         pixel.chan.b = (int16_t)(CLAMP(src[2], -1, 1) * 0x7fff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0x7fff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0x7fff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0x7fff)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16b16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(MAX2(pixel.chan.r, 0) >> 7); /* r */
+         dst[1] = (uint8_t)(MAX2(pixel.chan.g, 0) >> 7); /* g */
+         dst[2] = (uint8_t)(MAX2(pixel.chan.b, 0) >> 7); /* b */
+         dst[3] = 255; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_snorm pixel;
+         pixel.chan.r = (int16_t)(((uint32_t)src[0]) * 0x7fff / 0xff);
+         pixel.chan.g = (int16_t)(((uint32_t)src[1]) * 0x7fff / 0xff);
+         pixel.chan.b = (int16_t)(((uint32_t)src[2]) * 0x7fff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16a16_snorm {
+   uint64_t value;
+   struct {
+      int16_t r;
+      int16_t g;
+      int16_t b;
+      int16_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16a16_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0x7fff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0x7fff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0x7fff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0f/0x7fff)); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_snorm pixel;
+         pixel.chan.r = (int16_t)(CLAMP(src[0], -1, 1) * 0x7fff);
+         pixel.chan.g = (int16_t)(CLAMP(src[1], -1, 1) * 0x7fff);
+         pixel.chan.b = (int16_t)(CLAMP(src[2], -1, 1) * 0x7fff);
+         pixel.chan.a = (int16_t)(CLAMP(src[3], -1, 1) * 0x7fff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16a16_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0x7fff)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0x7fff)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0x7fff)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0f/0x7fff)); /* a */
+}
+
+static INLINE void
+util_format_r16g16b16a16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(MAX2(pixel.chan.r, 0) >> 7); /* r */
+         dst[1] = (uint8_t)(MAX2(pixel.chan.g, 0) >> 7); /* g */
+         dst[2] = (uint8_t)(MAX2(pixel.chan.b, 0) >> 7); /* b */
+         dst[3] = (uint8_t)(MAX2(pixel.chan.a, 0) >> 7); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_snorm pixel;
+         pixel.chan.r = (int16_t)(((uint32_t)src[0]) * 0x7fff / 0xff);
+         pixel.chan.g = (int16_t)(((uint32_t)src[1]) * 0x7fff / 0xff);
+         pixel.chan.b = (int16_t)(((uint32_t)src[2]) * 0x7fff / 0xff);
+         pixel.chan.a = (int16_t)(((uint32_t)src[3]) * 0x7fff / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16_sscaled {
+   uint16_t value;
+   struct {
+      int16_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r16_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (int16_t)(value) ;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)((int16_t)CLAMP(src[0], -32768, 32767)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (int16_t)(value) ;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (int16_t)(value) ;
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)((int16_t)(((uint32_t)src[0]) * 0x1 / 0xff)) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16_sscaled {
+   uint32_t value;
+   struct {
+      int16_t r;
+      int16_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 16) ) >> 16;
+         g = ((int32_t)(value) ) >> 16;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int16_t)CLAMP(src[0], -32768, 32767)) & 0xffff) ;
+         value |= (uint32_t)(((int16_t)CLAMP(src[1], -32768, 32767)) << 16) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 16) ) >> 16;
+         g = ((int32_t)(value) ) >> 16;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 16) ) >> 16;
+         g = ((int32_t)(value) ) >> 16;
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)CLAMP(g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int16_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0xffff) ;
+         value |= (uint32_t)(((int16_t)(((uint32_t)src[1]) * 0x1 / 0xff)) << 16) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16_sscaled {
+   struct {
+      int16_t r;
+      int16_t g;
+      int16_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_sscaled pixel;
+         pixel.chan.r = (int16_t)CLAMP(src[0], -32768, 32767);
+         pixel.chan.g = (int16_t)CLAMP(src[1], -32768, 32767);
+         pixel.chan.b = (int16_t)CLAMP(src[2], -32768, 32767);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r16g16b16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.b, 0, 1)) * 0xff / 0x1); /* b */
+         dst[3] = 255; /* a */
+         src += 6;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16_sscaled pixel;
+         pixel.chan.r = (int16_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (int16_t)(((uint32_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (int16_t)(((uint32_t)src[2]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 6;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r16g16b16a16_sscaled {
+   uint64_t value;
+   struct {
+      int16_t r;
+      int16_t g;
+      int16_t b;
+      int16_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r16g16b16a16_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_sscaled pixel;
+         pixel.chan.r = (int16_t)CLAMP(src[0], -32768, 32767);
+         pixel.chan.g = (int16_t)CLAMP(src[1], -32768, 32767);
+         pixel.chan.b = (int16_t)CLAMP(src[2], -32768, 32767);
+         pixel.chan.a = (int16_t)CLAMP(src[3], -32768, 32767);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r16g16b16a16_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = (float)pixel.chan.a; /* a */
+}
+
+static INLINE void
+util_format_r16g16b16a16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.b, 0, 1)) * 0xff / 0x1); /* b */
+         dst[3] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.a, 0, 1)) * 0xff / 0x1); /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r16g16b16a16_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r16g16b16a16_sscaled pixel;
+         pixel.chan.r = (int16_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (int16_t)(((uint32_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (int16_t)(((uint32_t)src[2]) * 0x1 / 0xff);
+         pixel.chan.a = (int16_t)(((uint32_t)src[3]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8_unorm {
+   uint8_t value;
+   struct {
+      uint8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= float_to_ubyte(src[0]);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= src[0];
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8_unorm {
+   uint16_t value;
+   struct {
+      uint8_t r;
+      uint8_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+         uint16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (value) & 0xff;
+         g = value >> 8;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (float_to_ubyte(src[0])) & 0xff;
+         value |= (float_to_ubyte(src[1])) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+         uint16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (value) & 0xff;
+         g = value >> 8;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+         uint16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (value) & 0xff;
+         g = value >> 8;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (src[0]) & 0xff;
+         value |= (src[1]) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8_unorm {
+   struct {
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = ubyte_to_float(pixel.chan.r); /* r */
+         dst[1] = ubyte_to_float(pixel.chan.g); /* g */
+         dst[2] = ubyte_to_float(pixel.chan.b); /* b */
+         dst[3] = 1; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_unorm pixel;
+         pixel.chan.r = float_to_ubyte(src[0]);
+         pixel.chan.g = float_to_ubyte(src[1]);
+         pixel.chan.b = float_to_ubyte(src[2]);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r8g8b8_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = ubyte_to_float(pixel.chan.r); /* r */
+         dst[1] = ubyte_to_float(pixel.chan.g); /* g */
+         dst[2] = ubyte_to_float(pixel.chan.b); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_unorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = pixel.chan.r; /* r */
+         dst[1] = pixel.chan.g; /* g */
+         dst[2] = pixel.chan.b; /* b */
+         dst[3] = 255; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_unorm pixel;
+         pixel.chan.r = src[0];
+         pixel.chan.g = src[1];
+         pixel.chan.b = src[2];
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8a8_unorm {
+   uint32_t value;
+   struct {
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8a8_unorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_unorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (float_to_ubyte(src[0])) & 0xff;
+         value |= ((float_to_ubyte(src[1])) & 0xff) << 8;
+         value |= ((float_to_ubyte(src[2])) & 0xff) << 16;
+         value |= (float_to_ubyte(src[3])) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_unorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = ubyte_to_float(r); /* r */
+         dst[1] = ubyte_to_float(g); /* g */
+         dst[2] = ubyte_to_float(b); /* b */
+         dst[3] = ubyte_to_float(a); /* a */
+}
+
+static INLINE void
+util_format_r8g8b8a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = r; /* r */
+         dst[1] = g; /* g */
+         dst[2] = b; /* b */
+         dst[3] = a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_unorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (src[0]) & 0xff;
+         value |= ((src[1]) & 0xff) << 8;
+         value |= ((src[2]) & 0xff) << 16;
+         value |= (src[3]) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8_uscaled {
+   uint8_t value;
+   struct {
+      uint8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r8_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)CLAMP(src[0], 0, 255);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         uint8_t r;
+         r = value;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8_uscaled {
+   uint16_t value;
+   struct {
+      uint8_t r;
+      uint8_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+         uint16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (value) & 0xff;
+         g = value >> 8;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint8_t)CLAMP(src[0], 0, 255)) & 0xff;
+         value |= ((uint8_t)CLAMP(src[1], 0, 255)) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+         uint16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (value) & 0xff;
+         g = value >> 8;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         uint16_t r;
+         uint16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = (value) & 0xff;
+         g = value >> 8;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= ((uint8_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0xff;
+         value |= ((uint8_t)(((uint32_t)src[1]) * 0x1 / 0xff)) << 8;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8_uscaled {
+   struct {
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_uscaled pixel;
+         pixel.chan.r = (uint8_t)CLAMP(src[0], 0, 255);
+         pixel.chan.g = (uint8_t)CLAMP(src[1], 0, 255);
+         pixel.chan.b = (uint8_t)CLAMP(src[2], 0, 255);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r8g8b8_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_uscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint32_t)MIN2(pixel.chan.r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(pixel.chan.g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MIN2(pixel.chan.b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = 255; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_uscaled pixel;
+         pixel.chan.r = (uint8_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (uint8_t)(((uint32_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (uint8_t)(((uint32_t)src[2]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8a8_uscaled {
+   uint32_t value;
+   struct {
+      uint8_t r;
+      uint8_t g;
+      uint8_t b;
+      uint8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8a8_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = (float)a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint8_t)CLAMP(src[0], 0, 255)) & 0xff;
+         value |= (((uint8_t)CLAMP(src[1], 0, 255)) & 0xff) << 8;
+         value |= (((uint8_t)CLAMP(src[2], 0, 255)) & 0xff) << 16;
+         value |= ((uint8_t)CLAMP(src[3], 0, 255)) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = (float)a; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8a8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+         uint32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0xff;
+         g = (value >> 8) & 0xff;
+         b = (value >> 16) & 0xff;
+         a = value >> 24;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MIN2(b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = (uint8_t)(((uint32_t)MIN2(a, 1)) * 0xff / 0x1); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint8_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0xff;
+         value |= (((uint8_t)(((uint32_t)src[1]) * 0x1 / 0xff)) & 0xff) << 8;
+         value |= (((uint8_t)(((uint32_t)src[2]) * 0x1 / 0xff)) & 0xff) << 16;
+         value |= ((uint8_t)(((uint32_t)src[3]) * 0x1 / 0xff)) << 24;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8_snorm {
+   uint8_t value;
+   struct {
+      int8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r8_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         int8_t r;
+         r = (int8_t)(value) ;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)((int8_t)(CLAMP(src[0], -1, 1) * 0x7f)) ;
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         int8_t r;
+         r = (int8_t)(value) ;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         int8_t r;
+         r = (int8_t)(value) ;
+         dst[0] = (uint8_t)(((uint32_t)MAX2(r, 0)) * 0xff / 0x7f); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)((int8_t)(src[0] >> 1)) ;
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8_snorm {
+   uint16_t value;
+   struct {
+      int8_t r;
+      int8_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 8) ) >> 8;
+         g = ((int16_t)(value) ) >> 8;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7f)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((int8_t)(CLAMP(src[0], -1, 1) * 0x7f)) & 0xff) ;
+         value |= (uint16_t)(((int8_t)(CLAMP(src[1], -1, 1) * 0x7f)) << 8) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 8) ) >> 8;
+         g = ((int16_t)(value) ) >> 8;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7f)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 8) ) >> 8;
+         g = ((int16_t)(value) ) >> 8;
+         dst[0] = (uint8_t)(((uint32_t)MAX2(r, 0)) * 0xff / 0x7f); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MAX2(g, 0)) * 0xff / 0x7f); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((int8_t)(src[0] >> 1)) & 0xff) ;
+         value |= (uint16_t)(((int8_t)(src[1] >> 1)) << 8) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8_snorm {
+   struct {
+      int8_t r;
+      int8_t g;
+      int8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0x7f)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0x7f)); /* b */
+         dst[3] = 1; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_snorm pixel;
+         pixel.chan.r = (int8_t)(CLAMP(src[0], -1, 1) * 0x7f);
+         pixel.chan.g = (int8_t)(CLAMP(src[1], -1, 1) * 0x7f);
+         pixel.chan.b = (int8_t)(CLAMP(src[2], -1, 1) * 0x7f);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r8g8b8_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0f/0x7f)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0f/0x7f)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_snorm pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint32_t)MAX2(pixel.chan.r, 0)) * 0xff / 0x7f); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MAX2(pixel.chan.g, 0)) * 0xff / 0x7f); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MAX2(pixel.chan.b, 0)) * 0xff / 0x7f); /* b */
+         dst[3] = 255; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_snorm pixel;
+         pixel.chan.r = (int8_t)(src[0] >> 1);
+         pixel.chan.g = (int8_t)(src[1] >> 1);
+         pixel.chan.b = (int8_t)(src[2] >> 1);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8a8_snorm {
+   uint32_t value;
+   struct {
+      int8_t r;
+      int8_t g;
+      int8_t b;
+      int8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8a8_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         int32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = ((int32_t)(value << 8) ) >> 24;
+         a = ((int32_t)(value) ) >> 24;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x7f)); /* b */
+         dst[3] = (float)(a * (1.0f/0x7f)); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int8_t)(CLAMP(src[0], -1, 1) * 0x7f)) & 0xff) ;
+         value |= (uint32_t)((((int8_t)(CLAMP(src[1], -1, 1) * 0x7f)) & 0xff) << 8) ;
+         value |= (uint32_t)((((int8_t)(CLAMP(src[2], -1, 1) * 0x7f)) & 0xff) << 16) ;
+         value |= (uint32_t)(((int8_t)(CLAMP(src[3], -1, 1) * 0x7f)) << 24) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         int32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = ((int32_t)(value << 8) ) >> 24;
+         a = ((int32_t)(value) ) >> 24;
+         dst[0] = (float)(r * (1.0f/0x7f)); /* r */
+         dst[1] = (float)(g * (1.0f/0x7f)); /* g */
+         dst[2] = (float)(b * (1.0f/0x7f)); /* b */
+         dst[3] = (float)(a * (1.0f/0x7f)); /* a */
+}
+
+static INLINE void
+util_format_r8g8b8a8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         int32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = ((int32_t)(value << 8) ) >> 24;
+         a = ((int32_t)(value) ) >> 24;
+         dst[0] = (uint8_t)(((uint32_t)MAX2(r, 0)) * 0xff / 0x7f); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MAX2(g, 0)) * 0xff / 0x7f); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MAX2(b, 0)) * 0xff / 0x7f); /* b */
+         dst[3] = (uint8_t)(((uint32_t)MAX2(a, 0)) * 0xff / 0x7f); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int8_t)(src[0] >> 1)) & 0xff) ;
+         value |= (uint32_t)((((int8_t)(src[1] >> 1)) & 0xff) << 8) ;
+         value |= (uint32_t)((((int8_t)(src[2] >> 1)) & 0xff) << 16) ;
+         value |= (uint32_t)(((int8_t)(src[3] >> 1)) << 24) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8_sscaled {
+   uint8_t value;
+   struct {
+      int8_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r8_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         int8_t r;
+         r = (int8_t)(value) ;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)((int8_t)CLAMP(src[0], -128, 127)) ;
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint8_t value = *(const uint8_t *)src;
+         int8_t r;
+         r = (int8_t)(value) ;
+         dst[0] = (float)r; /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = *(const uint8_t *)src;
+         int8_t r;
+         r = (int8_t)(value) ;
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 1;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint8_t value = 0;
+         value |= (uint8_t)((int8_t)(((uint32_t)src[0]) * 0x1 / 0xff)) ;
+         *(uint8_t *)dst = value;
+         src += 4;
+         dst += 1;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8_sscaled {
+   uint16_t value;
+   struct {
+      int8_t r;
+      int8_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 8) ) >> 8;
+         g = ((int16_t)(value) ) >> 8;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((int8_t)CLAMP(src[0], -128, 127)) & 0xff) ;
+         value |= (uint16_t)(((int8_t)CLAMP(src[1], -128, 127)) << 8) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 8) ) >> 8;
+         g = ((int16_t)(value) ) >> 8;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = *(const uint16_t *)src;
+         int16_t r;
+         int16_t g;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         r = ((int16_t)(value << 8) ) >> 8;
+         g = ((int16_t)(value) ) >> 8;
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)CLAMP(g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 2;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint16_t value = 0;
+         value |= (uint16_t)(((int8_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0xff) ;
+         value |= (uint16_t)(((int8_t)(((uint32_t)src[1]) * 0x1 / 0xff)) << 8) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap16(value);
+#endif
+         *(uint16_t *)dst = value;
+         src += 4;
+         dst += 2;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8_sscaled {
+   struct {
+      int8_t r;
+      int8_t g;
+      int8_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_sscaled pixel;
+         pixel.chan.r = (int8_t)CLAMP(src[0], -128, 127);
+         pixel.chan.g = (int8_t)CLAMP(src[1], -128, 127);
+         pixel.chan.b = (int8_t)CLAMP(src[2], -128, 127);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r8g8b8_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)pixel.chan.r; /* r */
+         dst[1] = (float)pixel.chan.g; /* g */
+         dst[2] = (float)pixel.chan.b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_sscaled pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)CLAMP(pixel.chan.b, 0, 1)) * 0xff / 0x1); /* b */
+         dst[3] = 255; /* a */
+         src += 3;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r8g8b8_sscaled pixel;
+         pixel.chan.r = (int8_t)(((uint32_t)src[0]) * 0x1 / 0xff);
+         pixel.chan.g = (int8_t)(((uint32_t)src[1]) * 0x1 / 0xff);
+         pixel.chan.b = (int8_t)(((uint32_t)src[2]) * 0x1 / 0xff);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 3;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r8g8b8a8_sscaled {
+   uint32_t value;
+   struct {
+      int8_t r;
+      int8_t g;
+      int8_t b;
+      int8_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r8g8b8a8_sscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         int32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = ((int32_t)(value << 8) ) >> 24;
+         a = ((int32_t)(value) ) >> 24;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = (float)a; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_sscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int8_t)CLAMP(src[0], -128, 127)) & 0xff) ;
+         value |= (uint32_t)((((int8_t)CLAMP(src[1], -128, 127)) & 0xff) << 8) ;
+         value |= (uint32_t)((((int8_t)CLAMP(src[2], -128, 127)) & 0xff) << 16) ;
+         value |= (uint32_t)(((int8_t)CLAMP(src[3], -128, 127)) << 24) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_sscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         int32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = ((int32_t)(value << 8) ) >> 24;
+         a = ((int32_t)(value) ) >> 24;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = (float)a; /* a */
+}
+
+static INLINE void
+util_format_r8g8b8a8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+         int32_t a;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 24) ) >> 24;
+         g = ((int32_t)(value << 16) ) >> 24;
+         b = ((int32_t)(value << 8) ) >> 24;
+         a = ((int32_t)(value) ) >> 24;
+         dst[0] = (uint8_t)(((uint32_t)CLAMP(r, 0, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)CLAMP(g, 0, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)CLAMP(b, 0, 1)) * 0xff / 0x1); /* b */
+         dst[3] = (uint8_t)(((uint32_t)CLAMP(a, 0, 1)) * 0xff / 0x1); /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r8g8b8a8_sscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((int8_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0xff) ;
+         value |= (uint32_t)((((int8_t)(((uint32_t)src[1]) * 0x1 / 0xff)) & 0xff) << 8) ;
+         value |= (uint32_t)((((int8_t)(((uint32_t)src[2]) * 0x1 / 0xff)) & 0xff) << 16) ;
+         value |= (uint32_t)(((int8_t)(((uint32_t)src[3]) * 0x1 / 0xff)) << 24) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32_fixed {
+   uint32_t value;
+   struct {
+      int32_t r;
+   } chan;
+};
+
+static INLINE void
+util_format_r32_fixed_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_fixed_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_fixed pixel;
+         pixel.chan.r = (int32_t)(CLAMP(src[0], -65536, 65535) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32_fixed_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)((CLAMP(pixel.chan.r, 0, 65536) * (1.0/0x10000)) * 0xff); /* r */
+         dst[1] = 0; /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32_fixed pixel;
+         pixel.chan.r = (int32_t)((float)(src[0] * (1.0f/0xff)) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32_fixed {
+   uint64_t value;
+   struct {
+      int32_t r;
+      int32_t g;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32_fixed_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x10000)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_fixed_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_fixed pixel;
+         pixel.chan.r = (int32_t)(CLAMP(src[0], -65536, 65535) * (double)0x10000);
+         pixel.chan.g = (int32_t)(CLAMP(src[1], -65536, 65535) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_fixed_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x10000)); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)((CLAMP(pixel.chan.r, 0, 65536) * (1.0/0x10000)) * 0xff); /* r */
+         dst[1] = (uint8_t)((CLAMP(pixel.chan.g, 0, 65536) * (1.0/0x10000)) * 0xff); /* g */
+         dst[2] = 0; /* b */
+         dst[3] = 255; /* a */
+         src += 8;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32_fixed pixel;
+         pixel.chan.r = (int32_t)((float)(src[0] * (1.0f/0xff)) * (double)0x10000);
+         pixel.chan.g = (int32_t)((float)(src[1] * (1.0f/0xff)) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 8;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32_fixed {
+   struct {
+      int32_t r;
+      int32_t g;
+      int32_t b;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32_fixed_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x10000)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x10000)); /* b */
+         dst[3] = 1; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_fixed_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_fixed pixel;
+         pixel.chan.r = (int32_t)(CLAMP(src[0], -65536, 65535) * (double)0x10000);
+         pixel.chan.g = (int32_t)(CLAMP(src[1], -65536, 65535) * (double)0x10000);
+         pixel.chan.b = (int32_t)(CLAMP(src[2], -65536, 65535) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_fixed_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x10000)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x10000)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r32g32b32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)((CLAMP(pixel.chan.r, 0, 65536) * (1.0/0x10000)) * 0xff); /* r */
+         dst[1] = (uint8_t)((CLAMP(pixel.chan.g, 0, 65536) * (1.0/0x10000)) * 0xff); /* g */
+         dst[2] = (uint8_t)((CLAMP(pixel.chan.b, 0, 65536) * (1.0/0x10000)) * 0xff); /* b */
+         dst[3] = 255; /* a */
+         src += 12;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32_fixed pixel;
+         pixel.chan.r = (int32_t)((float)(src[0] * (1.0f/0xff)) * (double)0x10000);
+         pixel.chan.g = (int32_t)((float)(src[1] * (1.0f/0xff)) * (double)0x10000);
+         pixel.chan.b = (int32_t)((float)(src[2] * (1.0f/0xff)) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 12;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r32g32b32a32_fixed {
+   struct {
+      int32_t r;
+      int32_t g;
+      int32_t b;
+      int32_t a;
+   } chan;
+};
+
+static INLINE void
+util_format_r32g32b32a32_fixed_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x10000)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x10000)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0/0x10000)); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_fixed_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_fixed pixel;
+         pixel.chan.r = (int32_t)(CLAMP(src[0], -65536, 65535) * (double)0x10000);
+         pixel.chan.g = (int32_t)(CLAMP(src[1], -65536, 65535) * (double)0x10000);
+         pixel.chan.b = (int32_t)(CLAMP(src[2], -65536, 65535) * (double)0x10000);
+         pixel.chan.a = (int32_t)(CLAMP(src[3], -65536, 65535) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_fixed_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         union util_format_r32g32b32a32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (float)(pixel.chan.r * (1.0/0x10000)); /* r */
+         dst[1] = (float)(pixel.chan.g * (1.0/0x10000)); /* g */
+         dst[2] = (float)(pixel.chan.b * (1.0/0x10000)); /* b */
+         dst[3] = (float)(pixel.chan.a * (1.0/0x10000)); /* a */
+}
+
+static INLINE void
+util_format_r32g32b32a32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_fixed pixel;
+         memcpy(&pixel, src, sizeof pixel);
+         dst[0] = (uint8_t)((CLAMP(pixel.chan.r, 0, 65536) * (1.0/0x10000)) * 0xff); /* r */
+         dst[1] = (uint8_t)((CLAMP(pixel.chan.g, 0, 65536) * (1.0/0x10000)) * 0xff); /* g */
+         dst[2] = (uint8_t)((CLAMP(pixel.chan.b, 0, 65536) * (1.0/0x10000)) * 0xff); /* b */
+         dst[3] = (uint8_t)((CLAMP(pixel.chan.a, 0, 65536) * (1.0/0x10000)) * 0xff); /* a */
+         src += 16;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r32g32b32a32_fixed_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         union util_format_r32g32b32a32_fixed pixel;
+         pixel.chan.r = (int32_t)((float)(src[0] * (1.0f/0xff)) * (double)0x10000);
+         pixel.chan.g = (int32_t)((float)(src[1] * (1.0f/0xff)) * (double)0x10000);
+         pixel.chan.b = (int32_t)((float)(src[2] * (1.0f/0xff)) * (double)0x10000);
+         pixel.chan.a = (int32_t)((float)(src[3] * (1.0f/0xff)) * (double)0x10000);
+         memcpy(dst, &pixel, sizeof pixel);
+         src += 4;
+         dst += 16;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r10g10b10x2_uscaled {
+   uint32_t value;
+   struct {
+      unsigned r:10;
+      unsigned g:10;
+      unsigned b:10;
+      unsigned x:2;
+   } chan;
+};
+
+static INLINE void
+util_format_r10g10b10x2_uscaled_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10x2_uscaled_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)CLAMP(src[0], 0, 1023)) & 0x3ff;
+         value |= (((uint32_t)CLAMP(src[1], 0, 1023)) & 0x3ff) << 10;
+         value |= (((uint32_t)CLAMP(src[2], 0, 1023)) & 0x3ff) << 20;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10x2_uscaled_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         dst[0] = (float)r; /* r */
+         dst[1] = (float)g; /* g */
+         dst[2] = (float)b; /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r10g10b10x2_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         uint32_t r;
+         uint32_t g;
+         uint32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = (value) & 0x3ff;
+         g = (value >> 10) & 0x3ff;
+         b = (value >> 20) & 0x3ff;
+         dst[0] = (uint8_t)(((uint32_t)MIN2(r, 1)) * 0xff / 0x1); /* r */
+         dst[1] = (uint8_t)(((uint32_t)MIN2(g, 1)) * 0xff / 0x1); /* g */
+         dst[2] = (uint8_t)(((uint32_t)MIN2(b, 1)) * 0xff / 0x1); /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10x2_uscaled_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= ((uint32_t)(((uint32_t)src[0]) * 0x1 / 0xff)) & 0x3ff;
+         value |= (((uint32_t)(((uint32_t)src[1]) * 0x1 / 0xff)) & 0x3ff) << 10;
+         value |= (((uint32_t)(((uint32_t)src[2]) * 0x1 / 0xff)) & 0x3ff) << 20;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+union util_format_r10g10b10x2_snorm {
+   uint32_t value;
+   struct {
+      int r:10;
+      int g:10;
+      int b:10;
+      unsigned x:2;
+   } chan;
+};
+
+static INLINE void
+util_format_r10g10b10x2_snorm_unpack_rgba_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 += 1) {
+      float *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 22) ) >> 22;
+         g = ((int32_t)(value << 12) ) >> 22;
+         b = ((int32_t)(value << 2) ) >> 22;
+         dst[0] = (float)(r * (1.0f/0x1ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1ff)); /* b */
+         dst[3] = 1; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10x2_snorm_pack_rgba_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 += 1) {
+      const float *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((uint32_t)(CLAMP(src[0], -1, 1) * 0x1ff)) & 0x3ff) ;
+         value |= (uint32_t)((((uint32_t)(CLAMP(src[1], -1, 1) * 0x1ff)) & 0x3ff) << 10) ;
+         value |= (uint32_t)((((uint32_t)(CLAMP(src[2], -1, 1) * 0x1ff)) & 0x3ff) << 20) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10x2_snorm_fetch_rgba_float(float *dst, const uint8_t *src, unsigned i, unsigned j)
+{
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 22) ) >> 22;
+         g = ((int32_t)(value << 12) ) >> 22;
+         b = ((int32_t)(value << 2) ) >> 22;
+         dst[0] = (float)(r * (1.0f/0x1ff)); /* r */
+         dst[1] = (float)(g * (1.0f/0x1ff)); /* g */
+         dst[2] = (float)(b * (1.0f/0x1ff)); /* b */
+         dst[3] = 1; /* a */
+}
+
+static INLINE void
+util_format_r10g10b10x2_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      uint8_t *dst = dst_row;
+      const uint8_t *src = src_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = *(const uint32_t *)src;
+         int32_t r;
+         int32_t g;
+         int32_t b;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         r = ((int32_t)(value << 22) ) >> 22;
+         g = ((int32_t)(value << 12) ) >> 22;
+         b = ((int32_t)(value << 2) ) >> 22;
+         dst[0] = (uint8_t)(MAX2(r, 0) >> 1); /* r */
+         dst[1] = (uint8_t)(MAX2(g, 0) >> 1); /* g */
+         dst[2] = (uint8_t)(MAX2(b, 0) >> 1); /* b */
+         dst[3] = 255; /* a */
+         src += 4;
+         dst += 4;
+      }
+      src_row += src_stride;
+      dst_row += dst_stride/sizeof(*dst_row);
+   }
+}
+
+static INLINE void
+util_format_r10g10b10x2_snorm_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;
+   for(y = 0; y < height; y += 1) {
+      const uint8_t *src = src_row;
+      uint8_t *dst = dst_row;
+      for(x = 0; x < width; x += 1) {
+         uint32_t value = 0;
+         value |= (uint32_t)(((uint32_t)(((uint32_t)src[0]) * 0x1ff / 0xff)) & 0x3ff) ;
+         value |= (uint32_t)((((uint32_t)(((uint32_t)src[1]) * 0x1ff / 0xff)) & 0x3ff) << 10) ;
+         value |= (uint32_t)((((uint32_t)(((uint32_t)src[2]) * 0x1ff / 0xff)) & 0x3ff) << 20) ;
+#ifdef PIPE_ARCH_BIG_ENDIAN
+         value = util_bswap32(value);
+#endif
+         *(uint32_t *)dst = value;
+         src += 4;
+         dst += 4;
+      }
+      dst_row += dst_stride;
+      src_row += src_stride/sizeof(*src_row);
+   }
+}
+
+const struct util_format_description
+util_format_none_description = {
+   PIPE_FORMAT_NONE,
+   "PIPE_FORMAT_NONE",
+   "none",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_none_unpack_rgba_8unorm,
+   &util_format_none_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_none_unpack_rgba_float,
+   &util_format_none_pack_rgba_float,
+   &util_format_none_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b8g8r8a8_unorm_description = {
+   PIPE_FORMAT_B8G8R8A8_UNORM,
+   "PIPE_FORMAT_B8G8R8A8_UNORM",
+   "b8g8r8a8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b8g8r8a8_unorm_unpack_rgba_8unorm,
+   &util_format_b8g8r8a8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b8g8r8a8_unorm_unpack_rgba_float,
+   &util_format_b8g8r8a8_unorm_pack_rgba_float,
+   &util_format_b8g8r8a8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b8g8r8x8_unorm_description = {
+   PIPE_FORMAT_B8G8R8X8_UNORM,
+   "PIPE_FORMAT_B8G8R8X8_UNORM",
+   "b8g8r8x8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = r */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b8g8r8x8_unorm_unpack_rgba_8unorm,
+   &util_format_b8g8r8x8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b8g8r8x8_unorm_unpack_rgba_float,
+   &util_format_b8g8r8x8_unorm_pack_rgba_float,
+   &util_format_b8g8r8x8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_a8r8g8b8_unorm_description = {
+   PIPE_FORMAT_A8R8G8B8_UNORM,
+   "PIPE_FORMAT_A8R8G8B8_UNORM",
+   "a8r8g8b8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = a */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = b */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Y,	/* r */
+      UTIL_FORMAT_SWIZZLE_Z,	/* g */
+      UTIL_FORMAT_SWIZZLE_W,	/* b */
+      UTIL_FORMAT_SWIZZLE_X	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_a8r8g8b8_unorm_unpack_rgba_8unorm,
+   &util_format_a8r8g8b8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_a8r8g8b8_unorm_unpack_rgba_float,
+   &util_format_a8r8g8b8_unorm_pack_rgba_float,
+   &util_format_a8r8g8b8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_x8r8g8b8_unorm_description = {
+   PIPE_FORMAT_X8R8G8B8_UNORM,
+   "PIPE_FORMAT_X8R8G8B8_UNORM",
+   "x8r8g8b8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8},	/* x = x */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = b */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Y,	/* r */
+      UTIL_FORMAT_SWIZZLE_Z,	/* g */
+      UTIL_FORMAT_SWIZZLE_W,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_x8r8g8b8_unorm_unpack_rgba_8unorm,
+   &util_format_x8r8g8b8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_x8r8g8b8_unorm_unpack_rgba_float,
+   &util_format_x8r8g8b8_unorm_pack_rgba_float,
+   &util_format_x8r8g8b8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_a8b8g8r8_unorm_description = {
+   PIPE_FORMAT_A8B8G8R8_UNORM,
+   "PIPE_FORMAT_A8B8G8R8_UNORM",
+   "a8b8g8r8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = a */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = r */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_W,	/* r */
+      UTIL_FORMAT_SWIZZLE_Z,	/* g */
+      UTIL_FORMAT_SWIZZLE_Y,	/* b */
+      UTIL_FORMAT_SWIZZLE_X	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_a8b8g8r8_unorm_unpack_rgba_8unorm,
+   &util_format_a8b8g8r8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_a8b8g8r8_unorm_unpack_rgba_float,
+   &util_format_a8b8g8r8_unorm_pack_rgba_float,
+   &util_format_a8b8g8r8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_x8b8g8r8_unorm_description = {
+   PIPE_FORMAT_X8B8G8R8_UNORM,
+   "PIPE_FORMAT_X8B8G8R8_UNORM",
+   "x8b8g8r8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8},	/* x = x */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = r */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_W,	/* r */
+      UTIL_FORMAT_SWIZZLE_Z,	/* g */
+      UTIL_FORMAT_SWIZZLE_Y,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_x8b8g8r8_unorm_unpack_rgba_8unorm,
+   &util_format_x8b8g8r8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_x8b8g8r8_unorm_unpack_rgba_float,
+   &util_format_x8b8g8r8_unorm_pack_rgba_float,
+   &util_format_x8b8g8r8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8x8_unorm_description = {
+   PIPE_FORMAT_R8G8B8X8_UNORM,
+   "PIPE_FORMAT_R8G8B8X8_UNORM",
+   "r8g8b8x8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = b */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8x8_unorm_unpack_rgba_8unorm,
+   &util_format_r8g8b8x8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8x8_unorm_unpack_rgba_float,
+   &util_format_r8g8b8x8_unorm_pack_rgba_float,
+   &util_format_r8g8b8x8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b5g5r5x1_unorm_description = {
+   PIPE_FORMAT_B5G5R5X1_UNORM,
+   "PIPE_FORMAT_B5G5R5X1_UNORM",
+   "b5g5r5x1_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* z = r */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 1}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b5g5r5x1_unorm_unpack_rgba_8unorm,
+   &util_format_b5g5r5x1_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b5g5r5x1_unorm_unpack_rgba_float,
+   &util_format_b5g5r5x1_unorm_pack_rgba_float,
+   &util_format_b5g5r5x1_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b5g5r5a1_unorm_description = {
+   PIPE_FORMAT_B5G5R5A1_UNORM,
+   "PIPE_FORMAT_B5G5R5A1_UNORM",
+   "b5g5r5a1_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* z = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 1}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b5g5r5a1_unorm_unpack_rgba_8unorm,
+   &util_format_b5g5r5a1_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b5g5r5a1_unorm_unpack_rgba_float,
+   &util_format_b5g5r5a1_unorm_pack_rgba_float,
+   &util_format_b5g5r5a1_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b4g4r4a4_unorm_description = {
+   PIPE_FORMAT_B4G4R4A4_UNORM,
+   "PIPE_FORMAT_B4G4R4A4_UNORM",
+   "b4g4r4a4_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* z = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b4g4r4a4_unorm_unpack_rgba_8unorm,
+   &util_format_b4g4r4a4_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b4g4r4a4_unorm_unpack_rgba_float,
+   &util_format_b4g4r4a4_unorm_pack_rgba_float,
+   &util_format_b4g4r4a4_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b4g4r4x4_unorm_description = {
+   PIPE_FORMAT_B4G4R4X4_UNORM,
+   "PIPE_FORMAT_B4G4R4X4_UNORM",
+   "b4g4r4x4_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* z = r */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 4}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b4g4r4x4_unorm_unpack_rgba_8unorm,
+   &util_format_b4g4r4x4_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b4g4r4x4_unorm_unpack_rgba_float,
+   &util_format_b4g4r4x4_unorm_pack_rgba_float,
+   &util_format_b4g4r4x4_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b5g6r5_unorm_description = {
+   PIPE_FORMAT_B5G6R5_UNORM,
+   "PIPE_FORMAT_B5G6R5_UNORM",
+   "b5g6r5_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 6},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 5},	/* z = r */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b5g6r5_unorm_unpack_rgba_8unorm,
+   &util_format_b5g6r5_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b5g6r5_unorm_unpack_rgba_float,
+   &util_format_b5g6r5_unorm_pack_rgba_float,
+   &util_format_b5g6r5_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r10g10b10a2_unorm_description = {
+   PIPE_FORMAT_R10G10B10A2_UNORM,
+   "PIPE_FORMAT_R10G10B10A2_UNORM",
+   "r10g10b10a2_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 10},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 10},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 10},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 2}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r10g10b10a2_unorm_unpack_rgba_8unorm,
+   &util_format_r10g10b10a2_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r10g10b10a2_unorm_unpack_rgba_float,
+   &util_format_r10g10b10a2_unorm_pack_rgba_float,
+   &util_format_r10g10b10a2_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b10g10r10a2_unorm_description = {
+   PIPE_FORMAT_B10G10R10A2_UNORM,
+   "PIPE_FORMAT_B10G10R10A2_UNORM",
+   "b10g10r10a2_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 10},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 10},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 10},	/* z = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 2}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_b10g10r10a2_unorm_unpack_rgba_8unorm,
+   &util_format_b10g10r10a2_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b10g10r10a2_unorm_unpack_rgba_float,
+   &util_format_b10g10r10a2_unorm_pack_rgba_float,
+   &util_format_b10g10r10a2_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_l8_unorm_description = {
+   PIPE_FORMAT_L8_UNORM,
+   "PIPE_FORMAT_L8_UNORM",
+   "l8_unorm",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = rgb */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_X,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_l8_unorm_unpack_rgba_8unorm,
+   &util_format_l8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_l8_unorm_unpack_rgba_float,
+   &util_format_l8_unorm_pack_rgba_float,
+   &util_format_l8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_a8_unorm_description = {
+   PIPE_FORMAT_A8_UNORM,
+   "PIPE_FORMAT_A8_UNORM",
+   "a8_unorm",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = a */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_0,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_X	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_a8_unorm_unpack_rgba_8unorm,
+   &util_format_a8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_a8_unorm_unpack_rgba_float,
+   &util_format_a8_unorm_pack_rgba_float,
+   &util_format_a8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_i8_unorm_description = {
+   PIPE_FORMAT_I8_UNORM,
+   "PIPE_FORMAT_I8_UNORM",
+   "i8_unorm",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = rgba */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_X,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_X	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_i8_unorm_unpack_rgba_8unorm,
+   &util_format_i8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_i8_unorm_unpack_rgba_float,
+   &util_format_i8_unorm_pack_rgba_float,
+   &util_format_i8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_l4a4_unorm_description = {
+   PIPE_FORMAT_L4A4_UNORM,
+   "PIPE_FORMAT_L4A4_UNORM",
+   "l4a4_unorm",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* x = rgb */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 4},	/* y = a */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_X,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_Y	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_l4a4_unorm_unpack_rgba_8unorm,
+   &util_format_l4a4_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_l4a4_unorm_unpack_rgba_float,
+   &util_format_l4a4_unorm_pack_rgba_float,
+   &util_format_l4a4_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_l8a8_unorm_description = {
+   PIPE_FORMAT_L8A8_UNORM,
+   "PIPE_FORMAT_L8A8_UNORM",
+   "l8a8_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = rgb */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = a */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_X,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_Y	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_l8a8_unorm_unpack_rgba_8unorm,
+   &util_format_l8a8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_l8a8_unorm_unpack_rgba_float,
+   &util_format_l8a8_unorm_pack_rgba_float,
+   &util_format_l8a8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_l16_unorm_description = {
+   PIPE_FORMAT_L16_UNORM,
+   "PIPE_FORMAT_L16_UNORM",
+   "l16_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* x = rgb */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_X,	/* g */
+      UTIL_FORMAT_SWIZZLE_X,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_l16_unorm_unpack_rgba_8unorm,
+   &util_format_l16_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_l16_unorm_unpack_rgba_float,
+   &util_format_l16_unorm_pack_rgba_float,
+   &util_format_l16_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_l8_srgb_description = {
+   PIPE_FORMAT_L8_SRGB,
+   "PIPE_FORMAT_L8_SRGB",
+   "l8_srgb",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = rgb */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_X,	/* sg */
+      UTIL_FORMAT_SWIZZLE_X,	/* sb */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_l8_srgb_unpack_rgba_8unorm,
+   &util_format_l8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_l8_srgb_unpack_rgba_float,
+   &util_format_l8_srgb_pack_rgba_float,
+   &util_format_l8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_l8a8_srgb_description = {
+   PIPE_FORMAT_L8A8_SRGB,
+   "PIPE_FORMAT_L8A8_SRGB",
+   "l8a8_srgb",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = rgb */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = a */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_X,	/* sg */
+      UTIL_FORMAT_SWIZZLE_X,	/* sb */
+      UTIL_FORMAT_SWIZZLE_Y	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_l8a8_srgb_unpack_rgba_8unorm,
+   &util_format_l8a8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_l8a8_srgb_unpack_rgba_float,
+   &util_format_l8a8_srgb_pack_rgba_float,
+   &util_format_l8a8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8_srgb_description = {
+   PIPE_FORMAT_R8G8B8_SRGB,
+   "PIPE_FORMAT_R8G8B8_SRGB",
+   "r8g8b8_srgb",
+   {1, 1, 24},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sb */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_r8g8b8_srgb_unpack_rgba_8unorm,
+   &util_format_r8g8b8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8_srgb_unpack_rgba_float,
+   &util_format_r8g8b8_srgb_pack_rgba_float,
+   &util_format_r8g8b8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8a8_srgb_description = {
+   PIPE_FORMAT_R8G8B8A8_SRGB,
+   "PIPE_FORMAT_R8G8B8A8_SRGB",
+   "r8g8b8a8_srgb",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sb */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_r8g8b8a8_srgb_unpack_rgba_8unorm,
+   &util_format_r8g8b8a8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8a8_srgb_unpack_rgba_float,
+   &util_format_r8g8b8a8_srgb_pack_rgba_float,
+   &util_format_r8g8b8a8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_a8b8g8r8_srgb_description = {
+   PIPE_FORMAT_A8B8G8R8_SRGB,
+   "PIPE_FORMAT_A8B8G8R8_SRGB",
+   "a8b8g8r8_srgb",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = a */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = r */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_W,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sb */
+      UTIL_FORMAT_SWIZZLE_X	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_a8b8g8r8_srgb_unpack_rgba_8unorm,
+   &util_format_a8b8g8r8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_a8b8g8r8_srgb_unpack_rgba_float,
+   &util_format_a8b8g8r8_srgb_pack_rgba_float,
+   &util_format_a8b8g8r8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_x8b8g8r8_srgb_description = {
+   PIPE_FORMAT_X8B8G8R8_SRGB,
+   "PIPE_FORMAT_X8B8G8R8_SRGB",
+   "x8b8g8r8_srgb",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8},	/* x = x */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = r */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_W,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sb */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_x8b8g8r8_srgb_unpack_rgba_8unorm,
+   &util_format_x8b8g8r8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_x8b8g8r8_srgb_unpack_rgba_float,
+   &util_format_x8b8g8r8_srgb_pack_rgba_float,
+   &util_format_x8b8g8r8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b8g8r8a8_srgb_description = {
+   PIPE_FORMAT_B8G8R8A8_SRGB,
+   "PIPE_FORMAT_B8G8R8A8_SRGB",
+   "b8g8r8a8_srgb",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_X,	/* sb */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_b8g8r8a8_srgb_unpack_rgba_8unorm,
+   &util_format_b8g8r8a8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b8g8r8a8_srgb_unpack_rgba_float,
+   &util_format_b8g8r8a8_srgb_pack_rgba_float,
+   &util_format_b8g8r8a8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_b8g8r8x8_srgb_description = {
+   PIPE_FORMAT_B8G8R8X8_SRGB,
+   "PIPE_FORMAT_B8G8R8X8_SRGB",
+   "b8g8r8x8_srgb",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = r */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Z,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_X,	/* sb */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_b8g8r8x8_srgb_unpack_rgba_8unorm,
+   &util_format_b8g8r8x8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_b8g8r8x8_srgb_unpack_rgba_float,
+   &util_format_b8g8r8x8_srgb_pack_rgba_float,
+   &util_format_b8g8r8x8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_a8r8g8b8_srgb_description = {
+   PIPE_FORMAT_A8R8G8B8_SRGB,
+   "PIPE_FORMAT_A8R8G8B8_SRGB",
+   "a8r8g8b8_srgb",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = a */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = b */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Y,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sg */
+      UTIL_FORMAT_SWIZZLE_W,	/* sb */
+      UTIL_FORMAT_SWIZZLE_X	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_a8r8g8b8_srgb_unpack_rgba_8unorm,
+   &util_format_a8r8g8b8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_a8r8g8b8_srgb_unpack_rgba_float,
+   &util_format_a8r8g8b8_srgb_pack_rgba_float,
+   &util_format_a8r8g8b8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_x8r8g8b8_srgb_description = {
+   PIPE_FORMAT_X8R8G8B8_SRGB,
+   "PIPE_FORMAT_X8R8G8B8_SRGB",
+   "x8r8g8b8_srgb",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8},	/* x = x */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = b */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Y,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sg */
+      UTIL_FORMAT_SWIZZLE_W,	/* sb */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_x8r8g8b8_srgb_unpack_rgba_8unorm,
+   &util_format_x8r8g8b8_srgb_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_x8r8g8b8_srgb_unpack_rgba_float,
+   &util_format_x8r8g8b8_srgb_pack_rgba_float,
+   &util_format_x8r8g8b8_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8sg8sb8ux8u_norm_description = {
+   PIPE_FORMAT_R8SG8SB8UX8U_NORM,
+   "PIPE_FORMAT_R8SG8SB8UX8U_NORM",
+   "r8sg8sb8ux8u_norm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   TRUE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = b */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8sg8sb8ux8u_norm_unpack_rgba_8unorm,
+   &util_format_r8sg8sb8ux8u_norm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8sg8sb8ux8u_norm_unpack_rgba_float,
+   &util_format_r8sg8sb8ux8u_norm_pack_rgba_float,
+   &util_format_r8sg8sb8ux8u_norm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r10sg10sb10sa2u_norm_description = {
+   PIPE_FORMAT_R10SG10SB10SA2U_NORM,
+   "PIPE_FORMAT_R10SG10SB10SA2U_NORM",
+   "r10sg10sb10sa2u_norm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   TRUE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 10},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 10},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 10},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 2}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r10sg10sb10sa2u_norm_unpack_rgba_8unorm,
+   &util_format_r10sg10sb10sa2u_norm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r10sg10sb10sa2u_norm_unpack_rgba_float,
+   &util_format_r10sg10sb10sa2u_norm_pack_rgba_float,
+   &util_format_r10sg10sb10sa2u_norm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r5sg5sb6u_norm_description = {
+   PIPE_FORMAT_R5SG5SB6U_NORM,
+   "PIPE_FORMAT_R5SG5SB6U_NORM",
+   "r5sg5sb6u_norm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   TRUE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 5},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 5},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 6},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r5sg5sb6u_norm_unpack_rgba_8unorm,
+   &util_format_r5sg5sb6u_norm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r5sg5sb6u_norm_unpack_rgba_float,
+   &util_format_r5sg5sb6u_norm_pack_rgba_float,
+   &util_format_r5sg5sb6u_norm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_s8_uscaled_description = {
+   PIPE_FORMAT_S8_USCALED,
+   "PIPE_FORMAT_S8_USCALED",
+   "s8_uscaled",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = s */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_NONE,	/* z */
+      UTIL_FORMAT_SWIZZLE_X,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   &util_format_s8_uscaled_unpack_s_8uscaled,
+   &util_format_s8_uscaled_pack_s_8uscaled
+};
+
+const struct util_format_description
+util_format_z16_unorm_description = {
+   PIPE_FORMAT_Z16_UNORM,
+   "PIPE_FORMAT_Z16_UNORM",
+   "z16_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* x = z */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* z */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_z16_unorm_unpack_z_32unorm,
+   &util_format_z16_unorm_pack_z_32unorm,
+   &util_format_z16_unorm_unpack_z_float,
+   &util_format_z16_unorm_pack_z_float,
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_z32_unorm_description = {
+   PIPE_FORMAT_Z32_UNORM,
+   "PIPE_FORMAT_Z32_UNORM",
+   "z32_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* x = z */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* z */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_z32_unorm_unpack_z_32unorm,
+   &util_format_z32_unorm_pack_z_32unorm,
+   &util_format_z32_unorm_unpack_z_float,
+   &util_format_z32_unorm_pack_z_float,
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_z32_float_description = {
+   PIPE_FORMAT_Z32_FLOAT,
+   "PIPE_FORMAT_Z32_FLOAT",
+   "z32_float",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* x = z */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* z */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_z32_float_unpack_z_32unorm,
+   &util_format_z32_float_pack_z_32unorm,
+   &util_format_z32_float_unpack_z_float,
+   &util_format_z32_float_pack_z_float,
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_z24_unorm_s8_uscaled_description = {
+   PIPE_FORMAT_Z24_UNORM_S8_USCALED,
+   "PIPE_FORMAT_Z24_UNORM_S8_USCALED",
+   "z24_unorm_s8_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   TRUE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 24},	/* x = z */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* y = s */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* z */
+      UTIL_FORMAT_SWIZZLE_Y,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_z24_unorm_s8_uscaled_unpack_z_32unorm,
+   &util_format_z24_unorm_s8_uscaled_pack_z_32unorm,
+   &util_format_z24_unorm_s8_uscaled_unpack_z_float,
+   &util_format_z24_unorm_s8_uscaled_pack_z_float,
+   &util_format_z24_unorm_s8_uscaled_unpack_s_8uscaled,
+   &util_format_z24_unorm_s8_uscaled_pack_s_8uscaled
+};
+
+const struct util_format_description
+util_format_s8_uscaled_z24_unorm_description = {
+   PIPE_FORMAT_S8_USCALED_Z24_UNORM,
+   "PIPE_FORMAT_S8_USCALED_Z24_UNORM",
+   "s8_uscaled_z24_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   TRUE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = s */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 24},	/* y = z */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Y,	/* z */
+      UTIL_FORMAT_SWIZZLE_X,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_s8_uscaled_z24_unorm_unpack_z_32unorm,
+   &util_format_s8_uscaled_z24_unorm_pack_z_32unorm,
+   &util_format_s8_uscaled_z24_unorm_unpack_z_float,
+   &util_format_s8_uscaled_z24_unorm_pack_z_float,
+   &util_format_s8_uscaled_z24_unorm_unpack_s_8uscaled,
+   &util_format_s8_uscaled_z24_unorm_pack_s_8uscaled
+};
+
+const struct util_format_description
+util_format_x24s8_uscaled_description = {
+   PIPE_FORMAT_X24S8_USCALED,
+   "PIPE_FORMAT_X24S8_USCALED",
+   "x24s8_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 24},	/* x = x */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* y = s */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_NONE,	/* z */
+      UTIL_FORMAT_SWIZZLE_Y,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   &util_format_x24s8_uscaled_unpack_s_8uscaled,
+   &util_format_x24s8_uscaled_pack_s_8uscaled
+};
+
+const struct util_format_description
+util_format_s8x24_uscaled_description = {
+   PIPE_FORMAT_S8X24_USCALED,
+   "PIPE_FORMAT_S8X24_USCALED",
+   "s8x24_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = s */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 24},	/* y = x */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_NONE,	/* z */
+      UTIL_FORMAT_SWIZZLE_X,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   &util_format_s8x24_uscaled_unpack_s_8uscaled,
+   &util_format_s8x24_uscaled_pack_s_8uscaled
+};
+
+const struct util_format_description
+util_format_z24x8_unorm_description = {
+   PIPE_FORMAT_Z24X8_UNORM,
+   "PIPE_FORMAT_Z24X8_UNORM",
+   "z24x8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 24},	/* x = z */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8},	/* y = x */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* z */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_z24x8_unorm_unpack_z_32unorm,
+   &util_format_z24x8_unorm_pack_z_32unorm,
+   &util_format_z24x8_unorm_unpack_z_float,
+   &util_format_z24x8_unorm_pack_z_float,
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_x8z24_unorm_description = {
+   PIPE_FORMAT_X8Z24_UNORM,
+   "PIPE_FORMAT_X8Z24_UNORM",
+   "x8z24_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8},	/* x = x */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 24},	/* y = z */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_Y,	/* z */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_x8z24_unorm_unpack_z_32unorm,
+   &util_format_x8z24_unorm_pack_z_32unorm,
+   &util_format_x8z24_unorm_unpack_z_float,
+   &util_format_x8z24_unorm_pack_z_float,
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_z32_float_s8x24_uscaled_description = {
+   PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED,
+   "PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED",
+   "z32_float_s8x24_uscaled",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   TRUE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* x = z */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* y = s */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 24},	/* z = x */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* z */
+      UTIL_FORMAT_SWIZZLE_Y,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   &util_format_z32_float_s8x24_uscaled_unpack_z_32unorm,
+   &util_format_z32_float_s8x24_uscaled_pack_z_32unorm,
+   &util_format_z32_float_s8x24_uscaled_unpack_z_float,
+   &util_format_z32_float_s8x24_uscaled_pack_z_float,
+   &util_format_z32_float_s8x24_uscaled_unpack_s_8uscaled,
+   &util_format_z32_float_s8x24_uscaled_pack_s_8uscaled
+};
+
+const struct util_format_description
+util_format_x32_s8x24_uscaled_description = {
+   PIPE_FORMAT_X32_S8X24_USCALED,
+   "PIPE_FORMAT_X32_S8X24_USCALED",
+   "x32_s8x24_uscaled",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 32},	/* x = x */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* y = s */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 24},	/* z = x */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_NONE,	/* z */
+      UTIL_FORMAT_SWIZZLE_Y,	/* s */
+      UTIL_FORMAT_SWIZZLE_NONE,	/* ignored */
+      UTIL_FORMAT_SWIZZLE_NONE	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_ZS,
+   NULL, /* unpack_rgba_8unorm */
+   NULL, /* pack_rgba_8unorm */
+   NULL, /* fetch_rgba_8unorm */
+   NULL, /* unpack_rgba_float */
+   NULL, /* pack_rgba_float */
+   NULL, /* fetch_rgba_float */
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   &util_format_x32_s8x24_uscaled_unpack_s_8uscaled,
+   &util_format_x32_s8x24_uscaled_pack_s_8uscaled
+};
+
+const struct util_format_description
+util_format_uyvy_description = {
+   PIPE_FORMAT_UYVY,
+   "PIPE_FORMAT_UYVY",
+   "uyvy",
+   {2, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_SUBSAMPLED,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 32},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* y */
+      UTIL_FORMAT_SWIZZLE_Y,	/* u */
+      UTIL_FORMAT_SWIZZLE_Z,	/* v */
+      UTIL_FORMAT_SWIZZLE_1	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_YUV,
+   &util_format_uyvy_unpack_rgba_8unorm,
+   &util_format_uyvy_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_uyvy_unpack_rgba_float,
+   &util_format_uyvy_pack_rgba_float,
+   &util_format_uyvy_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_yuyv_description = {
+   PIPE_FORMAT_YUYV,
+   "PIPE_FORMAT_YUYV",
+   "yuyv",
+   {2, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_SUBSAMPLED,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 32},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* y */
+      UTIL_FORMAT_SWIZZLE_Y,	/* u */
+      UTIL_FORMAT_SWIZZLE_Z,	/* v */
+      UTIL_FORMAT_SWIZZLE_1	/* ignored */
+   },
+   UTIL_FORMAT_COLORSPACE_YUV,
+   &util_format_yuyv_unpack_rgba_8unorm,
+   &util_format_yuyv_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_yuyv_unpack_rgba_float,
+   &util_format_yuyv_pack_rgba_float,
+   &util_format_yuyv_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8_b8g8_unorm_description = {
+   PIPE_FORMAT_R8G8_B8G8_UNORM,
+   "PIPE_FORMAT_R8G8_B8G8_UNORM",
+   "r8g8_b8g8_unorm",
+   {2, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_SUBSAMPLED,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 32},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8_b8g8_unorm_unpack_rgba_8unorm,
+   &util_format_r8g8_b8g8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8_b8g8_unorm_unpack_rgba_float,
+   &util_format_r8g8_b8g8_unorm_pack_rgba_float,
+   &util_format_r8g8_b8g8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_g8r8_g8b8_unorm_description = {
+   PIPE_FORMAT_G8R8_G8B8_UNORM,
+   "PIPE_FORMAT_G8R8_G8B8_UNORM",
+   "g8r8_g8b8_unorm",
+   {2, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_SUBSAMPLED,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 32},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_g8r8_g8b8_unorm_unpack_rgba_8unorm,
+   &util_format_g8r8_g8b8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_g8r8_g8b8_unorm_unpack_rgba_float,
+   &util_format_g8r8_g8b8_unorm_pack_rgba_float,
+   &util_format_g8r8_g8b8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r10g10b10a2_uscaled_description = {
+   PIPE_FORMAT_R10G10B10A2_USCALED,
+   "PIPE_FORMAT_R10G10B10A2_USCALED",
+   "r10g10b10a2_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 10},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 10},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 10},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 2}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r10g10b10a2_uscaled_unpack_rgba_8unorm,
+   &util_format_r10g10b10a2_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r10g10b10a2_uscaled_unpack_rgba_float,
+   &util_format_r10g10b10a2_uscaled_pack_rgba_float,
+   &util_format_r10g10b10a2_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r11g11b10_float_description = {
+   PIPE_FORMAT_R11G11B10_FLOAT,
+   "PIPE_FORMAT_R11G11B10_FLOAT",
+   "r11g11b10_float",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 11},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 11},	/* y = g */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 10},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r11g11b10_float_unpack_rgba_8unorm,
+   &util_format_r11g11b10_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r11g11b10_float_unpack_rgba_float,
+   &util_format_r11g11b10_float_pack_rgba_float,
+   &util_format_r11g11b10_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r9g9b9e5_float_description = {
+   PIPE_FORMAT_R9G9B9E5_FLOAT,
+   "PIPE_FORMAT_R9G9B9E5_FLOAT",
+   "r9g9b9e5_float",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_OTHER,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 32},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r9g9b9e5_float_unpack_rgba_8unorm,
+   &util_format_r9g9b9e5_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r9g9b9e5_float_unpack_rgba_float,
+   &util_format_r9g9b9e5_float_pack_rgba_float,
+   &util_format_r9g9b9e5_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r1_unorm_description = {
+   PIPE_FORMAT_R1_UNORM,
+   "PIPE_FORMAT_R1_UNORM",
+   "r1_unorm",
+   {8, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_OTHER,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 8},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r1_unorm_unpack_rgba_8unorm,
+   &util_format_r1_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r1_unorm_unpack_rgba_float,
+   &util_format_r1_unorm_pack_rgba_float,
+   &util_format_r1_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8bx_snorm_description = {
+   PIPE_FORMAT_R8G8Bx_SNORM,
+   "PIPE_FORMAT_R8G8Bx_SNORM",
+   "r8g8bx_snorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_OTHER,
+   2,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* x = x */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* y = y */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8bx_snorm_unpack_rgba_8unorm,
+   &util_format_r8g8bx_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8bx_snorm_unpack_rgba_float,
+   &util_format_r8g8bx_snorm_pack_rgba_float,
+   &util_format_r8g8bx_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt1_rgb_description = {
+   PIPE_FORMAT_DXT1_RGB,
+   "PIPE_FORMAT_DXT1_RGB",
+   "dxt1_rgb",
+   {4, 4, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 64},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_dxt1_rgb_unpack_rgba_8unorm,
+   &util_format_dxt1_rgb_pack_rgba_8unorm,
+   &util_format_dxt1_rgb_fetch_rgba_8unorm,
+   &util_format_dxt1_rgb_unpack_rgba_float,
+   &util_format_dxt1_rgb_pack_rgba_float,
+   &util_format_dxt1_rgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt1_rgba_description = {
+   PIPE_FORMAT_DXT1_RGBA,
+   "PIPE_FORMAT_DXT1_RGBA",
+   "dxt1_rgba",
+   {4, 4, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 64},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_dxt1_rgba_unpack_rgba_8unorm,
+   &util_format_dxt1_rgba_pack_rgba_8unorm,
+   &util_format_dxt1_rgba_fetch_rgba_8unorm,
+   &util_format_dxt1_rgba_unpack_rgba_float,
+   &util_format_dxt1_rgba_pack_rgba_float,
+   &util_format_dxt1_rgba_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt3_rgba_description = {
+   PIPE_FORMAT_DXT3_RGBA,
+   "PIPE_FORMAT_DXT3_RGBA",
+   "dxt3_rgba",
+   {4, 4, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 128},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_dxt3_rgba_unpack_rgba_8unorm,
+   &util_format_dxt3_rgba_pack_rgba_8unorm,
+   &util_format_dxt3_rgba_fetch_rgba_8unorm,
+   &util_format_dxt3_rgba_unpack_rgba_float,
+   &util_format_dxt3_rgba_pack_rgba_float,
+   &util_format_dxt3_rgba_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt5_rgba_description = {
+   PIPE_FORMAT_DXT5_RGBA,
+   "PIPE_FORMAT_DXT5_RGBA",
+   "dxt5_rgba",
+   {4, 4, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 128},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_dxt5_rgba_unpack_rgba_8unorm,
+   &util_format_dxt5_rgba_pack_rgba_8unorm,
+   &util_format_dxt5_rgba_fetch_rgba_8unorm,
+   &util_format_dxt5_rgba_unpack_rgba_float,
+   &util_format_dxt5_rgba_pack_rgba_float,
+   &util_format_dxt5_rgba_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt1_srgb_description = {
+   PIPE_FORMAT_DXT1_SRGB,
+   "PIPE_FORMAT_DXT1_SRGB",
+   "dxt1_srgb",
+   {4, 4, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 64},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sb */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_dxt1_srgb_unpack_rgba_8unorm,
+   &util_format_dxt1_srgb_pack_rgba_8unorm,
+   &util_format_dxt1_srgb_fetch_rgba_8unorm,
+   &util_format_dxt1_srgb_unpack_rgba_float,
+   &util_format_dxt1_srgb_pack_rgba_float,
+   &util_format_dxt1_srgb_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt1_srgba_description = {
+   PIPE_FORMAT_DXT1_SRGBA,
+   "PIPE_FORMAT_DXT1_SRGBA",
+   "dxt1_srgba",
+   {4, 4, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 64},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sb */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_dxt1_srgba_unpack_rgba_8unorm,
+   &util_format_dxt1_srgba_pack_rgba_8unorm,
+   &util_format_dxt1_srgba_fetch_rgba_8unorm,
+   &util_format_dxt1_srgba_unpack_rgba_float,
+   &util_format_dxt1_srgba_pack_rgba_float,
+   &util_format_dxt1_srgba_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt3_srgba_description = {
+   PIPE_FORMAT_DXT3_SRGBA,
+   "PIPE_FORMAT_DXT3_SRGBA",
+   "dxt3_srgba",
+   {4, 4, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 128},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sb */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_dxt3_srgba_unpack_rgba_8unorm,
+   &util_format_dxt3_srgba_pack_rgba_8unorm,
+   &util_format_dxt3_srgba_fetch_rgba_8unorm,
+   &util_format_dxt3_srgba_unpack_rgba_float,
+   &util_format_dxt3_srgba_pack_rgba_float,
+   &util_format_dxt3_srgba_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_dxt5_srgba_description = {
+   PIPE_FORMAT_DXT5_SRGBA,
+   "PIPE_FORMAT_DXT5_SRGBA",
+   "dxt5_srgba",
+   {4, 4, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_S3TC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 128},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* sr */
+      UTIL_FORMAT_SWIZZLE_Y,	/* sg */
+      UTIL_FORMAT_SWIZZLE_Z,	/* sb */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_SRGB,
+   &util_format_dxt5_srgba_unpack_rgba_8unorm,
+   &util_format_dxt5_srgba_pack_rgba_8unorm,
+   &util_format_dxt5_srgba_fetch_rgba_8unorm,
+   &util_format_dxt5_srgba_unpack_rgba_float,
+   &util_format_dxt5_srgba_pack_rgba_float,
+   &util_format_dxt5_srgba_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_rgtc1_unorm_description = {
+   PIPE_FORMAT_RGTC1_UNORM,
+   "PIPE_FORMAT_RGTC1_UNORM",
+   "rgtc1_unorm",
+   {4, 4, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_RGTC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 64},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_rgtc1_unorm_unpack_rgba_8unorm,
+   &util_format_rgtc1_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_rgtc1_unorm_unpack_rgba_float,
+   &util_format_rgtc1_unorm_pack_rgba_float,
+   &util_format_rgtc1_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_rgtc1_snorm_description = {
+   PIPE_FORMAT_RGTC1_SNORM,
+   "PIPE_FORMAT_RGTC1_SNORM",
+   "rgtc1_snorm",
+   {4, 4, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_RGTC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 64},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_rgtc1_snorm_unpack_rgba_8unorm,
+   &util_format_rgtc1_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_rgtc1_snorm_unpack_rgba_float,
+   &util_format_rgtc1_snorm_pack_rgba_float,
+   &util_format_rgtc1_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_rgtc2_unorm_description = {
+   PIPE_FORMAT_RGTC2_UNORM,
+   "PIPE_FORMAT_RGTC2_UNORM",
+   "rgtc2_unorm",
+   {4, 4, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_RGTC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 128},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_rgtc2_unorm_unpack_rgba_8unorm,
+   &util_format_rgtc2_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_rgtc2_unorm_unpack_rgba_float,
+   &util_format_rgtc2_unorm_pack_rgba_float,
+   &util_format_rgtc2_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_rgtc2_snorm_description = {
+   PIPE_FORMAT_RGTC2_SNORM,
+   "PIPE_FORMAT_RGTC2_SNORM",
+   "rgtc2_snorm",
+   {4, 4, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_RGTC,
+   1,	/* nr_channels */
+   FALSE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 128},	/* x = x */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_rgtc2_snorm_unpack_rgba_8unorm,
+   &util_format_rgtc2_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_rgtc2_snorm_unpack_rgba_float,
+   &util_format_rgtc2_snorm_pack_rgba_float,
+   &util_format_rgtc2_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r64_float_description = {
+   PIPE_FORMAT_R64_FLOAT,
+   "PIPE_FORMAT_R64_FLOAT",
+   "r64_float",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r64_float_unpack_rgba_8unorm,
+   &util_format_r64_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r64_float_unpack_rgba_float,
+   &util_format_r64_float_pack_rgba_float,
+   &util_format_r64_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r64g64_float_description = {
+   PIPE_FORMAT_R64G64_FLOAT,
+   "PIPE_FORMAT_R64G64_FLOAT",
+   "r64g64_float",
+   {1, 1, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r64g64_float_unpack_rgba_8unorm,
+   &util_format_r64g64_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r64g64_float_unpack_rgba_float,
+   &util_format_r64g64_float_pack_rgba_float,
+   &util_format_r64g64_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r64g64b64_float_description = {
+   PIPE_FORMAT_R64G64B64_FLOAT,
+   "PIPE_FORMAT_R64G64B64_FLOAT",
+   "r64g64b64_float",
+   {1, 1, 192},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* y = g */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r64g64b64_float_unpack_rgba_8unorm,
+   &util_format_r64g64b64_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r64g64b64_float_unpack_rgba_float,
+   &util_format_r64g64b64_float_pack_rgba_float,
+   &util_format_r64g64b64_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r64g64b64a64_float_description = {
+   PIPE_FORMAT_R64G64B64A64_FLOAT,
+   "PIPE_FORMAT_R64G64B64A64_FLOAT",
+   "r64g64b64a64_float",
+   {1, 1, 256},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* y = g */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64},	/* z = b */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 64}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r64g64b64a64_float_unpack_rgba_8unorm,
+   &util_format_r64g64b64a64_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r64g64b64a64_float_unpack_rgba_float,
+   &util_format_r64g64b64a64_float_pack_rgba_float,
+   &util_format_r64g64b64a64_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32_float_description = {
+   PIPE_FORMAT_R32_FLOAT,
+   "PIPE_FORMAT_R32_FLOAT",
+   "r32_float",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32_float_unpack_rgba_8unorm,
+   &util_format_r32_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32_float_unpack_rgba_float,
+   &util_format_r32_float_pack_rgba_float,
+   &util_format_r32_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32_float_description = {
+   PIPE_FORMAT_R32G32_FLOAT,
+   "PIPE_FORMAT_R32G32_FLOAT",
+   "r32g32_float",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32_float_unpack_rgba_8unorm,
+   &util_format_r32g32_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32_float_unpack_rgba_float,
+   &util_format_r32g32_float_pack_rgba_float,
+   &util_format_r32g32_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32_float_description = {
+   PIPE_FORMAT_R32G32B32_FLOAT,
+   "PIPE_FORMAT_R32G32B32_FLOAT",
+   "r32g32b32_float",
+   {1, 1, 96},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32_float_unpack_rgba_8unorm,
+   &util_format_r32g32b32_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32_float_unpack_rgba_float,
+   &util_format_r32g32b32_float_pack_rgba_float,
+   &util_format_r32g32b32_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32a32_float_description = {
+   PIPE_FORMAT_R32G32B32A32_FLOAT,
+   "PIPE_FORMAT_R32G32B32A32_FLOAT",
+   "r32g32b32a32_float",
+   {1, 1, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32},	/* z = b */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 32}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32a32_float_unpack_rgba_8unorm,
+   &util_format_r32g32b32a32_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32a32_float_unpack_rgba_float,
+   &util_format_r32g32b32a32_float_pack_rgba_float,
+   &util_format_r32g32b32a32_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32_unorm_description = {
+   PIPE_FORMAT_R32_UNORM,
+   "PIPE_FORMAT_R32_UNORM",
+   "r32_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32_unorm_unpack_rgba_8unorm,
+   &util_format_r32_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32_unorm_unpack_rgba_float,
+   &util_format_r32_unorm_pack_rgba_float,
+   &util_format_r32_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32_unorm_description = {
+   PIPE_FORMAT_R32G32_UNORM,
+   "PIPE_FORMAT_R32G32_UNORM",
+   "r32g32_unorm",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32_unorm_unpack_rgba_8unorm,
+   &util_format_r32g32_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32_unorm_unpack_rgba_float,
+   &util_format_r32g32_unorm_pack_rgba_float,
+   &util_format_r32g32_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32_unorm_description = {
+   PIPE_FORMAT_R32G32B32_UNORM,
+   "PIPE_FORMAT_R32G32B32_UNORM",
+   "r32g32b32_unorm",
+   {1, 1, 96},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32_unorm_unpack_rgba_8unorm,
+   &util_format_r32g32b32_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32_unorm_unpack_rgba_float,
+   &util_format_r32g32b32_unorm_pack_rgba_float,
+   &util_format_r32g32b32_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32a32_unorm_description = {
+   PIPE_FORMAT_R32G32B32A32_UNORM,
+   "PIPE_FORMAT_R32G32B32A32_UNORM",
+   "r32g32b32a32_unorm",
+   {1, 1, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 32}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32a32_unorm_unpack_rgba_8unorm,
+   &util_format_r32g32b32a32_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32a32_unorm_unpack_rgba_float,
+   &util_format_r32g32b32a32_unorm_pack_rgba_float,
+   &util_format_r32g32b32a32_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32_uscaled_description = {
+   PIPE_FORMAT_R32_USCALED,
+   "PIPE_FORMAT_R32_USCALED",
+   "r32_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32_uscaled_unpack_rgba_8unorm,
+   &util_format_r32_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32_uscaled_unpack_rgba_float,
+   &util_format_r32_uscaled_pack_rgba_float,
+   &util_format_r32_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32_uscaled_description = {
+   PIPE_FORMAT_R32G32_USCALED,
+   "PIPE_FORMAT_R32G32_USCALED",
+   "r32g32_uscaled",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32_uscaled_unpack_rgba_8unorm,
+   &util_format_r32g32_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32_uscaled_unpack_rgba_float,
+   &util_format_r32g32_uscaled_pack_rgba_float,
+   &util_format_r32g32_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32_uscaled_description = {
+   PIPE_FORMAT_R32G32B32_USCALED,
+   "PIPE_FORMAT_R32G32B32_USCALED",
+   "r32g32b32_uscaled",
+   {1, 1, 96},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32_uscaled_unpack_rgba_8unorm,
+   &util_format_r32g32b32_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32_uscaled_unpack_rgba_float,
+   &util_format_r32g32b32_uscaled_pack_rgba_float,
+   &util_format_r32g32b32_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32a32_uscaled_description = {
+   PIPE_FORMAT_R32G32B32A32_USCALED,
+   "PIPE_FORMAT_R32G32B32A32_USCALED",
+   "r32g32b32a32_uscaled",
+   {1, 1, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 32}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32a32_uscaled_unpack_rgba_8unorm,
+   &util_format_r32g32b32a32_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32a32_uscaled_unpack_rgba_float,
+   &util_format_r32g32b32a32_uscaled_pack_rgba_float,
+   &util_format_r32g32b32a32_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32_snorm_description = {
+   PIPE_FORMAT_R32_SNORM,
+   "PIPE_FORMAT_R32_SNORM",
+   "r32_snorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32_snorm_unpack_rgba_8unorm,
+   &util_format_r32_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32_snorm_unpack_rgba_float,
+   &util_format_r32_snorm_pack_rgba_float,
+   &util_format_r32_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32_snorm_description = {
+   PIPE_FORMAT_R32G32_SNORM,
+   "PIPE_FORMAT_R32G32_SNORM",
+   "r32g32_snorm",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32_snorm_unpack_rgba_8unorm,
+   &util_format_r32g32_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32_snorm_unpack_rgba_float,
+   &util_format_r32g32_snorm_pack_rgba_float,
+   &util_format_r32g32_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32_snorm_description = {
+   PIPE_FORMAT_R32G32B32_SNORM,
+   "PIPE_FORMAT_R32G32B32_SNORM",
+   "r32g32b32_snorm",
+   {1, 1, 96},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32_snorm_unpack_rgba_8unorm,
+   &util_format_r32g32b32_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32_snorm_unpack_rgba_float,
+   &util_format_r32g32b32_snorm_pack_rgba_float,
+   &util_format_r32g32b32_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32a32_snorm_description = {
+   PIPE_FORMAT_R32G32B32A32_SNORM,
+   "PIPE_FORMAT_R32G32B32A32_SNORM",
+   "r32g32b32a32_snorm",
+   {1, 1, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32},	/* z = b */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 32}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32a32_snorm_unpack_rgba_8unorm,
+   &util_format_r32g32b32a32_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32a32_snorm_unpack_rgba_float,
+   &util_format_r32g32b32a32_snorm_pack_rgba_float,
+   &util_format_r32g32b32a32_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32_sscaled_description = {
+   PIPE_FORMAT_R32_SSCALED,
+   "PIPE_FORMAT_R32_SSCALED",
+   "r32_sscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32_sscaled_unpack_rgba_8unorm,
+   &util_format_r32_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32_sscaled_unpack_rgba_float,
+   &util_format_r32_sscaled_pack_rgba_float,
+   &util_format_r32_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32_sscaled_description = {
+   PIPE_FORMAT_R32G32_SSCALED,
+   "PIPE_FORMAT_R32G32_SSCALED",
+   "r32g32_sscaled",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32_sscaled_unpack_rgba_8unorm,
+   &util_format_r32g32_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32_sscaled_unpack_rgba_float,
+   &util_format_r32g32_sscaled_pack_rgba_float,
+   &util_format_r32g32_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32_sscaled_description = {
+   PIPE_FORMAT_R32G32B32_SSCALED,
+   "PIPE_FORMAT_R32G32B32_SSCALED",
+   "r32g32b32_sscaled",
+   {1, 1, 96},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32_sscaled_unpack_rgba_8unorm,
+   &util_format_r32g32b32_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32_sscaled_unpack_rgba_float,
+   &util_format_r32g32b32_sscaled_pack_rgba_float,
+   &util_format_r32g32b32_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32a32_sscaled_description = {
+   PIPE_FORMAT_R32G32B32A32_SSCALED,
+   "PIPE_FORMAT_R32G32B32A32_SSCALED",
+   "r32g32b32a32_sscaled",
+   {1, 1, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32},	/* z = b */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 32}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32a32_sscaled_unpack_rgba_8unorm,
+   &util_format_r32g32b32a32_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32a32_sscaled_unpack_rgba_float,
+   &util_format_r32g32b32a32_sscaled_pack_rgba_float,
+   &util_format_r32g32b32a32_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16_float_description = {
+   PIPE_FORMAT_R16_FLOAT,
+   "PIPE_FORMAT_R16_FLOAT",
+   "r16_float",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16_float_unpack_rgba_8unorm,
+   &util_format_r16_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16_float_unpack_rgba_float,
+   &util_format_r16_float_pack_rgba_float,
+   &util_format_r16_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16_float_description = {
+   PIPE_FORMAT_R16G16_FLOAT,
+   "PIPE_FORMAT_R16G16_FLOAT",
+   "r16g16_float",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16_float_unpack_rgba_8unorm,
+   &util_format_r16g16_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16_float_unpack_rgba_float,
+   &util_format_r16g16_float_pack_rgba_float,
+   &util_format_r16g16_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16_float_description = {
+   PIPE_FORMAT_R16G16B16_FLOAT,
+   "PIPE_FORMAT_R16G16B16_FLOAT",
+   "r16g16b16_float",
+   {1, 1, 48},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16_float_unpack_rgba_8unorm,
+   &util_format_r16g16b16_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16_float_unpack_rgba_float,
+   &util_format_r16g16b16_float_pack_rgba_float,
+   &util_format_r16g16b16_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16a16_float_description = {
+   PIPE_FORMAT_R16G16B16A16_FLOAT,
+   "PIPE_FORMAT_R16G16B16A16_FLOAT",
+   "r16g16b16a16_float",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16},	/* z = b */
+      {UTIL_FORMAT_TYPE_FLOAT, FALSE, 16}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16a16_float_unpack_rgba_8unorm,
+   &util_format_r16g16b16a16_float_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16a16_float_unpack_rgba_float,
+   &util_format_r16g16b16a16_float_pack_rgba_float,
+   &util_format_r16g16b16a16_float_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16_unorm_description = {
+   PIPE_FORMAT_R16_UNORM,
+   "PIPE_FORMAT_R16_UNORM",
+   "r16_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16_unorm_unpack_rgba_8unorm,
+   &util_format_r16_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16_unorm_unpack_rgba_float,
+   &util_format_r16_unorm_pack_rgba_float,
+   &util_format_r16_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16_unorm_description = {
+   PIPE_FORMAT_R16G16_UNORM,
+   "PIPE_FORMAT_R16G16_UNORM",
+   "r16g16_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16_unorm_unpack_rgba_8unorm,
+   &util_format_r16g16_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16_unorm_unpack_rgba_float,
+   &util_format_r16g16_unorm_pack_rgba_float,
+   &util_format_r16g16_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16_unorm_description = {
+   PIPE_FORMAT_R16G16B16_UNORM,
+   "PIPE_FORMAT_R16G16B16_UNORM",
+   "r16g16b16_unorm",
+   {1, 1, 48},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16_unorm_unpack_rgba_8unorm,
+   &util_format_r16g16b16_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16_unorm_unpack_rgba_float,
+   &util_format_r16g16b16_unorm_pack_rgba_float,
+   &util_format_r16g16b16_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16a16_unorm_description = {
+   PIPE_FORMAT_R16G16B16A16_UNORM,
+   "PIPE_FORMAT_R16G16B16A16_UNORM",
+   "r16g16b16a16_unorm",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 16}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16a16_unorm_unpack_rgba_8unorm,
+   &util_format_r16g16b16a16_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16a16_unorm_unpack_rgba_float,
+   &util_format_r16g16b16a16_unorm_pack_rgba_float,
+   &util_format_r16g16b16a16_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16_uscaled_description = {
+   PIPE_FORMAT_R16_USCALED,
+   "PIPE_FORMAT_R16_USCALED",
+   "r16_uscaled",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16_uscaled_unpack_rgba_8unorm,
+   &util_format_r16_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16_uscaled_unpack_rgba_float,
+   &util_format_r16_uscaled_pack_rgba_float,
+   &util_format_r16_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16_uscaled_description = {
+   PIPE_FORMAT_R16G16_USCALED,
+   "PIPE_FORMAT_R16G16_USCALED",
+   "r16g16_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16_uscaled_unpack_rgba_8unorm,
+   &util_format_r16g16_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16_uscaled_unpack_rgba_float,
+   &util_format_r16g16_uscaled_pack_rgba_float,
+   &util_format_r16g16_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16_uscaled_description = {
+   PIPE_FORMAT_R16G16B16_USCALED,
+   "PIPE_FORMAT_R16G16B16_USCALED",
+   "r16g16b16_uscaled",
+   {1, 1, 48},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16_uscaled_unpack_rgba_8unorm,
+   &util_format_r16g16b16_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16_uscaled_unpack_rgba_float,
+   &util_format_r16g16b16_uscaled_pack_rgba_float,
+   &util_format_r16g16b16_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16a16_uscaled_description = {
+   PIPE_FORMAT_R16G16B16A16_USCALED,
+   "PIPE_FORMAT_R16G16B16A16_USCALED",
+   "r16g16b16a16_uscaled",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 16}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16a16_uscaled_unpack_rgba_8unorm,
+   &util_format_r16g16b16a16_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16a16_uscaled_unpack_rgba_float,
+   &util_format_r16g16b16a16_uscaled_pack_rgba_float,
+   &util_format_r16g16b16a16_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16_snorm_description = {
+   PIPE_FORMAT_R16_SNORM,
+   "PIPE_FORMAT_R16_SNORM",
+   "r16_snorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16_snorm_unpack_rgba_8unorm,
+   &util_format_r16_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16_snorm_unpack_rgba_float,
+   &util_format_r16_snorm_pack_rgba_float,
+   &util_format_r16_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16_snorm_description = {
+   PIPE_FORMAT_R16G16_SNORM,
+   "PIPE_FORMAT_R16G16_SNORM",
+   "r16g16_snorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16_snorm_unpack_rgba_8unorm,
+   &util_format_r16g16_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16_snorm_unpack_rgba_float,
+   &util_format_r16g16_snorm_pack_rgba_float,
+   &util_format_r16g16_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16_snorm_description = {
+   PIPE_FORMAT_R16G16B16_SNORM,
+   "PIPE_FORMAT_R16G16B16_SNORM",
+   "r16g16b16_snorm",
+   {1, 1, 48},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16_snorm_unpack_rgba_8unorm,
+   &util_format_r16g16b16_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16_snorm_unpack_rgba_float,
+   &util_format_r16g16b16_snorm_pack_rgba_float,
+   &util_format_r16g16b16_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16a16_snorm_description = {
+   PIPE_FORMAT_R16G16B16A16_SNORM,
+   "PIPE_FORMAT_R16G16B16A16_SNORM",
+   "r16g16b16a16_snorm",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16},	/* z = b */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 16}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16a16_snorm_unpack_rgba_8unorm,
+   &util_format_r16g16b16a16_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16a16_snorm_unpack_rgba_float,
+   &util_format_r16g16b16a16_snorm_pack_rgba_float,
+   &util_format_r16g16b16a16_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16_sscaled_description = {
+   PIPE_FORMAT_R16_SSCALED,
+   "PIPE_FORMAT_R16_SSCALED",
+   "r16_sscaled",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16_sscaled_unpack_rgba_8unorm,
+   &util_format_r16_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16_sscaled_unpack_rgba_float,
+   &util_format_r16_sscaled_pack_rgba_float,
+   &util_format_r16_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16_sscaled_description = {
+   PIPE_FORMAT_R16G16_SSCALED,
+   "PIPE_FORMAT_R16G16_SSCALED",
+   "r16g16_sscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16_sscaled_unpack_rgba_8unorm,
+   &util_format_r16g16_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16_sscaled_unpack_rgba_float,
+   &util_format_r16g16_sscaled_pack_rgba_float,
+   &util_format_r16g16_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16_sscaled_description = {
+   PIPE_FORMAT_R16G16B16_SSCALED,
+   "PIPE_FORMAT_R16G16B16_SSCALED",
+   "r16g16b16_sscaled",
+   {1, 1, 48},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16_sscaled_unpack_rgba_8unorm,
+   &util_format_r16g16b16_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16_sscaled_unpack_rgba_float,
+   &util_format_r16g16b16_sscaled_pack_rgba_float,
+   &util_format_r16g16b16_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r16g16b16a16_sscaled_description = {
+   PIPE_FORMAT_R16G16B16A16_SSCALED,
+   "PIPE_FORMAT_R16G16B16A16_SSCALED",
+   "r16g16b16a16_sscaled",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16},	/* z = b */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 16}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r16g16b16a16_sscaled_unpack_rgba_8unorm,
+   &util_format_r16g16b16a16_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r16g16b16a16_sscaled_unpack_rgba_float,
+   &util_format_r16g16b16a16_sscaled_pack_rgba_float,
+   &util_format_r16g16b16a16_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8_unorm_description = {
+   PIPE_FORMAT_R8_UNORM,
+   "PIPE_FORMAT_R8_UNORM",
+   "r8_unorm",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8_unorm_unpack_rgba_8unorm,
+   &util_format_r8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8_unorm_unpack_rgba_float,
+   &util_format_r8_unorm_pack_rgba_float,
+   &util_format_r8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8_unorm_description = {
+   PIPE_FORMAT_R8G8_UNORM,
+   "PIPE_FORMAT_R8G8_UNORM",
+   "r8g8_unorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8_unorm_unpack_rgba_8unorm,
+   &util_format_r8g8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8_unorm_unpack_rgba_float,
+   &util_format_r8g8_unorm_pack_rgba_float,
+   &util_format_r8g8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8_unorm_description = {
+   PIPE_FORMAT_R8G8B8_UNORM,
+   "PIPE_FORMAT_R8G8B8_UNORM",
+   "r8g8b8_unorm",
+   {1, 1, 24},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8_unorm_unpack_rgba_8unorm,
+   &util_format_r8g8b8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8_unorm_unpack_rgba_float,
+   &util_format_r8g8b8_unorm_pack_rgba_float,
+   &util_format_r8g8b8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8a8_unorm_description = {
+   PIPE_FORMAT_R8G8B8A8_UNORM,
+   "PIPE_FORMAT_R8G8B8A8_UNORM",
+   "r8g8b8a8_unorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, TRUE, 8}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8a8_unorm_unpack_rgba_8unorm,
+   &util_format_r8g8b8a8_unorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8a8_unorm_unpack_rgba_float,
+   &util_format_r8g8b8a8_unorm_pack_rgba_float,
+   &util_format_r8g8b8a8_unorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8_uscaled_description = {
+   PIPE_FORMAT_R8_USCALED,
+   "PIPE_FORMAT_R8_USCALED",
+   "r8_uscaled",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8_uscaled_unpack_rgba_8unorm,
+   &util_format_r8_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8_uscaled_unpack_rgba_float,
+   &util_format_r8_uscaled_pack_rgba_float,
+   &util_format_r8_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8_uscaled_description = {
+   PIPE_FORMAT_R8G8_USCALED,
+   "PIPE_FORMAT_R8G8_USCALED",
+   "r8g8_uscaled",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8_uscaled_unpack_rgba_8unorm,
+   &util_format_r8g8_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8_uscaled_unpack_rgba_float,
+   &util_format_r8g8_uscaled_pack_rgba_float,
+   &util_format_r8g8_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8_uscaled_description = {
+   PIPE_FORMAT_R8G8B8_USCALED,
+   "PIPE_FORMAT_R8G8B8_USCALED",
+   "r8g8b8_uscaled",
+   {1, 1, 24},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8_uscaled_unpack_rgba_8unorm,
+   &util_format_r8g8b8_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8_uscaled_unpack_rgba_float,
+   &util_format_r8g8b8_uscaled_pack_rgba_float,
+   &util_format_r8g8b8_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8a8_uscaled_description = {
+   PIPE_FORMAT_R8G8B8A8_USCALED,
+   "PIPE_FORMAT_R8G8B8A8_USCALED",
+   "r8g8b8a8_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8},	/* z = b */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 8}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8a8_uscaled_unpack_rgba_8unorm,
+   &util_format_r8g8b8a8_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8a8_uscaled_unpack_rgba_float,
+   &util_format_r8g8b8a8_uscaled_pack_rgba_float,
+   &util_format_r8g8b8a8_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8_snorm_description = {
+   PIPE_FORMAT_R8_SNORM,
+   "PIPE_FORMAT_R8_SNORM",
+   "r8_snorm",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8_snorm_unpack_rgba_8unorm,
+   &util_format_r8_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8_snorm_unpack_rgba_float,
+   &util_format_r8_snorm_pack_rgba_float,
+   &util_format_r8_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8_snorm_description = {
+   PIPE_FORMAT_R8G8_SNORM,
+   "PIPE_FORMAT_R8G8_SNORM",
+   "r8g8_snorm",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8_snorm_unpack_rgba_8unorm,
+   &util_format_r8g8_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8_snorm_unpack_rgba_float,
+   &util_format_r8g8_snorm_pack_rgba_float,
+   &util_format_r8g8_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8_snorm_description = {
+   PIPE_FORMAT_R8G8B8_SNORM,
+   "PIPE_FORMAT_R8G8B8_SNORM",
+   "r8g8b8_snorm",
+   {1, 1, 24},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8_snorm_unpack_rgba_8unorm,
+   &util_format_r8g8b8_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8_snorm_unpack_rgba_float,
+   &util_format_r8g8b8_snorm_pack_rgba_float,
+   &util_format_r8g8b8_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8a8_snorm_description = {
+   PIPE_FORMAT_R8G8B8A8_SNORM,
+   "PIPE_FORMAT_R8G8B8A8_SNORM",
+   "r8g8b8a8_snorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8},	/* z = b */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 8}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8a8_snorm_unpack_rgba_8unorm,
+   &util_format_r8g8b8a8_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8a8_snorm_unpack_rgba_float,
+   &util_format_r8g8b8a8_snorm_pack_rgba_float,
+   &util_format_r8g8b8a8_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8_sscaled_description = {
+   PIPE_FORMAT_R8_SSCALED,
+   "PIPE_FORMAT_R8_SSCALED",
+   "r8_sscaled",
+   {1, 1, 8},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8_sscaled_unpack_rgba_8unorm,
+   &util_format_r8_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8_sscaled_unpack_rgba_float,
+   &util_format_r8_sscaled_pack_rgba_float,
+   &util_format_r8_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8_sscaled_description = {
+   PIPE_FORMAT_R8G8_SSCALED,
+   "PIPE_FORMAT_R8G8_SSCALED",
+   "r8g8_sscaled",
+   {1, 1, 16},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8_sscaled_unpack_rgba_8unorm,
+   &util_format_r8g8_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8_sscaled_unpack_rgba_float,
+   &util_format_r8g8_sscaled_pack_rgba_float,
+   &util_format_r8g8_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8_sscaled_description = {
+   PIPE_FORMAT_R8G8B8_SSCALED,
+   "PIPE_FORMAT_R8G8B8_SSCALED",
+   "r8g8b8_sscaled",
+   {1, 1, 24},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8_sscaled_unpack_rgba_8unorm,
+   &util_format_r8g8b8_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8_sscaled_unpack_rgba_float,
+   &util_format_r8g8b8_sscaled_pack_rgba_float,
+   &util_format_r8g8b8_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r8g8b8a8_sscaled_description = {
+   PIPE_FORMAT_R8G8B8A8_SSCALED,
+   "PIPE_FORMAT_R8G8B8A8_SSCALED",
+   "r8g8b8a8_sscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8},	/* z = b */
+      {UTIL_FORMAT_TYPE_SIGNED, FALSE, 8}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r8g8b8a8_sscaled_unpack_rgba_8unorm,
+   &util_format_r8g8b8a8_sscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r8g8b8a8_sscaled_unpack_rgba_float,
+   &util_format_r8g8b8a8_sscaled_pack_rgba_float,
+   &util_format_r8g8b8a8_sscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32_fixed_description = {
+   PIPE_FORMAT_R32_FIXED,
+   "PIPE_FORMAT_R32_FIXED",
+   "r32_fixed",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   1,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* x = r */
+      {0, 0, 0},
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_0,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32_fixed_unpack_rgba_8unorm,
+   &util_format_r32_fixed_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32_fixed_unpack_rgba_float,
+   &util_format_r32_fixed_pack_rgba_float,
+   &util_format_r32_fixed_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32_fixed_description = {
+   PIPE_FORMAT_R32G32_FIXED,
+   "PIPE_FORMAT_R32G32_FIXED",
+   "r32g32_fixed",
+   {1, 1, 64},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   2,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* y = g */
+      {0, 0, 0},
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_0,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32_fixed_unpack_rgba_8unorm,
+   &util_format_r32g32_fixed_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32_fixed_unpack_rgba_float,
+   &util_format_r32g32_fixed_pack_rgba_float,
+   &util_format_r32g32_fixed_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32_fixed_description = {
+   PIPE_FORMAT_R32G32B32_FIXED,
+   "PIPE_FORMAT_R32G32B32_FIXED",
+   "r32g32b32_fixed",
+   {1, 1, 96},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   3,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* z = b */
+      {0, 0, 0}
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32_fixed_unpack_rgba_8unorm,
+   &util_format_r32g32b32_fixed_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32_fixed_unpack_rgba_float,
+   &util_format_r32g32b32_fixed_pack_rgba_float,
+   &util_format_r32g32b32_fixed_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r32g32b32a32_fixed_description = {
+   PIPE_FORMAT_R32G32B32A32_FIXED,
+   "PIPE_FORMAT_R32G32B32A32_FIXED",
+   "r32g32b32a32_fixed",
+   {1, 1, 128},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   TRUE,	/* is_array */
+   FALSE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* x = r */
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* y = g */
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32},	/* z = b */
+      {UTIL_FORMAT_TYPE_FIXED, FALSE, 32}	/* w = a */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_W	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r32g32b32a32_fixed_unpack_rgba_8unorm,
+   &util_format_r32g32b32a32_fixed_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r32g32b32a32_fixed_unpack_rgba_float,
+   &util_format_r32g32b32a32_fixed_pack_rgba_float,
+   &util_format_r32g32b32a32_fixed_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r10g10b10x2_uscaled_description = {
+   PIPE_FORMAT_R10G10B10X2_USCALED,
+   "PIPE_FORMAT_R10G10B10X2_USCALED",
+   "r10g10b10x2_uscaled",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 10},	/* x = r */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 10},	/* y = g */
+      {UTIL_FORMAT_TYPE_UNSIGNED, FALSE, 10},	/* z = b */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 2}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r10g10b10x2_uscaled_unpack_rgba_8unorm,
+   &util_format_r10g10b10x2_uscaled_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r10g10b10x2_uscaled_unpack_rgba_float,
+   &util_format_r10g10b10x2_uscaled_pack_rgba_float,
+   &util_format_r10g10b10x2_uscaled_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description
+util_format_r10g10b10x2_snorm_description = {
+   PIPE_FORMAT_R10G10B10X2_SNORM,
+   "PIPE_FORMAT_R10G10B10X2_SNORM",
+   "r10g10b10x2_snorm",
+   {1, 1, 32},	/* block */
+   UTIL_FORMAT_LAYOUT_PLAIN,
+   4,	/* nr_channels */
+   FALSE,	/* is_array */
+   TRUE,	/* is_bitmask */
+   FALSE,	/* is_mixed */
+   {
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 10},	/* x = r */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 10},	/* y = g */
+      {UTIL_FORMAT_TYPE_SIGNED, TRUE, 10},	/* z = b */
+      {UTIL_FORMAT_TYPE_VOID, FALSE, 2}	/* w = x */
+   },
+   {
+      UTIL_FORMAT_SWIZZLE_X,	/* r */
+      UTIL_FORMAT_SWIZZLE_Y,	/* g */
+      UTIL_FORMAT_SWIZZLE_Z,	/* b */
+      UTIL_FORMAT_SWIZZLE_1	/* a */
+   },
+   UTIL_FORMAT_COLORSPACE_RGB,
+   &util_format_r10g10b10x2_snorm_unpack_rgba_8unorm,
+   &util_format_r10g10b10x2_snorm_pack_rgba_8unorm,
+   NULL, /* fetch_rgba_8unorm */
+   &util_format_r10g10b10x2_snorm_unpack_rgba_float,
+   &util_format_r10g10b10x2_snorm_pack_rgba_float,
+   &util_format_r10g10b10x2_snorm_fetch_rgba_float,
+   NULL, /* unpack_z_32unorm */
+   NULL, /* pack_z_32unorm */
+   NULL, /* unpack_z_float */
+   NULL, /* pack_z_float */
+   NULL, /* unpack_s_8uscaled */
+   NULL /* pack_s_8uscaled */
+};
+
+const struct util_format_description *
+util_format_description(enum pipe_format format)
+{
+   if (format >= PIPE_FORMAT_COUNT) {
+      return NULL;
+   }
+
+   switch (format) {
+   case PIPE_FORMAT_NONE:
+      return &util_format_none_description;
+   case PIPE_FORMAT_B8G8R8A8_UNORM:
+      return &util_format_b8g8r8a8_unorm_description;
+   case PIPE_FORMAT_B8G8R8X8_UNORM:
+      return &util_format_b8g8r8x8_unorm_description;
+   case PIPE_FORMAT_A8R8G8B8_UNORM:
+      return &util_format_a8r8g8b8_unorm_description;
+   case PIPE_FORMAT_X8R8G8B8_UNORM:
+      return &util_format_x8r8g8b8_unorm_description;
+   case PIPE_FORMAT_A8B8G8R8_UNORM:
+      return &util_format_a8b8g8r8_unorm_description;
+   case PIPE_FORMAT_X8B8G8R8_UNORM:
+      return &util_format_x8b8g8r8_unorm_description;
+   case PIPE_FORMAT_R8G8B8X8_UNORM:
+      return &util_format_r8g8b8x8_unorm_description;
+   case PIPE_FORMAT_B5G5R5X1_UNORM:
+      return &util_format_b5g5r5x1_unorm_description;
+   case PIPE_FORMAT_B5G5R5A1_UNORM:
+      return &util_format_b5g5r5a1_unorm_description;
+   case PIPE_FORMAT_B4G4R4A4_UNORM:
+      return &util_format_b4g4r4a4_unorm_description;
+   case PIPE_FORMAT_B4G4R4X4_UNORM:
+      return &util_format_b4g4r4x4_unorm_description;
+   case PIPE_FORMAT_B5G6R5_UNORM:
+      return &util_format_b5g6r5_unorm_description;
+   case PIPE_FORMAT_R10G10B10A2_UNORM:
+      return &util_format_r10g10b10a2_unorm_description;
+   case PIPE_FORMAT_B10G10R10A2_UNORM:
+      return &util_format_b10g10r10a2_unorm_description;
+   case PIPE_FORMAT_L8_UNORM:
+      return &util_format_l8_unorm_description;
+   case PIPE_FORMAT_A8_UNORM:
+      return &util_format_a8_unorm_description;
+   case PIPE_FORMAT_I8_UNORM:
+      return &util_format_i8_unorm_description;
+   case PIPE_FORMAT_L4A4_UNORM:
+      return &util_format_l4a4_unorm_description;
+   case PIPE_FORMAT_L8A8_UNORM:
+      return &util_format_l8a8_unorm_description;
+   case PIPE_FORMAT_L16_UNORM:
+      return &util_format_l16_unorm_description;
+   case PIPE_FORMAT_L8_SRGB:
+      return &util_format_l8_srgb_description;
+   case PIPE_FORMAT_L8A8_SRGB:
+      return &util_format_l8a8_srgb_description;
+   case PIPE_FORMAT_R8G8B8_SRGB:
+      return &util_format_r8g8b8_srgb_description;
+   case PIPE_FORMAT_R8G8B8A8_SRGB:
+      return &util_format_r8g8b8a8_srgb_description;
+   case PIPE_FORMAT_A8B8G8R8_SRGB:
+      return &util_format_a8b8g8r8_srgb_description;
+   case PIPE_FORMAT_X8B8G8R8_SRGB:
+      return &util_format_x8b8g8r8_srgb_description;
+   case PIPE_FORMAT_B8G8R8A8_SRGB:
+      return &util_format_b8g8r8a8_srgb_description;
+   case PIPE_FORMAT_B8G8R8X8_SRGB:
+      return &util_format_b8g8r8x8_srgb_description;
+   case PIPE_FORMAT_A8R8G8B8_SRGB:
+      return &util_format_a8r8g8b8_srgb_description;
+   case PIPE_FORMAT_X8R8G8B8_SRGB:
+      return &util_format_x8r8g8b8_srgb_description;
+   case PIPE_FORMAT_R8SG8SB8UX8U_NORM:
+      return &util_format_r8sg8sb8ux8u_norm_description;
+   case PIPE_FORMAT_R10SG10SB10SA2U_NORM:
+      return &util_format_r10sg10sb10sa2u_norm_description;
+   case PIPE_FORMAT_R5SG5SB6U_NORM:
+      return &util_format_r5sg5sb6u_norm_description;
+   case PIPE_FORMAT_S8_USCALED:
+      return &util_format_s8_uscaled_description;
+   case PIPE_FORMAT_Z16_UNORM:
+      return &util_format_z16_unorm_description;
+   case PIPE_FORMAT_Z32_UNORM:
+      return &util_format_z32_unorm_description;
+   case PIPE_FORMAT_Z32_FLOAT:
+      return &util_format_z32_float_description;
+   case PIPE_FORMAT_Z24_UNORM_S8_USCALED:
+      return &util_format_z24_unorm_s8_uscaled_description;
+   case PIPE_FORMAT_S8_USCALED_Z24_UNORM:
+      return &util_format_s8_uscaled_z24_unorm_description;
+   case PIPE_FORMAT_X24S8_USCALED:
+      return &util_format_x24s8_uscaled_description;
+   case PIPE_FORMAT_S8X24_USCALED:
+      return &util_format_s8x24_uscaled_description;
+   case PIPE_FORMAT_Z24X8_UNORM:
+      return &util_format_z24x8_unorm_description;
+   case PIPE_FORMAT_X8Z24_UNORM:
+      return &util_format_x8z24_unorm_description;
+   case PIPE_FORMAT_Z32_FLOAT_S8X24_USCALED:
+      return &util_format_z32_float_s8x24_uscaled_description;
+   case PIPE_FORMAT_X32_S8X24_USCALED:
+      return &util_format_x32_s8x24_uscaled_description;
+   case PIPE_FORMAT_UYVY:
+      return &util_format_uyvy_description;
+   case PIPE_FORMAT_YUYV:
+      return &util_format_yuyv_description;
+   case PIPE_FORMAT_R8G8_B8G8_UNORM:
+      return &util_format_r8g8_b8g8_unorm_description;
+   case PIPE_FORMAT_G8R8_G8B8_UNORM:
+      return &util_format_g8r8_g8b8_unorm_description;
+   case PIPE_FORMAT_R10G10B10A2_USCALED:
+      return &util_format_r10g10b10a2_uscaled_description;
+   case PIPE_FORMAT_R11G11B10_FLOAT:
+      return &util_format_r11g11b10_float_description;
+   case PIPE_FORMAT_R9G9B9E5_FLOAT:
+      return &util_format_r9g9b9e5_float_description;
+   case PIPE_FORMAT_R1_UNORM:
+      return &util_format_r1_unorm_description;
+   case PIPE_FORMAT_R8G8Bx_SNORM:
+      return &util_format_r8g8bx_snorm_description;
+   case PIPE_FORMAT_DXT1_RGB:
+      return &util_format_dxt1_rgb_description;
+   case PIPE_FORMAT_DXT1_RGBA:
+      return &util_format_dxt1_rgba_description;
+   case PIPE_FORMAT_DXT3_RGBA:
+      return &util_format_dxt3_rgba_description;
+   case PIPE_FORMAT_DXT5_RGBA:
+      return &util_format_dxt5_rgba_description;
+   case PIPE_FORMAT_DXT1_SRGB:
+      return &util_format_dxt1_srgb_description;
+   case PIPE_FORMAT_DXT1_SRGBA:
+      return &util_format_dxt1_srgba_description;
+   case PIPE_FORMAT_DXT3_SRGBA:
+      return &util_format_dxt3_srgba_description;
+   case PIPE_FORMAT_DXT5_SRGBA:
+      return &util_format_dxt5_srgba_description;
+   case PIPE_FORMAT_RGTC1_UNORM:
+      return &util_format_rgtc1_unorm_description;
+   case PIPE_FORMAT_RGTC1_SNORM:
+      return &util_format_rgtc1_snorm_description;
+   case PIPE_FORMAT_RGTC2_UNORM:
+      return &util_format_rgtc2_unorm_description;
+   case PIPE_FORMAT_RGTC2_SNORM:
+      return &util_format_rgtc2_snorm_description;
+   case PIPE_FORMAT_R64_FLOAT:
+      return &util_format_r64_float_description;
+   case PIPE_FORMAT_R64G64_FLOAT:
+      return &util_format_r64g64_float_description;
+   case PIPE_FORMAT_R64G64B64_FLOAT:
+      return &util_format_r64g64b64_float_description;
+   case PIPE_FORMAT_R64G64B64A64_FLOAT:
+      return &util_format_r64g64b64a64_float_description;
+   case PIPE_FORMAT_R32_FLOAT:
+      return &util_format_r32_float_description;
+   case PIPE_FORMAT_R32G32_FLOAT:
+      return &util_format_r32g32_float_description;
+   case PIPE_FORMAT_R32G32B32_FLOAT:
+      return &util_format_r32g32b32_float_description;
+   case PIPE_FORMAT_R32G32B32A32_FLOAT:
+      return &util_format_r32g32b32a32_float_description;
+   case PIPE_FORMAT_R32_UNORM:
+      return &util_format_r32_unorm_description;
+   case PIPE_FORMAT_R32G32_UNORM:
+      return &util_format_r32g32_unorm_description;
+   case PIPE_FORMAT_R32G32B32_UNORM:
+      return &util_format_r32g32b32_unorm_description;
+   case PIPE_FORMAT_R32G32B32A32_UNORM:
+      return &util_format_r32g32b32a32_unorm_description;
+   case PIPE_FORMAT_R32_USCALED:
+      return &util_format_r32_uscaled_description;
+   case PIPE_FORMAT_R32G32_USCALED:
+      return &util_format_r32g32_uscaled_description;
+   case PIPE_FORMAT_R32G32B32_USCALED:
+      return &util_format_r32g32b32_uscaled_description;
+   case PIPE_FORMAT_R32G32B32A32_USCALED:
+      return &util_format_r32g32b32a32_uscaled_description;
+   case PIPE_FORMAT_R32_SNORM:
+      return &util_format_r32_snorm_description;
+   case PIPE_FORMAT_R32G32_SNORM:
+      return &util_format_r32g32_snorm_description;
+   case PIPE_FORMAT_R32G32B32_SNORM:
+      return &util_format_r32g32b32_snorm_description;
+   case PIPE_FORMAT_R32G32B32A32_SNORM:
+      return &util_format_r32g32b32a32_snorm_description;
+   case PIPE_FORMAT_R32_SSCALED:
+      return &util_format_r32_sscaled_description;
+   case PIPE_FORMAT_R32G32_SSCALED:
+      return &util_format_r32g32_sscaled_description;
+   case PIPE_FORMAT_R32G32B32_SSCALED:
+      return &util_format_r32g32b32_sscaled_description;
+   case PIPE_FORMAT_R32G32B32A32_SSCALED:
+      return &util_format_r32g32b32a32_sscaled_description;
+   case PIPE_FORMAT_R16_FLOAT:
+      return &util_format_r16_float_description;
+   case PIPE_FORMAT_R16G16_FLOAT:
+      return &util_format_r16g16_float_description;
+   case PIPE_FORMAT_R16G16B16_FLOAT:
+      return &util_format_r16g16b16_float_description;
+   case PIPE_FORMAT_R16G16B16A16_FLOAT:
+      return &util_format_r16g16b16a16_float_description;
+   case PIPE_FORMAT_R16_UNORM:
+      return &util_format_r16_unorm_description;
+   case PIPE_FORMAT_R16G16_UNORM:
+      return &util_format_r16g16_unorm_description;
+   case PIPE_FORMAT_R16G16B16_UNORM:
+      return &util_format_r16g16b16_unorm_description;
+   case PIPE_FORMAT_R16G16B16A16_UNORM:
+      return &util_format_r16g16b16a16_unorm_description;
+   case PIPE_FORMAT_R16_USCALED:
+      return &util_format_r16_uscaled_description;
+   case PIPE_FORMAT_R16G16_USCALED:
+      return &util_format_r16g16_uscaled_description;
+   case PIPE_FORMAT_R16G16B16_USCALED:
+      return &util_format_r16g16b16_uscaled_description;
+   case PIPE_FORMAT_R16G16B16A16_USCALED:
+      return &util_format_r16g16b16a16_uscaled_description;
+   case PIPE_FORMAT_R16_SNORM:
+      return &util_format_r16_snorm_description;
+   case PIPE_FORMAT_R16G16_SNORM:
+      return &util_format_r16g16_snorm_description;
+   case PIPE_FORMAT_R16G16B16_SNORM:
+      return &util_format_r16g16b16_snorm_description;
+   case PIPE_FORMAT_R16G16B16A16_SNORM:
+      return &util_format_r16g16b16a16_snorm_description;
+   case PIPE_FORMAT_R16_SSCALED:
+      return &util_format_r16_sscaled_description;
+   case PIPE_FORMAT_R16G16_SSCALED:
+      return &util_format_r16g16_sscaled_description;
+   case PIPE_FORMAT_R16G16B16_SSCALED:
+      return &util_format_r16g16b16_sscaled_description;
+   case PIPE_FORMAT_R16G16B16A16_SSCALED:
+      return &util_format_r16g16b16a16_sscaled_description;
+   case PIPE_FORMAT_R8_UNORM:
+      return &util_format_r8_unorm_description;
+   case PIPE_FORMAT_R8G8_UNORM:
+      return &util_format_r8g8_unorm_description;
+   case PIPE_FORMAT_R8G8B8_UNORM:
+      return &util_format_r8g8b8_unorm_description;
+   case PIPE_FORMAT_R8G8B8A8_UNORM:
+      return &util_format_r8g8b8a8_unorm_description;
+   case PIPE_FORMAT_R8_USCALED:
+      return &util_format_r8_uscaled_description;
+   case PIPE_FORMAT_R8G8_USCALED:
+      return &util_format_r8g8_uscaled_description;
+   case PIPE_FORMAT_R8G8B8_USCALED:
+      return &util_format_r8g8b8_uscaled_description;
+   case PIPE_FORMAT_R8G8B8A8_USCALED:
+      return &util_format_r8g8b8a8_uscaled_description;
+   case PIPE_FORMAT_R8_SNORM:
+      return &util_format_r8_snorm_description;
+   case PIPE_FORMAT_R8G8_SNORM:
+      return &util_format_r8g8_snorm_description;
+   case PIPE_FORMAT_R8G8B8_SNORM:
+      return &util_format_r8g8b8_snorm_description;
+   case PIPE_FORMAT_R8G8B8A8_SNORM:
+      return &util_format_r8g8b8a8_snorm_description;
+   case PIPE_FORMAT_R8_SSCALED:
+      return &util_format_r8_sscaled_description;
+   case PIPE_FORMAT_R8G8_SSCALED:
+      return &util_format_r8g8_sscaled_description;
+   case PIPE_FORMAT_R8G8B8_SSCALED:
+      return &util_format_r8g8b8_sscaled_description;
+   case PIPE_FORMAT_R8G8B8A8_SSCALED:
+      return &util_format_r8g8b8a8_sscaled_description;
+   case PIPE_FORMAT_R32_FIXED:
+      return &util_format_r32_fixed_description;
+   case PIPE_FORMAT_R32G32_FIXED:
+      return &util_format_r32g32_fixed_description;
+   case PIPE_FORMAT_R32G32B32_FIXED:
+      return &util_format_r32g32b32_fixed_description;
+   case PIPE_FORMAT_R32G32B32A32_FIXED:
+      return &util_format_r32g32b32a32_fixed_description;
+   case PIPE_FORMAT_R10G10B10X2_USCALED:
+      return &util_format_r10g10b10x2_uscaled_description;
+   case PIPE_FORMAT_R10G10B10X2_SNORM:
+      return &util_format_r10g10b10x2_snorm_description;
+   default:
+      return NULL;
+   }
+}
+
diff --git a/lib/libGL/gallium/libgallium/generated/u_half.c b/lib/libGL/gallium/libgallium/generated/u_half.c
new file mode 100644
index 000000000..fe3146d57
--- /dev/null
+++ b/lib/libGL/gallium/libgallium/generated/u_half.c
@@ -0,0 +1,3217 @@
+/* This file is autogenerated by u_half.py. Do not edit directly. */
+#include "util/u_half.h"
+
+const uint32_t util_half_to_float_mantissa_table[2048] = {
+	0x0,
+	0x33800000,
+	0x34000000,
+	0x34400000,
+	0x34800000,
+	0x34a00000,
+	0x34c00000,
+	0x34e00000,
+	0x35000000,
+	0x35100000,
+	0x35200000,
+	0x35300000,
+	0x35400000,
+	0x35500000,
+	0x35600000,
+	0x35700000,
+	0x35800000,
+	0x35880000,
+	0x35900000,
+	0x35980000,
+	0x35a00000,
+	0x35a80000,
+	0x35b00000,
+	0x35b80000,
+	0x35c00000,
+	0x35c80000,
+	0x35d00000,
+	0x35d80000,
+	0x35e00000,
+	0x35e80000,
+	0x35f00000,
+	0x35f80000,
+	0x36000000,
+	0x36040000,
+	0x36080000,
+	0x360c0000,
+	0x36100000,
+	0x36140000,
+	0x36180000,
+	0x361c0000,
+	0x36200000,
+	0x36240000,
+	0x36280000,
+	0x362c0000,
+	0x36300000,
+	0x36340000,
+	0x36380000,
+	0x363c0000,
+	0x36400000,
+	0x36440000,
+	0x36480000,
+	0x364c0000,
+	0x36500000,
+	0x36540000,
+	0x36580000,
+	0x365c0000,
+	0x36600000,
+	0x36640000,
+	0x36680000,
+	0x366c0000,
+	0x36700000,
+	0x36740000,
+	0x36780000,
+	0x367c0000,
+	0x36800000,
+	0x36820000,
+	0x36840000,
+	0x36860000,
+	0x36880000,
+	0x368a0000,
+	0x368c0000,
+	0x368e0000,
+	0x36900000,
+	0x36920000,
+	0x36940000,
+	0x36960000,
+	0x36980000,
+	0x369a0000,
+	0x369c0000,
+	0x369e0000,
+	0x36a00000,
+	0x36a20000,
+	0x36a40000,
+	0x36a60000,
+	0x36a80000,
+	0x36aa0000,
+	0x36ac0000,
+	0x36ae0000,
+	0x36b00000,
+	0x36b20000,
+	0x36b40000,
+	0x36b60000,
+	0x36b80000,
+	0x36ba0000,
+	0x36bc0000,
+	0x36be0000,
+	0x36c00000,
+	0x36c20000,
+	0x36c40000,
+	0x36c60000,
+	0x36c80000,
+	0x36ca0000,
+	0x36cc0000,
+	0x36ce0000,
+	0x36d00000,
+	0x36d20000,
+	0x36d40000,
+	0x36d60000,
+	0x36d80000,
+	0x36da0000,
+	0x36dc0000,
+	0x36de0000,
+	0x36e00000,
+	0x36e20000,
+	0x36e40000,
+	0x36e60000,
+	0x36e80000,
+	0x36ea0000,
+	0x36ec0000,
+	0x36ee0000,
+	0x36f00000,
+	0x36f20000,
+	0x36f40000,
+	0x36f60000,
+	0x36f80000,
+	0x36fa0000,
+	0x36fc0000,
+	0x36fe0000,
+	0x37000000,
+	0x37010000,
+	0x37020000,
+	0x37030000,
+	0x37040000,
+	0x37050000,
+	0x37060000,
+	0x37070000,
+	0x37080000,
+	0x37090000,
+	0x370a0000,
+	0x370b0000,
+	0x370c0000,
+	0x370d0000,
+	0x370e0000,
+	0x370f0000,
+	0x37100000,
+	0x37110000,
+	0x37120000,
+	0x37130000,
+	0x37140000,
+	0x37150000,
+	0x37160000,
+	0x37170000,
+	0x37180000,
+	0x37190000,
+	0x371a0000,
+	0x371b0000,
+	0x371c0000,
+	0x371d0000,
+	0x371e0000,
+	0x371f0000,
+	0x37200000,
+	0x37210000,
+	0x37220000,
+	0x37230000,
+	0x37240000,
+	0x37250000,
+	0x37260000,
+	0x37270000,
+	0x37280000,
+	0x37290000,
+	0x372a0000,
+	0x372b0000,
+	0x372c0000,
+	0x372d0000,
+	0x372e0000,
+	0x372f0000,
+	0x37300000,
+	0x37310000,
+	0x37320000,
+	0x37330000,
+	0x37340000,
+	0x37350000,
+	0x37360000,
+	0x37370000,
+	0x37380000,
+	0x37390000,
+	0x373a0000,
+	0x373b0000,
+	0x373c0000,
+	0x373d0000,
+	0x373e0000,
+	0x373f0000,
+	0x37400000,
+	0x37410000,
+	0x37420000,
+	0x37430000,
+	0x37440000,
+	0x37450000,
+	0x37460000,
+	0x37470000,
+	0x37480000,
+	0x37490000,
+	0x374a0000,
+	0x374b0000,
+	0x374c0000,
+	0x374d0000,
+	0x374e0000,
+	0x374f0000,
+	0x37500000,
+	0x37510000,
+	0x37520000,
+	0x37530000,
+	0x37540000,
+	0x37550000,
+	0x37560000,
+	0x37570000,
+	0x37580000,
+	0x37590000,
+	0x375a0000,
+	0x375b0000,
+	0x375c0000,
+	0x375d0000,
+	0x375e0000,
+	0x375f0000,
+	0x37600000,
+	0x37610000,
+	0x37620000,
+	0x37630000,
+	0x37640000,
+	0x37650000,
+	0x37660000,
+	0x37670000,
+	0x37680000,
+	0x37690000,
+	0x376a0000,
+	0x376b0000,
+	0x376c0000,
+	0x376d0000,
+	0x376e0000,
+	0x376f0000,
+	0x37700000,
+	0x37710000,
+	0x37720000,
+	0x37730000,
+	0x37740000,
+	0x37750000,
+	0x37760000,
+	0x37770000,
+	0x37780000,
+	0x37790000,
+	0x377a0000,
+	0x377b0000,
+	0x377c0000,
+	0x377d0000,
+	0x377e0000,
+	0x377f0000,
+	0x37800000,
+	0x37808000,
+	0x37810000,
+	0x37818000,
+	0x37820000,
+	0x37828000,
+	0x37830000,
+	0x37838000,
+	0x37840000,
+	0x37848000,
+	0x37850000,
+	0x37858000,
+	0x37860000,
+	0x37868000,
+	0x37870000,
+	0x37878000,
+	0x37880000,
+	0x37888000,
+	0x37890000,
+	0x37898000,
+	0x378a0000,
+	0x378a8000,
+	0x378b0000,
+	0x378b8000,
+	0x378c0000,
+	0x378c8000,
+	0x378d0000,
+	0x378d8000,
+	0x378e0000,
+	0x378e8000,
+	0x378f0000,
+	0x378f8000,
+	0x37900000,
+	0x37908000,
+	0x37910000,
+	0x37918000,
+	0x37920000,
+	0x37928000,
+	0x37930000,
+	0x37938000,
+	0x37940000,
+	0x37948000,
+	0x37950000,
+	0x37958000,
+	0x37960000,
+	0x37968000,
+	0x37970000,
+	0x37978000,
+	0x37980000,
+	0x37988000,
+	0x37990000,
+	0x37998000,
+	0x379a0000,
+	0x379a8000,
+	0x379b0000,
+	0x379b8000,
+	0x379c0000,
+	0x379c8000,
+	0x379d0000,
+	0x379d8000,
+	0x379e0000,
+	0x379e8000,
+	0x379f0000,
+	0x379f8000,
+	0x37a00000,
+	0x37a08000,
+	0x37a10000,
+	0x37a18000,
+	0x37a20000,
+	0x37a28000,
+	0x37a30000,
+	0x37a38000,
+	0x37a40000,
+	0x37a48000,
+	0x37a50000,
+	0x37a58000,
+	0x37a60000,
+	0x37a68000,
+	0x37a70000,
+	0x37a78000,
+	0x37a80000,
+	0x37a88000,
+	0x37a90000,
+	0x37a98000,
+	0x37aa0000,
+	0x37aa8000,
+	0x37ab0000,
+	0x37ab8000,
+	0x37ac0000,
+	0x37ac8000,
+	0x37ad0000,
+	0x37ad8000,
+	0x37ae0000,
+	0x37ae8000,
+	0x37af0000,
+	0x37af8000,
+	0x37b00000,
+	0x37b08000,
+	0x37b10000,
+	0x37b18000,
+	0x37b20000,
+	0x37b28000,
+	0x37b30000,
+	0x37b38000,
+	0x37b40000,
+	0x37b48000,
+	0x37b50000,
+	0x37b58000,
+	0x37b60000,
+	0x37b68000,
+	0x37b70000,
+	0x37b78000,
+	0x37b80000,
+	0x37b88000,
+	0x37b90000,
+	0x37b98000,
+	0x37ba0000,
+	0x37ba8000,
+	0x37bb0000,
+	0x37bb8000,
+	0x37bc0000,
+	0x37bc8000,
+	0x37bd0000,
+	0x37bd8000,
+	0x37be0000,
+	0x37be8000,
+	0x37bf0000,
+	0x37bf8000,
+	0x37c00000,
+	0x37c08000,
+	0x37c10000,
+	0x37c18000,
+	0x37c20000,
+	0x37c28000,
+	0x37c30000,
+	0x37c38000,
+	0x37c40000,
+	0x37c48000,
+	0x37c50000,
+	0x37c58000,
+	0x37c60000,
+	0x37c68000,
+	0x37c70000,
+	0x37c78000,
+	0x37c80000,
+	0x37c88000,
+	0x37c90000,
+	0x37c98000,
+	0x37ca0000,
+	0x37ca8000,
+	0x37cb0000,
+	0x37cb8000,
+	0x37cc0000,
+	0x37cc8000,
+	0x37cd0000,
+	0x37cd8000,
+	0x37ce0000,
+	0x37ce8000,
+	0x37cf0000,
+	0x37cf8000,
+	0x37d00000,
+	0x37d08000,
+	0x37d10000,
+	0x37d18000,
+	0x37d20000,
+	0x37d28000,
+	0x37d30000,
+	0x37d38000,
+	0x37d40000,
+	0x37d48000,
+	0x37d50000,
+	0x37d58000,
+	0x37d60000,
+	0x37d68000,
+	0x37d70000,
+	0x37d78000,
+	0x37d80000,
+	0x37d88000,
+	0x37d90000,
+	0x37d98000,
+	0x37da0000,
+	0x37da8000,
+	0x37db0000,
+	0x37db8000,
+	0x37dc0000,
+	0x37dc8000,
+	0x37dd0000,
+	0x37dd8000,
+	0x37de0000,
+	0x37de8000,
+	0x37df0000,
+	0x37df8000,
+	0x37e00000,
+	0x37e08000,
+	0x37e10000,
+	0x37e18000,
+	0x37e20000,
+	0x37e28000,
+	0x37e30000,
+	0x37e38000,
+	0x37e40000,
+	0x37e48000,
+	0x37e50000,
+	0x37e58000,
+	0x37e60000,
+	0x37e68000,
+	0x37e70000,
+	0x37e78000,
+	0x37e80000,
+	0x37e88000,
+	0x37e90000,
+	0x37e98000,
+	0x37ea0000,
+	0x37ea8000,
+	0x37eb0000,
+	0x37eb8000,
+	0x37ec0000,
+	0x37ec8000,
+	0x37ed0000,
+	0x37ed8000,
+	0x37ee0000,
+	0x37ee8000,
+	0x37ef0000,
+	0x37ef8000,
+	0x37f00000,
+	0x37f08000,
+	0x37f10000,
+	0x37f18000,
+	0x37f20000,
+	0x37f28000,
+	0x37f30000,
+	0x37f38000,
+	0x37f40000,
+	0x37f48000,
+	0x37f50000,
+	0x37f58000,
+	0x37f60000,
+	0x37f68000,
+	0x37f70000,
+	0x37f78000,
+	0x37f80000,
+	0x37f88000,
+	0x37f90000,
+	0x37f98000,
+	0x37fa0000,
+	0x37fa8000,
+	0x37fb0000,
+	0x37fb8000,
+	0x37fc0000,
+	0x37fc8000,
+	0x37fd0000,
+	0x37fd8000,
+	0x37fe0000,
+	0x37fe8000,
+	0x37ff0000,
+	0x37ff8000,
+	0x38000000,
+	0x38004000,
+	0x38008000,
+	0x3800c000,
+	0x38010000,
+	0x38014000,
+	0x38018000,
+	0x3801c000,
+	0x38020000,
+	0x38024000,
+	0x38028000,
+	0x3802c000,
+	0x38030000,
+	0x38034000,
+	0x38038000,
+	0x3803c000,
+	0x38040000,
+	0x38044000,
+	0x38048000,
+	0x3804c000,
+	0x38050000,
+	0x38054000,
+	0x38058000,
+	0x3805c000,
+	0x38060000,
+	0x38064000,
+	0x38068000,
+	0x3806c000,
+	0x38070000,
+	0x38074000,
+	0x38078000,
+	0x3807c000,
+	0x38080000,
+	0x38084000,
+	0x38088000,
+	0x3808c000,
+	0x38090000,
+	0x38094000,
+	0x38098000,
+	0x3809c000,
+	0x380a0000,
+	0x380a4000,
+	0x380a8000,
+	0x380ac000,
+	0x380b0000,
+	0x380b4000,
+	0x380b8000,
+	0x380bc000,
+	0x380c0000,
+	0x380c4000,
+	0x380c8000,
+	0x380cc000,
+	0x380d0000,
+	0x380d4000,
+	0x380d8000,
+	0x380dc000,
+	0x380e0000,
+	0x380e4000,
+	0x380e8000,
+	0x380ec000,
+	0x380f0000,
+	0x380f4000,
+	0x380f8000,
+	0x380fc000,
+	0x38100000,
+	0x38104000,
+	0x38108000,
+	0x3810c000,
+	0x38110000,
+	0x38114000,
+	0x38118000,
+	0x3811c000,
+	0x38120000,
+	0x38124000,
+	0x38128000,
+	0x3812c000,
+	0x38130000,
+	0x38134000,
+	0x38138000,
+	0x3813c000,
+	0x38140000,
+	0x38144000,
+	0x38148000,
+	0x3814c000,
+	0x38150000,
+	0x38154000,
+	0x38158000,
+	0x3815c000,
+	0x38160000,
+	0x38164000,
+	0x38168000,
+	0x3816c000,
+	0x38170000,
+	0x38174000,
+	0x38178000,
+	0x3817c000,
+	0x38180000,
+	0x38184000,
+	0x38188000,
+	0x3818c000,
+	0x38190000,
+	0x38194000,
+	0x38198000,
+	0x3819c000,
+	0x381a0000,
+	0x381a4000,
+	0x381a8000,
+	0x381ac000,
+	0x381b0000,
+	0x381b4000,
+	0x381b8000,
+	0x381bc000,
+	0x381c0000,
+	0x381c4000,
+	0x381c8000,
+	0x381cc000,
+	0x381d0000,
+	0x381d4000,
+	0x381d8000,
+	0x381dc000,
+	0x381e0000,
+	0x381e4000,
+	0x381e8000,
+	0x381ec000,
+	0x381f0000,
+	0x381f4000,
+	0x381f8000,
+	0x381fc000,
+	0x38200000,
+	0x38204000,
+	0x38208000,
+	0x3820c000,
+	0x38210000,
+	0x38214000,
+	0x38218000,
+	0x3821c000,
+	0x38220000,
+	0x38224000,
+	0x38228000,
+	0x3822c000,
+	0x38230000,
+	0x38234000,
+	0x38238000,
+	0x3823c000,
+	0x38240000,
+	0x38244000,
+	0x38248000,
+	0x3824c000,
+	0x38250000,
+	0x38254000,
+	0x38258000,
+	0x3825c000,
+	0x38260000,
+	0x38264000,
+	0x38268000,
+	0x3826c000,
+	0x38270000,
+	0x38274000,
+	0x38278000,
+	0x3827c000,
+	0x38280000,
+	0x38284000,
+	0x38288000,
+	0x3828c000,
+	0x38290000,
+	0x38294000,
+	0x38298000,
+	0x3829c000,
+	0x382a0000,
+	0x382a4000,
+	0x382a8000,
+	0x382ac000,
+	0x382b0000,
+	0x382b4000,
+	0x382b8000,
+	0x382bc000,
+	0x382c0000,
+	0x382c4000,
+	0x382c8000,
+	0x382cc000,
+	0x382d0000,
+	0x382d4000,
+	0x382d8000,
+	0x382dc000,
+	0x382e0000,
+	0x382e4000,
+	0x382e8000,
+	0x382ec000,
+	0x382f0000,
+	0x382f4000,
+	0x382f8000,
+	0x382fc000,
+	0x38300000,
+	0x38304000,
+	0x38308000,
+	0x3830c000,
+	0x38310000,
+	0x38314000,
+	0x38318000,
+	0x3831c000,
+	0x38320000,
+	0x38324000,
+	0x38328000,
+	0x3832c000,
+	0x38330000,
+	0x38334000,
+	0x38338000,
+	0x3833c000,
+	0x38340000,
+	0x38344000,
+	0x38348000,
+	0x3834c000,
+	0x38350000,
+	0x38354000,
+	0x38358000,
+	0x3835c000,
+	0x38360000,
+	0x38364000,
+	0x38368000,
+	0x3836c000,
+	0x38370000,
+	0x38374000,
+	0x38378000,
+	0x3837c000,
+	0x38380000,
+	0x38384000,
+	0x38388000,
+	0x3838c000,
+	0x38390000,
+	0x38394000,
+	0x38398000,
+	0x3839c000,
+	0x383a0000,
+	0x383a4000,
+	0x383a8000,
+	0x383ac000,
+	0x383b0000,
+	0x383b4000,
+	0x383b8000,
+	0x383bc000,
+	0x383c0000,
+	0x383c4000,
+	0x383c8000,
+	0x383cc000,
+	0x383d0000,
+	0x383d4000,
+	0x383d8000,
+	0x383dc000,
+	0x383e0000,
+	0x383e4000,
+	0x383e8000,
+	0x383ec000,
+	0x383f0000,
+	0x383f4000,
+	0x383f8000,
+	0x383fc000,
+	0x38400000,
+	0x38404000,
+	0x38408000,
+	0x3840c000,
+	0x38410000,
+	0x38414000,
+	0x38418000,
+	0x3841c000,
+	0x38420000,
+	0x38424000,
+	0x38428000,
+	0x3842c000,
+	0x38430000,
+	0x38434000,
+	0x38438000,
+	0x3843c000,
+	0x38440000,
+	0x38444000,
+	0x38448000,
+	0x3844c000,
+	0x38450000,
+	0x38454000,
+	0x38458000,
+	0x3845c000,
+	0x38460000,
+	0x38464000,
+	0x38468000,
+	0x3846c000,
+	0x38470000,
+	0x38474000,
+	0x38478000,
+	0x3847c000,
+	0x38480000,
+	0x38484000,
+	0x38488000,
+	0x3848c000,
+	0x38490000,
+	0x38494000,
+	0x38498000,
+	0x3849c000,
+	0x384a0000,
+	0x384a4000,
+	0x384a8000,
+	0x384ac000,
+	0x384b0000,
+	0x384b4000,
+	0x384b8000,
+	0x384bc000,
+	0x384c0000,
+	0x384c4000,
+	0x384c8000,
+	0x384cc000,
+	0x384d0000,
+	0x384d4000,
+	0x384d8000,
+	0x384dc000,
+	0x384e0000,
+	0x384e4000,
+	0x384e8000,
+	0x384ec000,
+	0x384f0000,
+	0x384f4000,
+	0x384f8000,
+	0x384fc000,
+	0x38500000,
+	0x38504000,
+	0x38508000,
+	0x3850c000,
+	0x38510000,
+	0x38514000,
+	0x38518000,
+	0x3851c000,
+	0x38520000,
+	0x38524000,
+	0x38528000,
+	0x3852c000,
+	0x38530000,
+	0x38534000,
+	0x38538000,
+	0x3853c000,
+	0x38540000,
+	0x38544000,
+	0x38548000,
+	0x3854c000,
+	0x38550000,
+	0x38554000,
+	0x38558000,
+	0x3855c000,
+	0x38560000,
+	0x38564000,
+	0x38568000,
+	0x3856c000,
+	0x38570000,
+	0x38574000,
+	0x38578000,
+	0x3857c000,
+	0x38580000,
+	0x38584000,
+	0x38588000,
+	0x3858c000,
+	0x38590000,
+	0x38594000,
+	0x38598000,
+	0x3859c000,
+	0x385a0000,
+	0x385a4000,
+	0x385a8000,
+	0x385ac000,
+	0x385b0000,
+	0x385b4000,
+	0x385b8000,
+	0x385bc000,
+	0x385c0000,
+	0x385c4000,
+	0x385c8000,
+	0x385cc000,
+	0x385d0000,
+	0x385d4000,
+	0x385d8000,
+	0x385dc000,
+	0x385e0000,
+	0x385e4000,
+	0x385e8000,
+	0x385ec000,
+	0x385f0000,
+	0x385f4000,
+	0x385f8000,
+	0x385fc000,
+	0x38600000,
+	0x38604000,
+	0x38608000,
+	0x3860c000,
+	0x38610000,
+	0x38614000,
+	0x38618000,
+	0x3861c000,
+	0x38620000,
+	0x38624000,
+	0x38628000,
+	0x3862c000,
+	0x38630000,
+	0x38634000,
+	0x38638000,
+	0x3863c000,
+	0x38640000,
+	0x38644000,
+	0x38648000,
+	0x3864c000,
+	0x38650000,
+	0x38654000,
+	0x38658000,
+	0x3865c000,
+	0x38660000,
+	0x38664000,
+	0x38668000,
+	0x3866c000,
+	0x38670000,
+	0x38674000,
+	0x38678000,
+	0x3867c000,
+	0x38680000,
+	0x38684000,
+	0x38688000,
+	0x3868c000,
+	0x38690000,
+	0x38694000,
+	0x38698000,
+	0x3869c000,
+	0x386a0000,
+	0x386a4000,
+	0x386a8000,
+	0x386ac000,
+	0x386b0000,
+	0x386b4000,
+	0x386b8000,
+	0x386bc000,
+	0x386c0000,
+	0x386c4000,
+	0x386c8000,
+	0x386cc000,
+	0x386d0000,
+	0x386d4000,
+	0x386d8000,
+	0x386dc000,
+	0x386e0000,
+	0x386e4000,
+	0x386e8000,
+	0x386ec000,
+	0x386f0000,
+	0x386f4000,
+	0x386f8000,
+	0x386fc000,
+	0x38700000,
+	0x38704000,
+	0x38708000,
+	0x3870c000,
+	0x38710000,
+	0x38714000,
+	0x38718000,
+	0x3871c000,
+	0x38720000,
+	0x38724000,
+	0x38728000,
+	0x3872c000,
+	0x38730000,
+	0x38734000,
+	0x38738000,
+	0x3873c000,
+	0x38740000,
+	0x38744000,
+	0x38748000,
+	0x3874c000,
+	0x38750000,
+	0x38754000,
+	0x38758000,
+	0x3875c000,
+	0x38760000,
+	0x38764000,
+	0x38768000,
+	0x3876c000,
+	0x38770000,
+	0x38774000,
+	0x38778000,
+	0x3877c000,
+	0x38780000,
+	0x38784000,
+	0x38788000,
+	0x3878c000,
+	0x38790000,
+	0x38794000,
+	0x38798000,
+	0x3879c000,
+	0x387a0000,
+	0x387a4000,
+	0x387a8000,
+	0x387ac000,
+	0x387b0000,
+	0x387b4000,
+	0x387b8000,
+	0x387bc000,
+	0x387c0000,
+	0x387c4000,
+	0x387c8000,
+	0x387cc000,
+	0x387d0000,
+	0x387d4000,
+	0x387d8000,
+	0x387dc000,
+	0x387e0000,
+	0x387e4000,
+	0x387e8000,
+	0x387ec000,
+	0x387f0000,
+	0x387f4000,
+	0x387f8000,
+	0x387fc000,
+	0x0,
+	0x2000,
+	0x4000,
+	0x6000,
+	0x8000,
+	0xa000,
+	0xc000,
+	0xe000,
+	0x10000,
+	0x12000,
+	0x14000,
+	0x16000,
+	0x18000,
+	0x1a000,
+	0x1c000,
+	0x1e000,
+	0x20000,
+	0x22000,
+	0x24000,
+	0x26000,
+	0x28000,
+	0x2a000,
+	0x2c000,
+	0x2e000,
+	0x30000,
+	0x32000,
+	0x34000,
+	0x36000,
+	0x38000,
+	0x3a000,
+	0x3c000,
+	0x3e000,
+	0x40000,
+	0x42000,
+	0x44000,
+	0x46000,
+	0x48000,
+	0x4a000,
+	0x4c000,
+	0x4e000,
+	0x50000,
+	0x52000,
+	0x54000,
+	0x56000,
+	0x58000,
+	0x5a000,
+	0x5c000,
+	0x5e000,
+	0x60000,
+	0x62000,
+	0x64000,
+	0x66000,
+	0x68000,
+	0x6a000,
+	0x6c000,
+	0x6e000,
+	0x70000,
+	0x72000,
+	0x74000,
+	0x76000,
+	0x78000,
+	0x7a000,
+	0x7c000,
+	0x7e000,
+	0x80000,
+	0x82000,
+	0x84000,
+	0x86000,
+	0x88000,
+	0x8a000,
+	0x8c000,
+	0x8e000,
+	0x90000,
+	0x92000,
+	0x94000,
+	0x96000,
+	0x98000,
+	0x9a000,
+	0x9c000,
+	0x9e000,
+	0xa0000,
+	0xa2000,
+	0xa4000,
+	0xa6000,
+	0xa8000,
+	0xaa000,
+	0xac000,
+	0xae000,
+	0xb0000,
+	0xb2000,
+	0xb4000,
+	0xb6000,
+	0xb8000,
+	0xba000,
+	0xbc000,
+	0xbe000,
+	0xc0000,
+	0xc2000,
+	0xc4000,
+	0xc6000,
+	0xc8000,
+	0xca000,
+	0xcc000,
+	0xce000,
+	0xd0000,
+	0xd2000,
+	0xd4000,
+	0xd6000,
+	0xd8000,
+	0xda000,
+	0xdc000,
+	0xde000,
+	0xe0000,
+	0xe2000,
+	0xe4000,
+	0xe6000,
+	0xe8000,
+	0xea000,
+	0xec000,
+	0xee000,
+	0xf0000,
+	0xf2000,
+	0xf4000,
+	0xf6000,
+	0xf8000,
+	0xfa000,
+	0xfc000,
+	0xfe000,
+	0x100000,
+	0x102000,
+	0x104000,
+	0x106000,
+	0x108000,
+	0x10a000,
+	0x10c000,
+	0x10e000,
+	0x110000,
+	0x112000,
+	0x114000,
+	0x116000,
+	0x118000,
+	0x11a000,
+	0x11c000,
+	0x11e000,
+	0x120000,
+	0x122000,
+	0x124000,
+	0x126000,
+	0x128000,
+	0x12a000,
+	0x12c000,
+	0x12e000,
+	0x130000,
+	0x132000,
+	0x134000,
+	0x136000,
+	0x138000,
+	0x13a000,
+	0x13c000,
+	0x13e000,
+	0x140000,
+	0x142000,
+	0x144000,
+	0x146000,
+	0x148000,
+	0x14a000,
+	0x14c000,
+	0x14e000,
+	0x150000,
+	0x152000,
+	0x154000,
+	0x156000,
+	0x158000,
+	0x15a000,
+	0x15c000,
+	0x15e000,
+	0x160000,
+	0x162000,
+	0x164000,
+	0x166000,
+	0x168000,
+	0x16a000,
+	0x16c000,
+	0x16e000,
+	0x170000,
+	0x172000,
+	0x174000,
+	0x176000,
+	0x178000,
+	0x17a000,
+	0x17c000,
+	0x17e000,
+	0x180000,
+	0x182000,
+	0x184000,
+	0x186000,
+	0x188000,
+	0x18a000,
+	0x18c000,
+	0x18e000,
+	0x190000,
+	0x192000,
+	0x194000,
+	0x196000,
+	0x198000,
+	0x19a000,
+	0x19c000,
+	0x19e000,
+	0x1a0000,
+	0x1a2000,
+	0x1a4000,
+	0x1a6000,
+	0x1a8000,
+	0x1aa000,
+	0x1ac000,
+	0x1ae000,
+	0x1b0000,
+	0x1b2000,
+	0x1b4000,
+	0x1b6000,
+	0x1b8000,
+	0x1ba000,
+	0x1bc000,
+	0x1be000,
+	0x1c0000,
+	0x1c2000,
+	0x1c4000,
+	0x1c6000,
+	0x1c8000,
+	0x1ca000,
+	0x1cc000,
+	0x1ce000,
+	0x1d0000,
+	0x1d2000,
+	0x1d4000,
+	0x1d6000,
+	0x1d8000,
+	0x1da000,
+	0x1dc000,
+	0x1de000,
+	0x1e0000,
+	0x1e2000,
+	0x1e4000,
+	0x1e6000,
+	0x1e8000,
+	0x1ea000,
+	0x1ec000,
+	0x1ee000,
+	0x1f0000,
+	0x1f2000,
+	0x1f4000,
+	0x1f6000,
+	0x1f8000,
+	0x1fa000,
+	0x1fc000,
+	0x1fe000,
+	0x200000,
+	0x202000,
+	0x204000,
+	0x206000,
+	0x208000,
+	0x20a000,
+	0x20c000,
+	0x20e000,
+	0x210000,
+	0x212000,
+	0x214000,
+	0x216000,
+	0x218000,
+	0x21a000,
+	0x21c000,
+	0x21e000,
+	0x220000,
+	0x222000,
+	0x224000,
+	0x226000,
+	0x228000,
+	0x22a000,
+	0x22c000,
+	0x22e000,
+	0x230000,
+	0x232000,
+	0x234000,
+	0x236000,
+	0x238000,
+	0x23a000,
+	0x23c000,
+	0x23e000,
+	0x240000,
+	0x242000,
+	0x244000,
+	0x246000,
+	0x248000,
+	0x24a000,
+	0x24c000,
+	0x24e000,
+	0x250000,
+	0x252000,
+	0x254000,
+	0x256000,
+	0x258000,
+	0x25a000,
+	0x25c000,
+	0x25e000,
+	0x260000,
+	0x262000,
+	0x264000,
+	0x266000,
+	0x268000,
+	0x26a000,
+	0x26c000,
+	0x26e000,
+	0x270000,
+	0x272000,
+	0x274000,
+	0x276000,
+	0x278000,
+	0x27a000,
+	0x27c000,
+	0x27e000,
+	0x280000,
+	0x282000,
+	0x284000,
+	0x286000,
+	0x288000,
+	0x28a000,
+	0x28c000,
+	0x28e000,
+	0x290000,
+	0x292000,
+	0x294000,
+	0x296000,
+	0x298000,
+	0x29a000,
+	0x29c000,
+	0x29e000,
+	0x2a0000,
+	0x2a2000,
+	0x2a4000,
+	0x2a6000,
+	0x2a8000,
+	0x2aa000,
+	0x2ac000,
+	0x2ae000,
+	0x2b0000,
+	0x2b2000,
+	0x2b4000,
+	0x2b6000,
+	0x2b8000,
+	0x2ba000,
+	0x2bc000,
+	0x2be000,
+	0x2c0000,
+	0x2c2000,
+	0x2c4000,
+	0x2c6000,
+	0x2c8000,
+	0x2ca000,
+	0x2cc000,
+	0x2ce000,
+	0x2d0000,
+	0x2d2000,
+	0x2d4000,
+	0x2d6000,
+	0x2d8000,
+	0x2da000,
+	0x2dc000,
+	0x2de000,
+	0x2e0000,
+	0x2e2000,
+	0x2e4000,
+	0x2e6000,
+	0x2e8000,
+	0x2ea000,
+	0x2ec000,
+	0x2ee000,
+	0x2f0000,
+	0x2f2000,
+	0x2f4000,
+	0x2f6000,
+	0x2f8000,
+	0x2fa000,
+	0x2fc000,
+	0x2fe000,
+	0x300000,
+	0x302000,
+	0x304000,
+	0x306000,
+	0x308000,
+	0x30a000,
+	0x30c000,
+	0x30e000,
+	0x310000,
+	0x312000,
+	0x314000,
+	0x316000,
+	0x318000,
+	0x31a000,
+	0x31c000,
+	0x31e000,
+	0x320000,
+	0x322000,
+	0x324000,
+	0x326000,
+	0x328000,
+	0x32a000,
+	0x32c000,
+	0x32e000,
+	0x330000,
+	0x332000,
+	0x334000,
+	0x336000,
+	0x338000,
+	0x33a000,
+	0x33c000,
+	0x33e000,
+	0x340000,
+	0x342000,
+	0x344000,
+	0x346000,
+	0x348000,
+	0x34a000,
+	0x34c000,
+	0x34e000,
+	0x350000,
+	0x352000,
+	0x354000,
+	0x356000,
+	0x358000,
+	0x35a000,
+	0x35c000,
+	0x35e000,
+	0x360000,
+	0x362000,
+	0x364000,
+	0x366000,
+	0x368000,
+	0x36a000,
+	0x36c000,
+	0x36e000,
+	0x370000,
+	0x372000,
+	0x374000,
+	0x376000,
+	0x378000,
+	0x37a000,
+	0x37c000,
+	0x37e000,
+	0x380000,
+	0x382000,
+	0x384000,
+	0x386000,
+	0x388000,
+	0x38a000,
+	0x38c000,
+	0x38e000,
+	0x390000,
+	0x392000,
+	0x394000,
+	0x396000,
+	0x398000,
+	0x39a000,
+	0x39c000,
+	0x39e000,
+	0x3a0000,
+	0x3a2000,
+	0x3a4000,
+	0x3a6000,
+	0x3a8000,
+	0x3aa000,
+	0x3ac000,
+	0x3ae000,
+	0x3b0000,
+	0x3b2000,
+	0x3b4000,
+	0x3b6000,
+	0x3b8000,
+	0x3ba000,
+	0x3bc000,
+	0x3be000,
+	0x3c0000,
+	0x3c2000,
+	0x3c4000,
+	0x3c6000,
+	0x3c8000,
+	0x3ca000,
+	0x3cc000,
+	0x3ce000,
+	0x3d0000,
+	0x3d2000,
+	0x3d4000,
+	0x3d6000,
+	0x3d8000,
+	0x3da000,
+	0x3dc000,
+	0x3de000,
+	0x3e0000,
+	0x3e2000,
+	0x3e4000,
+	0x3e6000,
+	0x3e8000,
+	0x3ea000,
+	0x3ec000,
+	0x3ee000,
+	0x3f0000,
+	0x3f2000,
+	0x3f4000,
+	0x3f6000,
+	0x3f8000,
+	0x3fa000,
+	0x3fc000,
+	0x3fe000,
+	0x400000,
+	0x402000,
+	0x404000,
+	0x406000,
+	0x408000,
+	0x40a000,
+	0x40c000,
+	0x40e000,
+	0x410000,
+	0x412000,
+	0x414000,
+	0x416000,
+	0x418000,
+	0x41a000,
+	0x41c000,
+	0x41e000,
+	0x420000,
+	0x422000,
+	0x424000,
+	0x426000,
+	0x428000,
+	0x42a000,
+	0x42c000,
+	0x42e000,
+	0x430000,
+	0x432000,
+	0x434000,
+	0x436000,
+	0x438000,
+	0x43a000,
+	0x43c000,
+	0x43e000,
+	0x440000,
+	0x442000,
+	0x444000,
+	0x446000,
+	0x448000,
+	0x44a000,
+	0x44c000,
+	0x44e000,
+	0x450000,
+	0x452000,
+	0x454000,
+	0x456000,
+	0x458000,
+	0x45a000,
+	0x45c000,
+	0x45e000,
+	0x460000,
+	0x462000,
+	0x464000,
+	0x466000,
+	0x468000,
+	0x46a000,
+	0x46c000,
+	0x46e000,
+	0x470000,
+	0x472000,
+	0x474000,
+	0x476000,
+	0x478000,
+	0x47a000,
+	0x47c000,
+	0x47e000,
+	0x480000,
+	0x482000,
+	0x484000,
+	0x486000,
+	0x488000,
+	0x48a000,
+	0x48c000,
+	0x48e000,
+	0x490000,
+	0x492000,
+	0x494000,
+	0x496000,
+	0x498000,
+	0x49a000,
+	0x49c000,
+	0x49e000,
+	0x4a0000,
+	0x4a2000,
+	0x4a4000,
+	0x4a6000,
+	0x4a8000,
+	0x4aa000,
+	0x4ac000,
+	0x4ae000,
+	0x4b0000,
+	0x4b2000,
+	0x4b4000,
+	0x4b6000,
+	0x4b8000,
+	0x4ba000,
+	0x4bc000,
+	0x4be000,
+	0x4c0000,
+	0x4c2000,
+	0x4c4000,
+	0x4c6000,
+	0x4c8000,
+	0x4ca000,
+	0x4cc000,
+	0x4ce000,
+	0x4d0000,
+	0x4d2000,
+	0x4d4000,
+	0x4d6000,
+	0x4d8000,
+	0x4da000,
+	0x4dc000,
+	0x4de000,
+	0x4e0000,
+	0x4e2000,
+	0x4e4000,
+	0x4e6000,
+	0x4e8000,
+	0x4ea000,
+	0x4ec000,
+	0x4ee000,
+	0x4f0000,
+	0x4f2000,
+	0x4f4000,
+	0x4f6000,
+	0x4f8000,
+	0x4fa000,
+	0x4fc000,
+	0x4fe000,
+	0x500000,
+	0x502000,
+	0x504000,
+	0x506000,
+	0x508000,
+	0x50a000,
+	0x50c000,
+	0x50e000,
+	0x510000,
+	0x512000,
+	0x514000,
+	0x516000,
+	0x518000,
+	0x51a000,
+	0x51c000,
+	0x51e000,
+	0x520000,
+	0x522000,
+	0x524000,
+	0x526000,
+	0x528000,
+	0x52a000,
+	0x52c000,
+	0x52e000,
+	0x530000,
+	0x532000,
+	0x534000,
+	0x536000,
+	0x538000,
+	0x53a000,
+	0x53c000,
+	0x53e000,
+	0x540000,
+	0x542000,
+	0x544000,
+	0x546000,
+	0x548000,
+	0x54a000,
+	0x54c000,
+	0x54e000,
+	0x550000,
+	0x552000,
+	0x554000,
+	0x556000,
+	0x558000,
+	0x55a000,
+	0x55c000,
+	0x55e000,
+	0x560000,
+	0x562000,
+	0x564000,
+	0x566000,
+	0x568000,
+	0x56a000,
+	0x56c000,
+	0x56e000,
+	0x570000,
+	0x572000,
+	0x574000,
+	0x576000,
+	0x578000,
+	0x57a000,
+	0x57c000,
+	0x57e000,
+	0x580000,
+	0x582000,
+	0x584000,
+	0x586000,
+	0x588000,
+	0x58a000,
+	0x58c000,
+	0x58e000,
+	0x590000,
+	0x592000,
+	0x594000,
+	0x596000,
+	0x598000,
+	0x59a000,
+	0x59c000,
+	0x59e000,
+	0x5a0000,
+	0x5a2000,
+	0x5a4000,
+	0x5a6000,
+	0x5a8000,
+	0x5aa000,
+	0x5ac000,
+	0x5ae000,
+	0x5b0000,
+	0x5b2000,
+	0x5b4000,
+	0x5b6000,
+	0x5b8000,
+	0x5ba000,
+	0x5bc000,
+	0x5be000,
+	0x5c0000,
+	0x5c2000,
+	0x5c4000,
+	0x5c6000,
+	0x5c8000,
+	0x5ca000,
+	0x5cc000,
+	0x5ce000,
+	0x5d0000,
+	0x5d2000,
+	0x5d4000,
+	0x5d6000,
+	0x5d8000,
+	0x5da000,
+	0x5dc000,
+	0x5de000,
+	0x5e0000,
+	0x5e2000,
+	0x5e4000,
+	0x5e6000,
+	0x5e8000,
+	0x5ea000,
+	0x5ec000,
+	0x5ee000,
+	0x5f0000,
+	0x5f2000,
+	0x5f4000,
+	0x5f6000,
+	0x5f8000,
+	0x5fa000,
+	0x5fc000,
+	0x5fe000,
+	0x600000,
+	0x602000,
+	0x604000,
+	0x606000,
+	0x608000,
+	0x60a000,
+	0x60c000,
+	0x60e000,
+	0x610000,
+	0x612000,
+	0x614000,
+	0x616000,
+	0x618000,
+	0x61a000,
+	0x61c000,
+	0x61e000,
+	0x620000,
+	0x622000,
+	0x624000,
+	0x626000,
+	0x628000,
+	0x62a000,
+	0x62c000,
+	0x62e000,
+	0x630000,
+	0x632000,
+	0x634000,
+	0x636000,
+	0x638000,
+	0x63a000,
+	0x63c000,
+	0x63e000,
+	0x640000,
+	0x642000,
+	0x644000,
+	0x646000,
+	0x648000,
+	0x64a000,
+	0x64c000,
+	0x64e000,
+	0x650000,
+	0x652000,
+	0x654000,
+	0x656000,
+	0x658000,
+	0x65a000,
+	0x65c000,
+	0x65e000,
+	0x660000,
+	0x662000,
+	0x664000,
+	0x666000,
+	0x668000,
+	0x66a000,
+	0x66c000,
+	0x66e000,
+	0x670000,
+	0x672000,
+	0x674000,
+	0x676000,
+	0x678000,
+	0x67a000,
+	0x67c000,
+	0x67e000,
+	0x680000,
+	0x682000,
+	0x684000,
+	0x686000,
+	0x688000,
+	0x68a000,
+	0x68c000,
+	0x68e000,
+	0x690000,
+	0x692000,
+	0x694000,
+	0x696000,
+	0x698000,
+	0x69a000,
+	0x69c000,
+	0x69e000,
+	0x6a0000,
+	0x6a2000,
+	0x6a4000,
+	0x6a6000,
+	0x6a8000,
+	0x6aa000,
+	0x6ac000,
+	0x6ae000,
+	0x6b0000,
+	0x6b2000,
+	0x6b4000,
+	0x6b6000,
+	0x6b8000,
+	0x6ba000,
+	0x6bc000,
+	0x6be000,
+	0x6c0000,
+	0x6c2000,
+	0x6c4000,
+	0x6c6000,
+	0x6c8000,
+	0x6ca000,
+	0x6cc000,
+	0x6ce000,
+	0x6d0000,
+	0x6d2000,
+	0x6d4000,
+	0x6d6000,
+	0x6d8000,
+	0x6da000,
+	0x6dc000,
+	0x6de000,
+	0x6e0000,
+	0x6e2000,
+	0x6e4000,
+	0x6e6000,
+	0x6e8000,
+	0x6ea000,
+	0x6ec000,
+	0x6ee000,
+	0x6f0000,
+	0x6f2000,
+	0x6f4000,
+	0x6f6000,
+	0x6f8000,
+	0x6fa000,
+	0x6fc000,
+	0x6fe000,
+	0x700000,
+	0x702000,
+	0x704000,
+	0x706000,
+	0x708000,
+	0x70a000,
+	0x70c000,
+	0x70e000,
+	0x710000,
+	0x712000,
+	0x714000,
+	0x716000,
+	0x718000,
+	0x71a000,
+	0x71c000,
+	0x71e000,
+	0x720000,
+	0x722000,
+	0x724000,
+	0x726000,
+	0x728000,
+	0x72a000,
+	0x72c000,
+	0x72e000,
+	0x730000,
+	0x732000,
+	0x734000,
+	0x736000,
+	0x738000,
+	0x73a000,
+	0x73c000,
+	0x73e000,
+	0x740000,
+	0x742000,
+	0x744000,
+	0x746000,
+	0x748000,
+	0x74a000,
+	0x74c000,
+	0x74e000,
+	0x750000,
+	0x752000,
+	0x754000,
+	0x756000,
+	0x758000,
+	0x75a000,
+	0x75c000,
+	0x75e000,
+	0x760000,
+	0x762000,
+	0x764000,
+	0x766000,
+	0x768000,
+	0x76a000,
+	0x76c000,
+	0x76e000,
+	0x770000,
+	0x772000,
+	0x774000,
+	0x776000,
+	0x778000,
+	0x77a000,
+	0x77c000,
+	0x77e000,
+	0x780000,
+	0x782000,
+	0x784000,
+	0x786000,
+	0x788000,
+	0x78a000,
+	0x78c000,
+	0x78e000,
+	0x790000,
+	0x792000,
+	0x794000,
+	0x796000,
+	0x798000,
+	0x79a000,
+	0x79c000,
+	0x79e000,
+	0x7a0000,
+	0x7a2000,
+	0x7a4000,
+	0x7a6000,
+	0x7a8000,
+	0x7aa000,
+	0x7ac000,
+	0x7ae000,
+	0x7b0000,
+	0x7b2000,
+	0x7b4000,
+	0x7b6000,
+	0x7b8000,
+	0x7ba000,
+	0x7bc000,
+	0x7be000,
+	0x7c0000,
+	0x7c2000,
+	0x7c4000,
+	0x7c6000,
+	0x7c8000,
+	0x7ca000,
+	0x7cc000,
+	0x7ce000,
+	0x7d0000,
+	0x7d2000,
+	0x7d4000,
+	0x7d6000,
+	0x7d8000,
+	0x7da000,
+	0x7dc000,
+	0x7de000,
+	0x7e0000,
+	0x7e2000,
+	0x7e4000,
+	0x7e6000,
+	0x7e8000,
+	0x7ea000,
+	0x7ec000,
+	0x7ee000,
+	0x7f0000,
+	0x7f2000,
+	0x7f4000,
+	0x7f6000,
+	0x7f8000,
+	0x7fa000,
+	0x7fc000,
+	0x7fe000,
+};
+
+const uint32_t util_half_to_float_exponent_table[64] = {
+	0x0,
+	0x38800000,
+	0x39000000,
+	0x39800000,
+	0x3a000000,
+	0x3a800000,
+	0x3b000000,
+	0x3b800000,
+	0x3c000000,
+	0x3c800000,
+	0x3d000000,
+	0x3d800000,
+	0x3e000000,
+	0x3e800000,
+	0x3f000000,
+	0x3f800000,
+	0x40000000,
+	0x40800000,
+	0x41000000,
+	0x41800000,
+	0x42000000,
+	0x42800000,
+	0x43000000,
+	0x43800000,
+	0x44000000,
+	0x44800000,
+	0x45000000,
+	0x45800000,
+	0x46000000,
+	0x46800000,
+	0x47000000,
+	0x7f800000,
+	0x80000000,
+	0xb8800000,
+	0xb9000000,
+	0xb9800000,
+	0xba000000,
+	0xba800000,
+	0xbb000000,
+	0xbb800000,
+	0xbc000000,
+	0xbc800000,
+	0xbd000000,
+	0xbd800000,
+	0xbe000000,
+	0xbe800000,
+	0xbf000000,
+	0xbf800000,
+	0xc0000000,
+	0xc0800000,
+	0xc1000000,
+	0xc1800000,
+	0xc2000000,
+	0xc2800000,
+	0xc3000000,
+	0xc3800000,
+	0xc4000000,
+	0xc4800000,
+	0xc5000000,
+	0xc5800000,
+	0xc6000000,
+	0xc6800000,
+	0xc7000000,
+	0xff800000,
+};
+
+const uint32_t util_half_to_float_offset_table[64] = {
+	0x0,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x0,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+	0x400,
+};
+
+const uint16_t util_float_to_half_base_table[512] = {
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x0,
+	0x1,
+	0x2,
+	0x4,
+	0x8,
+	0x10,
+	0x20,
+	0x40,
+	0x80,
+	0x100,
+	0x200,
+	0x400,
+	0x800,
+	0xc00,
+	0x1000,
+	0x1400,
+	0x1800,
+	0x1c00,
+	0x2000,
+	0x2400,
+	0x2800,
+	0x2c00,
+	0x3000,
+	0x3400,
+	0x3800,
+	0x3c00,
+	0x4000,
+	0x4400,
+	0x4800,
+	0x4c00,
+	0x5000,
+	0x5400,
+	0x5800,
+	0x5c00,
+	0x6000,
+	0x6400,
+	0x6800,
+	0x6c00,
+	0x7000,
+	0x7400,
+	0x7800,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x7c00,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8000,
+	0x8001,
+	0x8002,
+	0x8004,
+	0x8008,
+	0x8010,
+	0x8020,
+	0x8040,
+	0x8080,
+	0x8100,
+	0x8200,
+	0x8400,
+	0x8800,
+	0x8c00,
+	0x9000,
+	0x9400,
+	0x9800,
+	0x9c00,
+	0xa000,
+	0xa400,
+	0xa800,
+	0xac00,
+	0xb000,
+	0xb400,
+	0xb800,
+	0xbc00,
+	0xc000,
+	0xc400,
+	0xc800,
+	0xcc00,
+	0xd000,
+	0xd400,
+	0xd800,
+	0xdc00,
+	0xe000,
+	0xe400,
+	0xe800,
+	0xec00,
+	0xf000,
+	0xf400,
+	0xf800,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+	0xfc00,
+};
+
+const uint8_t util_float_to_half_shift_table[512] = {
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x17,
+	0x16,
+	0x15,
+	0x14,
+	0x13,
+	0x12,
+	0x11,
+	0x10,
+	0xf,
+	0xe,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0xd,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x17,
+	0x16,
+	0x15,
+	0x14,
+	0x13,
+	0x12,
+	0x11,
+	0x10,
+	0xf,
+	0xe,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0xd,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0x18,
+	0xd,
+};
diff --git a/lib/libGL/gallium/libgallium/generated/u_indices_gen.c b/lib/libGL/gallium/libgallium/generated/u_indices_gen.c
new file mode 100644
index 000000000..3c981e5d7
--- /dev/null
+++ b/lib/libGL/gallium/libgallium/generated/u_indices_gen.c
@@ -0,0 +1,5129 @@
+/* File automatically generated by indices.py */
+
+/*
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
+ * VMWARE AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+
+
+/**
+ * @file
+ * Functions to translate and generate index lists
+ */
+
+#include "indices/u_indices.h"
+#include "indices/u_indices_priv.h"
+#include "pipe/p_compiler.h"
+#include "util/u_debug.h"
+#include "pipe/p_defines.h"
+#include "util/u_memory.h"
+
+
+static unsigned out_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return OUT_UINT;
+   case 2: return OUT_USHORT;
+   default: assert(0); return OUT_USHORT;
+   }
+}
+
+static unsigned in_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return IN_UINT;
+   case 2: return IN_USHORT;
+   case 1: return IN_UBYTE;
+   default: assert(0); return IN_UBYTE;
+   }
+}
+
+
+static u_translate_func translate[IN_COUNT][OUT_COUNT][PV_COUNT][PV_COUNT][PRIM_COUNT];
+static u_generate_func  generate[OUT_COUNT][PV_COUNT][PV_COUNT][PRIM_COUNT];
+
+
+
+static void generate_points_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)(i);
+   }
+}
+static void generate_lines_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)(i);
+      (out+i)[1] = (ushort)(i+1);
+   }
+}
+static void generate_linestrip_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1);
+   }
+}
+static void generate_lineloop_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1);
+   }
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(0);
+}
+static void generate_tris_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)(i);
+      (out+i)[1] = (ushort)(i+1);
+      (out+i)[2] = (ushort)(i+2);
+   }
+}
+static void generate_tristrip_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1+(i&1));
+      (out+j)[2] = (ushort)(i+2-(i&1));
+   }
+}
+static void generate_trifan_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j)[2] = (ushort)(i+2);
+   }
+}
+static void generate_quads_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)(i+0);
+      (out+j+0)[1] = (ushort)(i+1);
+      (out+j+0)[2] = (ushort)(i+3);
+      (out+j+3)[0] = (ushort)(i+1);
+      (out+j+3)[1] = (ushort)(i+2);
+      (out+j+3)[2] = (ushort)(i+3);
+   }
+}
+static void generate_quadstrip_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)(i+2);
+      (out+j+0)[1] = (ushort)(i+0);
+      (out+j+0)[2] = (ushort)(i+3);
+      (out+j+3)[0] = (ushort)(i+0);
+      (out+j+3)[1] = (ushort)(i+1);
+      (out+j+3)[2] = (ushort)(i+3);
+   }
+}
+static void generate_polygon_ushort_first2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j)[2] = (ushort)(i+2);
+   }
+}
+static void generate_points_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)(i);
+   }
+}
+static void generate_lines_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)(i+1);
+      (out+i)[1] = (ushort)(i);
+   }
+}
+static void generate_linestrip_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i+1);
+      (out+j)[1] = (ushort)(i);
+   }
+}
+static void generate_lineloop_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i+1);
+      (out+j)[1] = (ushort)(i);
+   }
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i);
+}
+static void generate_tris_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)(i+1);
+      (out+i)[1] = (ushort)(i+2);
+      (out+i)[2] = (ushort)(i);
+   }
+}
+static void generate_tristrip_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i+1+(i&1));
+      (out+j)[1] = (ushort)(i+2-(i&1));
+      (out+j)[2] = (ushort)(i);
+   }
+}
+static void generate_trifan_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i+1);
+      (out+j)[1] = (ushort)(i+2);
+      (out+j)[2] = (ushort)(0);
+   }
+}
+static void generate_quads_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)(i+1);
+      (out+j+0)[1] = (ushort)(i+3);
+      (out+j+0)[2] = (ushort)(i+0);
+      (out+j+3)[0] = (ushort)(i+2);
+      (out+j+3)[1] = (ushort)(i+3);
+      (out+j+3)[2] = (ushort)(i+1);
+   }
+}
+static void generate_quadstrip_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)(i+0);
+      (out+j+0)[1] = (ushort)(i+3);
+      (out+j+0)[2] = (ushort)(i+2);
+      (out+j+3)[0] = (ushort)(i+1);
+      (out+j+3)[1] = (ushort)(i+3);
+      (out+j+3)[2] = (ushort)(i+0);
+   }
+}
+static void generate_polygon_ushort_first2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i+1);
+      (out+j)[1] = (ushort)(i+2);
+      (out+j)[2] = (ushort)(0);
+   }
+}
+static void generate_points_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)(i);
+   }
+}
+static void generate_lines_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)(i+1);
+      (out+i)[1] = (ushort)(i);
+   }
+}
+static void generate_linestrip_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i+1);
+      (out+j)[1] = (ushort)(i);
+   }
+}
+static void generate_lineloop_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i+1);
+      (out+j)[1] = (ushort)(i);
+   }
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i);
+}
+static void generate_tris_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)(i+2);
+      (out+i)[1] = (ushort)(i);
+      (out+i)[2] = (ushort)(i+1);
+   }
+}
+static void generate_tristrip_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i+2);
+      (out+j)[1] = (ushort)(i+(i&1));
+      (out+j)[2] = (ushort)(i+1-(i&1));
+   }
+}
+static void generate_trifan_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i+2);
+      (out+j)[1] = (ushort)(0);
+      (out+j)[2] = (ushort)(i+1);
+   }
+}
+static void generate_quads_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)(i+3);
+      (out+j+0)[1] = (ushort)(i+0);
+      (out+j+0)[2] = (ushort)(i+1);
+      (out+j+3)[0] = (ushort)(i+3);
+      (out+j+3)[1] = (ushort)(i+1);
+      (out+j+3)[2] = (ushort)(i+2);
+   }
+}
+static void generate_quadstrip_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)(i+3);
+      (out+j+0)[1] = (ushort)(i+2);
+      (out+j+0)[2] = (ushort)(i+0);
+      (out+j+3)[0] = (ushort)(i+3);
+      (out+j+3)[1] = (ushort)(i+0);
+      (out+j+3)[2] = (ushort)(i+1);
+   }
+}
+static void generate_polygon_ushort_last2first(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j)[2] = (ushort)(i+2);
+   }
+}
+static void generate_points_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)(i);
+   }
+}
+static void generate_lines_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)(i);
+      (out+i)[1] = (ushort)(i+1);
+   }
+}
+static void generate_linestrip_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1);
+   }
+}
+static void generate_lineloop_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1);
+   }
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(0);
+}
+static void generate_tris_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)(i);
+      (out+i)[1] = (ushort)(i+1);
+      (out+i)[2] = (ushort)(i+2);
+   }
+}
+static void generate_tristrip_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i+(i&1));
+      (out+j)[1] = (ushort)(i+1-(i&1));
+      (out+j)[2] = (ushort)(i+2);
+   }
+}
+static void generate_trifan_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j)[2] = (ushort)(i+2);
+   }
+}
+static void generate_quads_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)(i+0);
+      (out+j+0)[1] = (ushort)(i+1);
+      (out+j+0)[2] = (ushort)(i+3);
+      (out+j+3)[0] = (ushort)(i+1);
+      (out+j+3)[1] = (ushort)(i+2);
+      (out+j+3)[2] = (ushort)(i+3);
+   }
+}
+static void generate_quadstrip_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)(i+2);
+      (out+j+0)[1] = (ushort)(i+0);
+      (out+j+0)[2] = (ushort)(i+3);
+      (out+j+3)[0] = (ushort)(i+0);
+      (out+j+3)[1] = (ushort)(i+1);
+      (out+j+3)[2] = (ushort)(i+3);
+   }
+}
+static void generate_polygon_ushort_last2last(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)(i+1);
+      (out+j)[1] = (ushort)(i+2);
+      (out+j)[2] = (ushort)(0);
+   }
+}
+static void generate_points_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)(i);
+   }
+}
+static void generate_lines_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)(i);
+      (out+i)[1] = (uint)(i+1);
+   }
+}
+static void generate_linestrip_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1);
+   }
+}
+static void generate_lineloop_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1);
+   }
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(0);
+}
+static void generate_tris_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)(i);
+      (out+i)[1] = (uint)(i+1);
+      (out+i)[2] = (uint)(i+2);
+   }
+}
+static void generate_tristrip_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1+(i&1));
+      (out+j)[2] = (uint)(i+2-(i&1));
+   }
+}
+static void generate_trifan_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      (out+j)[2] = (uint)(i+2);
+   }
+}
+static void generate_quads_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)(i+0);
+      (out+j+0)[1] = (uint)(i+1);
+      (out+j+0)[2] = (uint)(i+3);
+      (out+j+3)[0] = (uint)(i+1);
+      (out+j+3)[1] = (uint)(i+2);
+      (out+j+3)[2] = (uint)(i+3);
+   }
+}
+static void generate_quadstrip_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)(i+2);
+      (out+j+0)[1] = (uint)(i+0);
+      (out+j+0)[2] = (uint)(i+3);
+      (out+j+3)[0] = (uint)(i+0);
+      (out+j+3)[1] = (uint)(i+1);
+      (out+j+3)[2] = (uint)(i+3);
+   }
+}
+static void generate_polygon_uint_first2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      (out+j)[2] = (uint)(i+2);
+   }
+}
+static void generate_points_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)(i);
+   }
+}
+static void generate_lines_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)(i+1);
+      (out+i)[1] = (uint)(i);
+   }
+}
+static void generate_linestrip_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)(i+1);
+      (out+j)[1] = (uint)(i);
+   }
+}
+static void generate_lineloop_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)(i+1);
+      (out+j)[1] = (uint)(i);
+   }
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i);
+}
+static void generate_tris_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)(i+1);
+      (out+i)[1] = (uint)(i+2);
+      (out+i)[2] = (uint)(i);
+   }
+}
+static void generate_tristrip_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i+1+(i&1));
+      (out+j)[1] = (uint)(i+2-(i&1));
+      (out+j)[2] = (uint)(i);
+   }
+}
+static void generate_trifan_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i+1);
+      (out+j)[1] = (uint)(i+2);
+      (out+j)[2] = (uint)(0);
+   }
+}
+static void generate_quads_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)(i+1);
+      (out+j+0)[1] = (uint)(i+3);
+      (out+j+0)[2] = (uint)(i+0);
+      (out+j+3)[0] = (uint)(i+2);
+      (out+j+3)[1] = (uint)(i+3);
+      (out+j+3)[2] = (uint)(i+1);
+   }
+}
+static void generate_quadstrip_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)(i+0);
+      (out+j+0)[1] = (uint)(i+3);
+      (out+j+0)[2] = (uint)(i+2);
+      (out+j+3)[0] = (uint)(i+1);
+      (out+j+3)[1] = (uint)(i+3);
+      (out+j+3)[2] = (uint)(i+0);
+   }
+}
+static void generate_polygon_uint_first2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i+1);
+      (out+j)[1] = (uint)(i+2);
+      (out+j)[2] = (uint)(0);
+   }
+}
+static void generate_points_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)(i);
+   }
+}
+static void generate_lines_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)(i+1);
+      (out+i)[1] = (uint)(i);
+   }
+}
+static void generate_linestrip_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)(i+1);
+      (out+j)[1] = (uint)(i);
+   }
+}
+static void generate_lineloop_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)(i+1);
+      (out+j)[1] = (uint)(i);
+   }
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i);
+}
+static void generate_tris_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)(i+2);
+      (out+i)[1] = (uint)(i);
+      (out+i)[2] = (uint)(i+1);
+   }
+}
+static void generate_tristrip_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i+2);
+      (out+j)[1] = (uint)(i+(i&1));
+      (out+j)[2] = (uint)(i+1-(i&1));
+   }
+}
+static void generate_trifan_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i+2);
+      (out+j)[1] = (uint)(0);
+      (out+j)[2] = (uint)(i+1);
+   }
+}
+static void generate_quads_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)(i+3);
+      (out+j+0)[1] = (uint)(i+0);
+      (out+j+0)[2] = (uint)(i+1);
+      (out+j+3)[0] = (uint)(i+3);
+      (out+j+3)[1] = (uint)(i+1);
+      (out+j+3)[2] = (uint)(i+2);
+   }
+}
+static void generate_quadstrip_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)(i+3);
+      (out+j+0)[1] = (uint)(i+2);
+      (out+j+0)[2] = (uint)(i+0);
+      (out+j+3)[0] = (uint)(i+3);
+      (out+j+3)[1] = (uint)(i+0);
+      (out+j+3)[2] = (uint)(i+1);
+   }
+}
+static void generate_polygon_uint_last2first(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      (out+j)[2] = (uint)(i+2);
+   }
+}
+static void generate_points_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)(i);
+   }
+}
+static void generate_lines_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)(i);
+      (out+i)[1] = (uint)(i+1);
+   }
+}
+static void generate_linestrip_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1);
+   }
+}
+static void generate_lineloop_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1);
+   }
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(0);
+}
+static void generate_tris_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)(i);
+      (out+i)[1] = (uint)(i+1);
+      (out+i)[2] = (uint)(i+2);
+   }
+}
+static void generate_tristrip_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i+(i&1));
+      (out+j)[1] = (uint)(i+1-(i&1));
+      (out+j)[2] = (uint)(i+2);
+   }
+}
+static void generate_trifan_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      (out+j)[2] = (uint)(i+2);
+   }
+}
+static void generate_quads_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)(i+0);
+      (out+j+0)[1] = (uint)(i+1);
+      (out+j+0)[2] = (uint)(i+3);
+      (out+j+3)[0] = (uint)(i+1);
+      (out+j+3)[1] = (uint)(i+2);
+      (out+j+3)[2] = (uint)(i+3);
+   }
+}
+static void generate_quadstrip_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)(i+2);
+      (out+j+0)[1] = (uint)(i+0);
+      (out+j+0)[2] = (uint)(i+3);
+      (out+j+3)[0] = (uint)(i+0);
+      (out+j+3)[1] = (uint)(i+1);
+      (out+j+3)[2] = (uint)(i+3);
+   }
+}
+static void generate_polygon_uint_last2last(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)(i+1);
+      (out+j)[1] = (uint)(i+2);
+      (out+j)[2] = (uint)(0);
+   }
+}
+static void translate_points_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_linestrip_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_lineloop_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[0];
+}
+static void translate_tris_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+      (out+i)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_tristrip_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1+(i&1)];
+      (out+j)[2] = (ushort)in[i+2-(i&1)];
+   }
+}
+static void translate_trifan_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quads_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+1];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+2];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_quadstrip_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+2];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+0];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_polygon_ubyte2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_points_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i];
+   }
+}
+static void translate_linestrip_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+}
+static void translate_lineloop_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i];
+}
+static void translate_tris_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i+2];
+      (out+i)[2] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1+(i&1)];
+      (out+j)[1] = (ushort)in[i+2-(i&1)];
+      (out+j)[2] = (ushort)in[i];
+   }
+}
+static void translate_trifan_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_quads_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+1];
+      (out+j+0)[1] = (ushort)in[i+3];
+      (out+j+0)[2] = (ushort)in[i+0];
+      (out+j+3)[0] = (ushort)in[i+2];
+      (out+j+3)[1] = (ushort)in[i+3];
+      (out+j+3)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_quadstrip_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+3];
+      (out+j+0)[2] = (ushort)in[i+2];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+3];
+      (out+j+3)[2] = (ushort)in[i+0];
+   }
+}
+static void translate_polygon_ubyte2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_points_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i];
+   }
+}
+static void translate_linestrip_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+}
+static void translate_lineloop_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i];
+}
+static void translate_tris_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i+2];
+      (out+i)[1] = (ushort)in[i];
+      (out+i)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_tristrip_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+(i&1)];
+      (out+j)[2] = (ushort)in[i+1-(i&1)];
+   }
+}
+static void translate_trifan_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[0];
+      (out+j)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_quads_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+3];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+1];
+      (out+j+3)[0] = (ushort)in[i+3];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quadstrip_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+3];
+      (out+j+0)[1] = (ushort)in[i+2];
+      (out+j+0)[2] = (ushort)in[i+0];
+      (out+j+3)[0] = (ushort)in[i+3];
+      (out+j+3)[1] = (ushort)in[i+0];
+      (out+j+3)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_polygon_ubyte2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_points_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_linestrip_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_lineloop_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[0];
+}
+static void translate_tris_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+      (out+i)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_tristrip_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+(i&1)];
+      (out+j)[1] = (ushort)in[i+1-(i&1)];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_trifan_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quads_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+1];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+2];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_quadstrip_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+2];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+0];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_polygon_ubyte2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_points_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+   }
+}
+static void translate_linestrip_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+}
+static void translate_lineloop_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[0];
+}
+static void translate_tris_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+      (out+i)[2] = (uint)in[i+2];
+   }
+}
+static void translate_tristrip_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1+(i&1)];
+      (out+j)[2] = (uint)in[i+2-(i&1)];
+   }
+}
+static void translate_trifan_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quads_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+1];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+2];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_quadstrip_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+2];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+0];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_polygon_ubyte2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_points_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i];
+   }
+}
+static void translate_linestrip_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+}
+static void translate_lineloop_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i];
+}
+static void translate_tris_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i+2];
+      (out+i)[2] = (uint)in[i];
+   }
+}
+static void translate_tristrip_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1+(i&1)];
+      (out+j)[1] = (uint)in[i+2-(i&1)];
+      (out+j)[2] = (uint)in[i];
+   }
+}
+static void translate_trifan_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_quads_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+1];
+      (out+j+0)[1] = (uint)in[i+3];
+      (out+j+0)[2] = (uint)in[i+0];
+      (out+j+3)[0] = (uint)in[i+2];
+      (out+j+3)[1] = (uint)in[i+3];
+      (out+j+3)[2] = (uint)in[i+1];
+   }
+}
+static void translate_quadstrip_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+3];
+      (out+j+0)[2] = (uint)in[i+2];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+3];
+      (out+j+3)[2] = (uint)in[i+0];
+   }
+}
+static void translate_polygon_ubyte2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_points_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i];
+   }
+}
+static void translate_linestrip_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+}
+static void translate_lineloop_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i];
+}
+static void translate_tris_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i+2];
+      (out+i)[1] = (uint)in[i];
+      (out+i)[2] = (uint)in[i+1];
+   }
+}
+static void translate_tristrip_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+(i&1)];
+      (out+j)[2] = (uint)in[i+1-(i&1)];
+   }
+}
+static void translate_trifan_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[0];
+      (out+j)[2] = (uint)in[i+1];
+   }
+}
+static void translate_quads_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+3];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+1];
+      (out+j+3)[0] = (uint)in[i+3];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quadstrip_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+3];
+      (out+j+0)[1] = (uint)in[i+2];
+      (out+j+0)[2] = (uint)in[i+0];
+      (out+j+3)[0] = (uint)in[i+3];
+      (out+j+3)[1] = (uint)in[i+0];
+      (out+j+3)[2] = (uint)in[i+1];
+   }
+}
+static void translate_polygon_ubyte2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_points_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+   }
+}
+static void translate_linestrip_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+}
+static void translate_lineloop_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[0];
+}
+static void translate_tris_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+      (out+i)[2] = (uint)in[i+2];
+   }
+}
+static void translate_tristrip_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+(i&1)];
+      (out+j)[1] = (uint)in[i+1-(i&1)];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_trifan_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quads_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+1];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+2];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_quadstrip_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+2];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+0];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_polygon_ubyte2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_points_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_linestrip_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_lineloop_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[0];
+}
+static void translate_tris_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+      (out+i)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_tristrip_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1+(i&1)];
+      (out+j)[2] = (ushort)in[i+2-(i&1)];
+   }
+}
+static void translate_trifan_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quads_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+1];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+2];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_quadstrip_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+2];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+0];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_polygon_ushort2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_points_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i];
+   }
+}
+static void translate_linestrip_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+}
+static void translate_lineloop_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i];
+}
+static void translate_tris_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i+2];
+      (out+i)[2] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1+(i&1)];
+      (out+j)[1] = (ushort)in[i+2-(i&1)];
+      (out+j)[2] = (ushort)in[i];
+   }
+}
+static void translate_trifan_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_quads_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+1];
+      (out+j+0)[1] = (ushort)in[i+3];
+      (out+j+0)[2] = (ushort)in[i+0];
+      (out+j+3)[0] = (ushort)in[i+2];
+      (out+j+3)[1] = (ushort)in[i+3];
+      (out+j+3)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_quadstrip_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+3];
+      (out+j+0)[2] = (ushort)in[i+2];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+3];
+      (out+j+3)[2] = (ushort)in[i+0];
+   }
+}
+static void translate_polygon_ushort2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_points_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i];
+   }
+}
+static void translate_linestrip_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+}
+static void translate_lineloop_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i];
+}
+static void translate_tris_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i+2];
+      (out+i)[1] = (ushort)in[i];
+      (out+i)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_tristrip_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+(i&1)];
+      (out+j)[2] = (ushort)in[i+1-(i&1)];
+   }
+}
+static void translate_trifan_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[0];
+      (out+j)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_quads_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+3];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+1];
+      (out+j+3)[0] = (ushort)in[i+3];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quadstrip_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+3];
+      (out+j+0)[1] = (ushort)in[i+2];
+      (out+j+0)[2] = (ushort)in[i+0];
+      (out+j+3)[0] = (ushort)in[i+3];
+      (out+j+3)[1] = (ushort)in[i+0];
+      (out+j+3)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_polygon_ushort2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_points_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_linestrip_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_lineloop_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[0];
+}
+static void translate_tris_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+      (out+i)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_tristrip_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+(i&1)];
+      (out+j)[1] = (ushort)in[i+1-(i&1)];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_trifan_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quads_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+1];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+2];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_quadstrip_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+2];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+0];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_polygon_ushort2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_points_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+   }
+}
+static void translate_linestrip_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+}
+static void translate_lineloop_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[0];
+}
+static void translate_tris_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+      (out+i)[2] = (uint)in[i+2];
+   }
+}
+static void translate_tristrip_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1+(i&1)];
+      (out+j)[2] = (uint)in[i+2-(i&1)];
+   }
+}
+static void translate_trifan_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quads_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+1];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+2];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_quadstrip_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+2];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+0];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_polygon_ushort2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_points_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i];
+   }
+}
+static void translate_linestrip_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+}
+static void translate_lineloop_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i];
+}
+static void translate_tris_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i+2];
+      (out+i)[2] = (uint)in[i];
+   }
+}
+static void translate_tristrip_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1+(i&1)];
+      (out+j)[1] = (uint)in[i+2-(i&1)];
+      (out+j)[2] = (uint)in[i];
+   }
+}
+static void translate_trifan_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_quads_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+1];
+      (out+j+0)[1] = (uint)in[i+3];
+      (out+j+0)[2] = (uint)in[i+0];
+      (out+j+3)[0] = (uint)in[i+2];
+      (out+j+3)[1] = (uint)in[i+3];
+      (out+j+3)[2] = (uint)in[i+1];
+   }
+}
+static void translate_quadstrip_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+3];
+      (out+j+0)[2] = (uint)in[i+2];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+3];
+      (out+j+3)[2] = (uint)in[i+0];
+   }
+}
+static void translate_polygon_ushort2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_points_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i];
+   }
+}
+static void translate_linestrip_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+}
+static void translate_lineloop_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i];
+}
+static void translate_tris_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i+2];
+      (out+i)[1] = (uint)in[i];
+      (out+i)[2] = (uint)in[i+1];
+   }
+}
+static void translate_tristrip_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+(i&1)];
+      (out+j)[2] = (uint)in[i+1-(i&1)];
+   }
+}
+static void translate_trifan_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[0];
+      (out+j)[2] = (uint)in[i+1];
+   }
+}
+static void translate_quads_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+3];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+1];
+      (out+j+3)[0] = (uint)in[i+3];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quadstrip_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+3];
+      (out+j+0)[1] = (uint)in[i+2];
+      (out+j+0)[2] = (uint)in[i+0];
+      (out+j+3)[0] = (uint)in[i+3];
+      (out+j+3)[1] = (uint)in[i+0];
+      (out+j+3)[2] = (uint)in[i+1];
+   }
+}
+static void translate_polygon_ushort2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_points_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+   }
+}
+static void translate_linestrip_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+}
+static void translate_lineloop_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[0];
+}
+static void translate_tris_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+      (out+i)[2] = (uint)in[i+2];
+   }
+}
+static void translate_tristrip_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+(i&1)];
+      (out+j)[1] = (uint)in[i+1-(i&1)];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_trifan_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quads_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+1];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+2];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_quadstrip_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+2];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+0];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_polygon_ushort2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_points_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_linestrip_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_lineloop_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[0];
+}
+static void translate_tris_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+      (out+i)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_tristrip_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1+(i&1)];
+      (out+j)[2] = (ushort)in[i+2-(i&1)];
+   }
+}
+static void translate_trifan_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quads_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+1];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+2];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_quadstrip_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+2];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+0];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_polygon_uint2ushort_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_points_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i];
+   }
+}
+static void translate_linestrip_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+}
+static void translate_lineloop_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i];
+}
+static void translate_tris_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i+2];
+      (out+i)[2] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1+(i&1)];
+      (out+j)[1] = (ushort)in[i+2-(i&1)];
+      (out+j)[2] = (ushort)in[i];
+   }
+}
+static void translate_trifan_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_quads_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+1];
+      (out+j+0)[1] = (ushort)in[i+3];
+      (out+j+0)[2] = (ushort)in[i+0];
+      (out+j+3)[0] = (ushort)in[i+2];
+      (out+j+3)[1] = (ushort)in[i+3];
+      (out+j+3)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_quadstrip_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+3];
+      (out+j+0)[2] = (ushort)in[i+2];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+3];
+      (out+j+3)[2] = (ushort)in[i+0];
+   }
+}
+static void translate_polygon_uint2ushort_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_points_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i+1];
+      (out+i)[1] = (ushort)in[i];
+   }
+}
+static void translate_linestrip_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+}
+static void translate_lineloop_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i];
+   }
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i];
+}
+static void translate_tris_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i+2];
+      (out+i)[1] = (ushort)in[i];
+      (out+i)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_tristrip_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+(i&1)];
+      (out+j)[2] = (ushort)in[i+1-(i&1)];
+   }
+}
+static void translate_trifan_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[0];
+      (out+j)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_quads_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+3];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+1];
+      (out+j+3)[0] = (ushort)in[i+3];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quadstrip_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+3];
+      (out+j+0)[1] = (ushort)in[i+2];
+      (out+j+0)[2] = (ushort)in[i+0];
+      (out+j+3)[0] = (ushort)in[i+3];
+      (out+j+3)[1] = (ushort)in[i+0];
+      (out+j+3)[2] = (ushort)in[i+1];
+   }
+}
+static void translate_polygon_uint2ushort_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_points_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (ushort)in[i];
+   }
+}
+static void translate_lines_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_linestrip_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+}
+static void translate_lineloop_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+   }
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[0];
+}
+static void translate_tris_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (ushort)in[i];
+      (out+i)[1] = (ushort)in[i+1];
+      (out+i)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_tristrip_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+(i&1)];
+      (out+j)[1] = (ushort)in[i+1-(i&1)];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_trifan_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j)[2] = (ushort)in[i+2];
+   }
+}
+static void translate_quads_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (ushort)in[i+0];
+      (out+j+0)[1] = (ushort)in[i+1];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+1];
+      (out+j+3)[1] = (ushort)in[i+2];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_quadstrip_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (ushort)in[i+2];
+      (out+j+0)[1] = (ushort)in[i+0];
+      (out+j+0)[2] = (ushort)in[i+3];
+      (out+j+3)[0] = (ushort)in[i+0];
+      (out+j+3)[1] = (ushort)in[i+1];
+      (out+j+3)[2] = (ushort)in[i+3];
+   }
+}
+static void translate_polygon_uint2ushort_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (ushort)in[i+1];
+      (out+j)[1] = (ushort)in[i+2];
+      (out+j)[2] = (ushort)in[0];
+   }
+}
+static void translate_points_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+   }
+}
+static void translate_linestrip_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+}
+static void translate_lineloop_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[0];
+}
+static void translate_tris_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+      (out+i)[2] = (uint)in[i+2];
+   }
+}
+static void translate_tristrip_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1+(i&1)];
+      (out+j)[2] = (uint)in[i+2-(i&1)];
+   }
+}
+static void translate_trifan_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quads_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+1];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+2];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_quadstrip_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+2];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+0];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_polygon_uint2uint_first2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_points_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i];
+   }
+}
+static void translate_linestrip_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+}
+static void translate_lineloop_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i];
+}
+static void translate_tris_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i+2];
+      (out+i)[2] = (uint)in[i];
+   }
+}
+static void translate_tristrip_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1+(i&1)];
+      (out+j)[1] = (uint)in[i+2-(i&1)];
+      (out+j)[2] = (uint)in[i];
+   }
+}
+static void translate_trifan_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_quads_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+1];
+      (out+j+0)[1] = (uint)in[i+3];
+      (out+j+0)[2] = (uint)in[i+0];
+      (out+j+3)[0] = (uint)in[i+2];
+      (out+j+3)[1] = (uint)in[i+3];
+      (out+j+3)[2] = (uint)in[i+1];
+   }
+}
+static void translate_quadstrip_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+3];
+      (out+j+0)[2] = (uint)in[i+2];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+3];
+      (out+j+3)[2] = (uint)in[i+0];
+   }
+}
+static void translate_polygon_uint2uint_first2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+static void translate_points_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i+1];
+      (out+i)[1] = (uint)in[i];
+   }
+}
+static void translate_linestrip_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+}
+static void translate_lineloop_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i];
+   }
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i];
+}
+static void translate_tris_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i+2];
+      (out+i)[1] = (uint)in[i];
+      (out+i)[2] = (uint)in[i+1];
+   }
+}
+static void translate_tristrip_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+(i&1)];
+      (out+j)[2] = (uint)in[i+1-(i&1)];
+   }
+}
+static void translate_trifan_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[0];
+      (out+j)[2] = (uint)in[i+1];
+   }
+}
+static void translate_quads_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+3];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+1];
+      (out+j+3)[0] = (uint)in[i+3];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quadstrip_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+3];
+      (out+j+0)[1] = (uint)in[i+2];
+      (out+j+0)[2] = (uint)in[i+0];
+      (out+j+3)[0] = (uint)in[i+3];
+      (out+j+3)[1] = (uint)in[i+0];
+      (out+j+3)[2] = (uint)in[i+1];
+   }
+}
+static void translate_polygon_uint2uint_last2first(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_points_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i++) { 
+      (out+i)[0] = (uint)in[i];
+   }
+}
+static void translate_lines_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=2) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+   }
+}
+static void translate_linestrip_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+}
+static void translate_lineloop_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr - 2; j+=2, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+   }
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[0];
+}
+static void translate_tris_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (i = 0; i < nr; i+=3) { 
+      (out+i)[0] = (uint)in[i];
+      (out+i)[1] = (uint)in[i+1];
+      (out+i)[2] = (uint)in[i+2];
+   }
+}
+static void translate_tristrip_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+(i&1)];
+      (out+j)[1] = (uint)in[i+1-(i&1)];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_trifan_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j)[2] = (uint)in[i+2];
+   }
+}
+static void translate_quads_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=4) { 
+      (out+j+0)[0] = (uint)in[i+0];
+      (out+j+0)[1] = (uint)in[i+1];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+1];
+      (out+j+3)[1] = (uint)in[i+2];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_quadstrip_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=2) { 
+      (out+j+0)[0] = (uint)in[i+2];
+      (out+j+0)[1] = (uint)in[i+0];
+      (out+j+0)[2] = (uint)in[i+3];
+      (out+j+3)[0] = (uint)in[i+0];
+      (out+j+3)[1] = (uint)in[i+1];
+      (out+j+3)[2] = (uint)in[i+3];
+   }
+}
+static void translate_polygon_uint2uint_last2last(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=3, i++) { 
+      (out+j)[0] = (uint)in[i+1];
+      (out+j)[1] = (uint)in[i+2];
+      (out+j)[2] = (uint)in[0];
+   }
+}
+void u_index_init( void )
+{
+  static int firsttime = 1;
+  if (!firsttime) return;
+  firsttime = 0;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = generate_points_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = generate_lines_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = generate_tris_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = generate_quads_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = generate_polygon_ushort_first2first;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = generate_points_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = generate_lines_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = generate_tris_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = generate_quads_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_ushort_first2last;
+generate[OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = generate_polygon_ushort_first2last;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = generate_points_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = generate_lines_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = generate_tris_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = generate_quads_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = generate_polygon_ushort_last2first;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = generate_points_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = generate_lines_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = generate_tris_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = generate_quads_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_ushort_last2last;
+generate[OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = generate_polygon_ushort_last2last;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = generate_points_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = generate_lines_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = generate_tris_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = generate_quads_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = generate_polygon_uint_first2first;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = generate_points_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = generate_lines_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = generate_tris_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = generate_quads_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_uint_first2last;
+generate[OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = generate_polygon_uint_first2last;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = generate_points_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = generate_lines_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = generate_tris_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = generate_quads_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = generate_polygon_uint_last2first;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = generate_points_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = generate_lines_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = generate_linestrip_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = generate_lineloop_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = generate_tris_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = generate_quads_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_uint_last2last;
+generate[OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = generate_polygon_uint_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2ushort_first2first;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2ushort_first2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2ushort_last2first;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2ushort_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2uint_first2first;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2uint_first2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2uint_last2first;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2uint_last2last;
+translate[IN_UBYTE][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2uint_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2ushort_first2first;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2ushort_first2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2ushort_last2first;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2ushort_last2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2uint_first2first;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2uint_first2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2uint_last2first;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2uint_last2last;
+translate[IN_USHORT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_ushort2uint_last2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_uint2ushort_first2first;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_uint2ushort_first2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_uint2ushort_last2first;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2ushort_last2last;
+translate[IN_UINT][OUT_USHORT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_uint2ushort_last2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_uint2uint_first2first;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_FIRST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_uint2uint_first2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POINTS] = translate_points_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINES] = translate_lines_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUADS] = translate_quads_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_FIRST][PIPE_PRIM_POLYGON] = translate_polygon_uint2uint_last2first;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POINTS] = translate_points_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINES] = translate_lines_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_STRIP] = translate_linestrip_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_LINE_LOOP] = translate_lineloop_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLES] = translate_tris_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUADS] = translate_quads_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2uint_last2last;
+translate[IN_UINT][OUT_UINT][PV_LAST][PV_LAST][PIPE_PRIM_POLYGON] = translate_polygon_uint2uint_last2last;
+}
+#include "indices/u_indices.c"
diff --git a/lib/libGL/gallium/libgallium/generated/u_unfilled_gen.c b/lib/libGL/gallium/libgallium/generated/u_unfilled_gen.c
new file mode 100644
index 000000000..93897c98d
--- /dev/null
+++ b/lib/libGL/gallium/libgallium/generated/u_unfilled_gen.c
@@ -0,0 +1,992 @@
+/* File automatically generated by u_unfilled_gen.py */
+
+/*
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
+ * VMWARE AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+
+
+/**
+ * @file
+ * Functions to translate and generate index lists
+ */
+
+#include "indices/u_indices.h"
+#include "indices/u_indices_priv.h"
+#include "pipe/p_compiler.h"
+#include "util/u_debug.h"
+#include "pipe/p_defines.h"
+#include "util/u_memory.h"
+
+
+static unsigned out_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return OUT_UINT;
+   case 2: return OUT_USHORT;
+   default: assert(0); return OUT_USHORT;
+   }
+}
+
+static unsigned in_size_idx( unsigned index_size )
+{
+   switch (index_size) {
+   case 4: return IN_UINT;
+   case 2: return IN_USHORT;
+   case 1: return IN_UBYTE;
+   default: assert(0); return IN_UBYTE;
+   }
+}
+
+
+static u_generate_func generate_line[OUT_COUNT][PRIM_COUNT];
+static u_translate_func translate_line[IN_COUNT][OUT_COUNT][PRIM_COUNT];
+
+
+static void generate_tris_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(i);
+   }
+}
+static void generate_tristrip_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)(i);
+      (out+j)[1] = (ushort)(i+1/*+(i&1)*/);
+      (out+j+2)[0] = (ushort)(i+1/*+(i&1)*/);
+      (out+j+2)[1] = (ushort)(i+2/*-(i&1)*/);
+      (out+j+4)[0] = (ushort)(i+2/*-(i&1)*/);
+      (out+j+4)[1] = (ushort)(i);
+   }
+}
+static void generate_trifan_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(0);
+   }
+}
+static void generate_quads_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (ushort)(i+0);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(i+3);
+      (out+j+6)[0] = (ushort)(i+3);
+      (out+j+6)[1] = (ushort)(i+0);
+   }
+}
+static void generate_quadstrip_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (ushort)(i+2);
+      (out+j)[1] = (ushort)(i+0);
+      (out+j+2)[0] = (ushort)(i+0);
+      (out+j+2)[1] = (ushort)(i+1);
+      (out+j+4)[0] = (ushort)(i+1);
+      (out+j+4)[1] = (ushort)(i+3);
+      (out+j+6)[0] = (ushort)(i+3);
+      (out+j+6)[1] = (ushort)(i+2);
+   }
+}
+static void generate_polygon_ushort(
+    unsigned nr,
+    void *_out )
+{
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)(0);
+      (out+j)[1] = (ushort)(i+1);
+      (out+j+2)[0] = (ushort)(i+1);
+      (out+j+2)[1] = (ushort)(i+2);
+      (out+j+4)[0] = (ushort)(i+2);
+      (out+j+4)[1] = (ushort)(0);
+   }
+}
+static void generate_tris_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(i);
+   }
+}
+static void generate_tristrip_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)(i);
+      (out+j)[1] = (uint)(i+1/*+(i&1)*/);
+      (out+j+2)[0] = (uint)(i+1/*+(i&1)*/);
+      (out+j+2)[1] = (uint)(i+2/*-(i&1)*/);
+      (out+j+4)[0] = (uint)(i+2/*-(i&1)*/);
+      (out+j+4)[1] = (uint)(i);
+   }
+}
+static void generate_trifan_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(0);
+   }
+}
+static void generate_quads_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (uint)(i+0);
+      (out+j)[1] = (uint)(i+1);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(i+3);
+      (out+j+6)[0] = (uint)(i+3);
+      (out+j+6)[1] = (uint)(i+0);
+   }
+}
+static void generate_quadstrip_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (uint)(i+2);
+      (out+j)[1] = (uint)(i+0);
+      (out+j+2)[0] = (uint)(i+0);
+      (out+j+2)[1] = (uint)(i+1);
+      (out+j+4)[0] = (uint)(i+1);
+      (out+j+4)[1] = (uint)(i+3);
+      (out+j+6)[0] = (uint)(i+3);
+      (out+j+6)[1] = (uint)(i+2);
+   }
+}
+static void generate_polygon_uint(
+    unsigned nr,
+    void *_out )
+{
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)(0);
+      (out+j)[1] = (uint)(i+1);
+      (out+j+2)[0] = (uint)(i+1);
+      (out+j+2)[1] = (uint)(i+2);
+      (out+j+4)[0] = (uint)(i+2);
+      (out+j+4)[1] = (uint)(0);
+   }
+}
+static void translate_tris_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_trifan_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_quads_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (ushort)in[i+0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i+3];
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+0];
+   }
+}
+static void translate_quadstrip_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+0];
+      (out+j+2)[0] = (ushort)in[i+0];
+      (out+j+2)[1] = (ushort)in[i+1];
+      (out+j+4)[0] = (ushort)in[i+1];
+      (out+j+4)[1] = (ushort)in[i+3];
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+2];
+   }
+}
+static void translate_polygon_ubyte2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_tris_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_tristrip_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_trifan_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_quads_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (uint)in[i+0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i+3];
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+0];
+   }
+}
+static void translate_quadstrip_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+0];
+      (out+j+2)[0] = (uint)in[i+0];
+      (out+j+2)[1] = (uint)in[i+1];
+      (out+j+4)[0] = (uint)in[i+1];
+      (out+j+4)[1] = (uint)in[i+3];
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+2];
+   }
+}
+static void translate_polygon_ubyte2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ubyte*in = (const ubyte*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_tris_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_trifan_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_quads_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (ushort)in[i+0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i+3];
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+0];
+   }
+}
+static void translate_quadstrip_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+0];
+      (out+j+2)[0] = (ushort)in[i+0];
+      (out+j+2)[1] = (ushort)in[i+1];
+      (out+j+4)[0] = (ushort)in[i+1];
+      (out+j+4)[1] = (ushort)in[i+3];
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+2];
+   }
+}
+static void translate_polygon_ushort2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_tris_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_tristrip_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_trifan_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_quads_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (uint)in[i+0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i+3];
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+0];
+   }
+}
+static void translate_quadstrip_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+0];
+      (out+j+2)[0] = (uint)in[i+0];
+      (out+j+2)[1] = (uint)in[i+1];
+      (out+j+4)[0] = (uint)in[i+1];
+      (out+j+4)[1] = (uint)in[i+3];
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+2];
+   }
+}
+static void translate_polygon_ushort2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const ushort*in = (const ushort*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_tris_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_tristrip_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[i];
+      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (ushort)in[i];
+   }
+}
+static void translate_trifan_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_quads_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (ushort)in[i+0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[i+3];
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+0];
+   }
+}
+static void translate_quadstrip_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (ushort)in[i+2];
+      (out+j)[1] = (ushort)in[i+0];
+      (out+j+2)[0] = (ushort)in[i+0];
+      (out+j+2)[1] = (ushort)in[i+1];
+      (out+j+4)[0] = (ushort)in[i+1];
+      (out+j+4)[1] = (ushort)in[i+3];
+      (out+j+6)[0] = (ushort)in[i+3];
+      (out+j+6)[1] = (ushort)in[i+2];
+   }
+}
+static void translate_polygon_uint2ushort(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  ushort *out = (ushort*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (ushort)in[0];
+      (out+j)[1] = (ushort)in[i+1];
+      (out+j+2)[0] = (ushort)in[i+1];
+      (out+j+2)[1] = (ushort)in[i+2];
+      (out+j+4)[0] = (ushort)in[i+2];
+      (out+j+4)[1] = (ushort)in[0];
+   }
+}
+static void translate_tris_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i+=3) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_tristrip_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[i];
+      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
+      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
+      (out+j+4)[1] = (uint)in[i];
+   }
+}
+static void translate_trifan_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+static void translate_quads_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=4) { 
+      (out+j)[0] = (uint)in[i+0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[i+3];
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+0];
+   }
+}
+static void translate_quadstrip_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=8, i+=2) { 
+      (out+j)[0] = (uint)in[i+2];
+      (out+j)[1] = (uint)in[i+0];
+      (out+j+2)[0] = (uint)in[i+0];
+      (out+j+2)[1] = (uint)in[i+1];
+      (out+j+4)[0] = (uint)in[i+1];
+      (out+j+4)[1] = (uint)in[i+3];
+      (out+j+6)[0] = (uint)in[i+3];
+      (out+j+6)[1] = (uint)in[i+2];
+   }
+}
+static void translate_polygon_uint2uint(
+    const void * _in,
+    unsigned nr,
+    void *_out )
+{
+  const uint*in = (const uint*)_in;
+  uint *out = (uint*)_out;
+  unsigned i, j;
+  (void)j;
+  for (j = i = 0; j < nr; j+=6, i++) { 
+      (out+j)[0] = (uint)in[0];
+      (out+j)[1] = (uint)in[i+1];
+      (out+j+2)[0] = (uint)in[i+1];
+      (out+j+2)[1] = (uint)in[i+2];
+      (out+j+4)[0] = (uint)in[i+2];
+      (out+j+4)[1] = (uint)in[0];
+   }
+}
+void u_unfilled_init( void )
+{
+  static int firsttime = 1;
+  if (!firsttime) return;
+  firsttime = 0;
+generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLES] = generate_tris_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_QUADS] = generate_quads_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_ushort;
+generate_line[OUT_USHORT][PIPE_PRIM_POLYGON] = generate_polygon_ushort;
+generate_line[OUT_UINT][PIPE_PRIM_TRIANGLES] = generate_tris_uint;
+generate_line[OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_uint;
+generate_line[OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_uint;
+generate_line[OUT_UINT][PIPE_PRIM_QUADS] = generate_quads_uint;
+generate_line[OUT_UINT][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_uint;
+generate_line[OUT_UINT][PIPE_PRIM_POLYGON] = generate_polygon_uint;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2ushort;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2uint;
+translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2uint;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2ushort;
+translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_ushort2ushort;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2uint;
+translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_ushort2uint;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2ushort;
+translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_uint2ushort;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2uint;
+translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_uint2uint;
+}
+#include "indices/u_unfilled_indices.c"
diff --git a/lib/libGL/gallium/libmesagallium/Makefile b/lib/libGL/gallium/libmesagallium/Makefile
new file mode 100644
index 000000000..ce9db71d6
--- /dev/null
+++ b/lib/libGL/gallium/libmesagallium/Makefile
@@ -0,0 +1,310 @@
+# $OpenBSD: Makefile,v 1.1 2012/07/13 12:43:34 mpi Exp $
+
+LIB=	mesagallium
+
+MESA=		${.CURDIR}/../../../../dist/Mesa/src/mesa
+MAPI=		${.CURDIR}/../../../../dist/Mesa/src/mapi
+GLSL=		${.CURDIR}/../../../../dist/Mesa/src/glsl
+GALLIUM=	${.CURDIR}/../../../../dist/Mesa/src/gallium
+
+MESA_INCLUDE=	${.CURDIR}/../../../../dist/Mesa/include
+
+CPPFLAGS+= \
+	-I. \
+	-I${MAPI} \
+	-I${MESA_INCLUDE} \
+	-I${MESA_INCLUDE}/GL/internal \
+	-I${MESA} \
+	-I${MESA}/main \
+	-I${MESA}/math \
+	-I${MESA}/drivers/dri/common \
+	-I${GLSL} \
+	-I${GALLIUM}/auxiliary \
+	-I${GALLIUM}/include \
+	-I${GALLIUM}/state_trackers/dri/common \
+	-I${X11BASE}/include \
+	-DPTHREADS \
+	-DHAVE_ALIAS \
+	-D_REENTRANT \
+	-DIN_DRI_DRIVER \
+	-DHAVE_POSIX_MEMALIGN \
+	-D__NOT_HAVE_DRM_H
+
+.if ${MACHINE_ARCH} == "i386"
+CPPFLAGS += \
+	-I${MESA}/x86 \
+	-DUSE_X86_ASM \
+	-DUSE_MMX_ASM \
+	-DUSE_3DNOW_ASM \
+	-DUSE_SSE_ASM
+.elif ${MACHINE_ARCH} == "amd64"
+CPPFLAGS += \
+	-I${MESA}/x86 \
+	-D__GLX_ALIGN64 -DUSE_X86_64_ASM
+.endif
+
+# this is part of MAIN_SOURCES
+MAIN_ES_SOURCES = \
+	api_exec_es1.c \
+	api_exec_es2.c
+
+# main
+MAIN_SOURCES = \
+	api_arrayelt.c \
+	api_exec.c \
+	api_loopback.c \
+	api_noop.c \
+	api_validate.c \
+	accum.c \
+	arbprogram.c \
+	atifragshader.c \
+	attrib.c \
+	arrayobj.c \
+	blend.c \
+	bufferobj.c \
+	buffers.c \
+	clear.c \
+	clip.c \
+	colortab.c \
+	condrender.c \
+	context.c \
+	convolve.c \
+	cpuinfo.c \
+	debug.c \
+	depth.c \
+	depthstencil.c \
+	dlist.c \
+	dlopen.c \
+	drawpix.c \
+	drawtex.c \
+	enable.c \
+	enums.c \
+	eval.c \
+	execmem.c \
+	extensions.c \
+	fbobject.c \
+	feedback.c \
+	ffvertex_prog.c \
+	fog.c \
+	formats.c \
+	framebuffer.c \
+	get.c \
+	getstring.c \
+	hash.c \
+	hint.c \
+	histogram.c \
+	image.c \
+	imports.c \
+	light.c \
+	lines.c \
+	matrix.c \
+	mipmap.c \
+	mm.c \
+	multisample.c \
+	nvprogram.c \
+	pack.c \
+	pixel.c \
+	pixelstore.c \
+	pixeltransfer.c \
+	points.c \
+	polygon.c \
+	queryobj.c \
+	querymatrix.c \
+	rastpos.c \
+	readpix.c \
+	remap.c \
+	renderbuffer.c \
+	scissor.c \
+	shaderapi.c \
+	shaderobj.c \
+	shared.c \
+	state.c \
+	stencil.c \
+	syncobj.c \
+	texcompress.c \
+	texcompress_s3tc.c \
+	texcompress_fxt1.c \
+	texenv.c \
+	texenvprogram.c \
+	texfetch.c \
+	texformat.c \
+	texgen.c \
+	texgetimage.c \
+	teximage.c \
+	texobj.c \
+	texpal.c \
+	texparam.c \
+	texrender.c \
+	texstate.c \
+	texstore.c \
+	transformfeedback.c \
+	uniforms.c \
+	varray.c \
+	version.c \
+	viewport.c \
+	vtxfmt.c \
+	${MAIN_ES_SOURCES}
+
+# math
+MATH_SOURCES= \
+	m_debug_clip.c \
+	m_debug_norm.c \
+	m_debug_xform.c \
+	m_eval.c \
+	m_matrix.c \
+	m_translate.c \
+	m_vector.c
+
+VBO_SOURCES = \
+	vbo_context.c \
+	vbo_exec.c \
+	vbo_exec_api.c \
+	vbo_exec_array.c \
+	vbo_exec_draw.c \
+	vbo_exec_eval.c \
+	vbo_rebase.c \
+	vbo_split.c \
+	vbo_split_copy.c \
+	vbo_split_inplace.c \
+	vbo_save.c \
+	vbo_save_api.c \
+	vbo_save_draw.c \
+	vbo_save_loopback.c 
+
+STATETRACKER_SOURCES = \
+	st_atom.c \
+	st_atom_blend.c \
+	st_atom_clip.c \
+	st_atom_constbuf.c \
+	st_atom_depth.c \
+	st_atom_framebuffer.c \
+	st_atom_msaa.c \
+	st_atom_pixeltransfer.c \
+	st_atom_sampler.c \
+	st_atom_scissor.c \
+	st_atom_shader.c \
+	st_atom_rasterizer.c \
+	st_atom_stipple.c \
+	st_atom_texture.c \
+	st_atom_viewport.c \
+	st_cb_accum.c \
+	st_cb_bitmap.c \
+	st_cb_blit.c \
+	st_cb_bufferobjects.c \
+	st_cb_clear.c \
+	st_cb_condrender.c \
+	st_cb_flush.c \
+	st_cb_drawpixels.c \
+	st_cb_drawtex.c \
+	st_cb_eglimage.c \
+	st_cb_fbo.c \
+	st_cb_feedback.c \
+	st_cb_program.c \
+	st_cb_queryobj.c \
+	st_cb_rasterpos.c \
+	st_cb_readpixels.c \
+	st_cb_strings.c \
+	st_cb_texture.c \
+	st_cb_viewport.c \
+	st_cb_xformfb.c \
+	st_context.c \
+	st_debug.c \
+	st_draw.c \
+	st_draw_feedback.c \
+	st_extensions.c \
+	st_format.c \
+	st_gen_mipmap.c \
+	st_manager.c \
+	st_mesa_to_tgsi.c \
+	st_program.c \
+	st_texture.c
+
+PROGRAM_SOURCES = \
+	arbprogparse.c \
+	hash_table.c \
+	lex.yy.c \
+	nvfragparse.c \
+	nvvertparse.c \
+	program.c \
+	program_parse.tab.c \
+	program_parse_extra.c \
+	prog_cache.c \
+	prog_execute.c \
+	prog_instruction.c \
+	prog_noise.c \
+	prog_optimize.c \
+	prog_parameter.c \
+	prog_parameter_layout.c \
+	prog_print.c \
+	prog_statevars.c \
+	prog_uniform.c \
+	programopt.c \
+	register_allocate.c \
+	symbol_table.c
+
+SHADER_CXX_SOURCES = \
+	ir_to_mesa.cpp \
+	sampler.cpp
+
+.if ${MACHINE_ARCH} == i386
+ASM_C_SOURCES= \
+	common_x86.c \
+.elif ${MACHINE_ARCH} == powerpc
+ASM_C_SOURCES= \
+	common_ppc.c
+.endif
+
+# drisw
+DRISW_SOURCES = \
+	dri_sw_winsys.c \
+        dri_context.c \
+        dri_screen.c \
+        dri_drawable.c \
+        drisw.c
+
+MESAGALLIUM_SOURCES = \
+	${MAIN_SOURCES}		\
+	${MATH_SOURCES}		\
+	${VBO_SOURCES}		\
+	${PROGRAM_SOURCES}	\
+	${STATETRACKER_SOURCES} \
+	${ASM_C_SOURCES}	\
+	${DRISW_SOURCES}
+
+MESAGALLIUM_CXX_SOURCES = \
+	${SHADER_CXX_SOURCES}
+
+SRCS=	${MESAGALLIUM_SOURCES}	\
+	${MESAGALLIUM_CXX_SOURCES} \
+	${ASM_SOURCES} \
+	_atexit.c
+
+all: lib${LIB}_pic.a
+
+obj: _xenocara_obj
+
+install:
+	@echo "Not installing lib${LIB}"
+
+clean: 
+	rm -f lib${LIB}_pic.a ${OBJS}
+
+cleandir:	clean
+
+.include <bsd.xorg.mk>
+
+.if ${MACHINE_ARCH} == i386
+.PATH: ${MESA}/x86
+.endif
+.if ${MACHINE_ARCH} == powerpc
+.PATH: ${MESA}/ppc
+.endif
+.PATH: ${MESA}/main
+.PATH: ${MESA}/math
+.PATH: ${MESA}/state_tracker
+.PATH: ${MESA}/program
+.PATH: ${MESA}/vbo
+.PATH: ${GALLIUM}/winsys/sw/dri
+.PATH: ${GALLIUM}/state_trackers/dri/sw
+.PATH: ${GALLIUM}/state_trackers/dri/common
+.PATH: ${.CURDIR}/generated
diff --git a/lib/libGL/gallium/libmesagallium/_atexit.c b/lib/libGL/gallium/libmesagallium/_atexit.c
new file mode 100644
index 000000000..5e2e7f03f
--- /dev/null
+++ b/lib/libGL/gallium/libmesagallium/_atexit.c
@@ -0,0 +1,12 @@
+/**
+ ** Hack for OpenBSD atexit
+ ** Author:  Matthew Dempsky <matthew@dempsky.org>
+ **/
+extern int __cxa_atexit(void (*)(void *), void *, void *);
+extern void *__dso_handle;
+
+int __attribute__((visibility("hidden")))
+atexit(void (*fn)(void))
+{
+   return (__cxa_atexit((void (*)(void *))fn, 0, &__dso_handle));
+}
diff --git a/lib/libGL/gallium/libmesagallium/generated/Makefile b/lib/libGL/gallium/libmesagallium/generated/Makefile
new file mode 100644
index 000000000..4f9ed94a3
--- /dev/null
+++ b/lib/libGL/gallium/libmesagallium/generated/Makefile
@@ -0,0 +1,31 @@
+# $OpenBSD: Makefile,v 1.1 2012/07/13 12:43:34 mpi Exp $
+
+MESA=	${.CURDIR}/../../../../dist/Mesa/src/mesa
+
+PYTHON=		python${PYTHON_VERSION}
+PYTHON_FLAGS=	-t -O -O
+
+GENERATED= api_exec_es1.c api_exec_es2.c
+
+all: ${GENERATED}
+
+obj:
+
+depend:
+
+install:
+
+clean distclean:
+	rm -f ${GENERATED}
+
+api_exec_es1.c: APIspec.xml es_generator.py APIspecutil.py APIspec.py
+	$(PYTHON) $(PYTHON_FLAGS) ${MESA}/main/es_generator.py -S ${MESA}/main/APIspec.xml -V GLES1.1 > $@
+
+api_exec_es2.c: APIspec.xml es_generator.py APIspecutil.py APIspec.py
+	$(PYTHON) $(PYTHON_FLAGS) ${MESA}/main/es_generator.py -S ${MESA}/main/APIspec.xml -V GLES2.0 > $@
+
+.SUFFIXES: .py .xml
+
+.PATH:	${MESA}/main
+
+.include <bsd.xorg.mk>
diff --git a/lib/libGL/gallium/libmesagallium/generated/api_exec_es1.c b/lib/libGL/gallium/libmesagallium/generated/api_exec_es1.c
new file mode 100644
index 000000000..54b3f0d62
--- /dev/null
+++ b/lib/libGL/gallium/libmesagallium/generated/api_exec_es1.c
@@ -0,0 +1,4821 @@
+/* DO NOT EDIT *************************************************
+ * THIS FILE AUTOMATICALLY GENERATED BY THE es_generator.py SCRIPT
+ * API specification file:   /share/OpenBSD/xenocara/lib/libGL/libmesa/generated/../../../../dist/Mesa/src/mesa/main/APIspec.xml
+ * GLES version:             GLES1.1
+ * date:                     2011-05-15 15:41:33
+ */
+
+
+#include "GLES/gl.h"
+#include "GLES/glext.h"
+#include "main/mfeatures.h"
+
+#if FEATURE_ES1
+
+
+/* These types are needed for the Mesa veneer, but are not defined in
+ * the standard GLES headers.
+ */
+typedef double GLdouble;
+typedef double GLclampd;
+
+/* Mesa error handling requires these */
+extern void *_mesa_get_current_context(void);
+extern void _mesa_error(void *ctx, GLenum error, const char *fmtString, ... );
+
+#include "main/compiler.h"
+#include "main/api_exec.h"
+#include "main/remap.h"
+
+/* cannot include main/dispatch.h here */
+#ifdef IN_DRI_DRIVER
+#define _GLAPI_USE_REMAP_TABLE
+#endif
+/* glapi uses GLAPIENTRY while GLES headers define GL_APIENTRY */
+#ifndef GLAPIENTRY
+#define GLAPIENTRY GL_APIENTRY
+#endif
+#include "es1api/glapi/glapitable.h"
+#include "es1api/glapi/glapioffsets.h"
+#include "es1api/glapi/glapidispatch.h"
+
+#if FEATURE_remap_table
+
+#if !FEATURE_GL
+int driDispatchRemapTable[driDispatchRemapTable_size];
+#endif
+
+#define need_MESA_remap_table
+
+#include "es1api/main/remap_helper.h"
+
+void
+_mesa_init_remap_table_es1(void)
+{
+   _mesa_do_init_remap_table(_mesa_function_pool,
+                             driDispatchRemapTable_size,
+                             MESA_remap_table_functions);
+}
+
+void
+_mesa_map_static_functions_es1(void)
+{
+}
+
+#endif
+
+typedef void (*_glapi_proc)(void); /* generic function pointer */
+
+/*************************************************************
+ * Generated functions begin here
+ */
+
+extern void GL_APIENTRY _mesa_ActiveTextureARB(GLenum texture);
+static void GL_APIENTRY _es_ActiveTexture(GLenum texture)
+{
+    switch(texture) {
+        case GL_TEXTURE0:
+        case GL_TEXTURE1:
+        case GL_TEXTURE2:
+        case GL_TEXTURE3:
+        case GL_TEXTURE4:
+        case GL_TEXTURE5:
+        case GL_TEXTURE6:
+        case GL_TEXTURE7:
+        case GL_TEXTURE8:
+        case GL_TEXTURE9:
+        case GL_TEXTURE10:
+        case GL_TEXTURE11:
+        case GL_TEXTURE12:
+        case GL_TEXTURE13:
+        case GL_TEXTURE14:
+        case GL_TEXTURE15:
+        case GL_TEXTURE16:
+        case GL_TEXTURE17:
+        case GL_TEXTURE18:
+        case GL_TEXTURE19:
+        case GL_TEXTURE20:
+        case GL_TEXTURE21:
+        case GL_TEXTURE22:
+        case GL_TEXTURE23:
+        case GL_TEXTURE24:
+        case GL_TEXTURE25:
+        case GL_TEXTURE26:
+        case GL_TEXTURE27:
+        case GL_TEXTURE28:
+        case GL_TEXTURE29:
+        case GL_TEXTURE30:
+        case GL_TEXTURE31:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glActiveTexture(texture=0x%x)", texture);
+            return;
+    }
+
+    _mesa_ActiveTextureARB(texture);
+}
+
+extern void GL_APIENTRY _mesa_AlphaFunc(GLenum func, GLclampf ref);
+static void GL_APIENTRY _es_AlphaFunc(GLenum func, GLclampf ref)
+{
+    switch(func) {
+        case GL_NEVER:
+        case GL_LESS:
+        case GL_EQUAL:
+        case GL_LEQUAL:
+        case GL_GREATER:
+        case GL_NOTEQUAL:
+        case GL_GEQUAL:
+        case GL_ALWAYS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glAlphaFunc(func=0x%x)", func);
+            return;
+    }
+
+    _mesa_AlphaFunc(func, ref);
+}
+
+extern void GL_APIENTRY _mesa_AlphaFunc(GLenum func, GLclampf  ref);
+static void GL_APIENTRY _es_AlphaFuncx(GLenum func, GLclampx ref)
+{
+    GLclampf converted_ref;
+
+    switch(func) {
+        case GL_NEVER:
+        case GL_LESS:
+        case GL_EQUAL:
+        case GL_LEQUAL:
+        case GL_GREATER:
+        case GL_NOTEQUAL:
+        case GL_GEQUAL:
+        case GL_ALWAYS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glAlphaFuncx(func=0x%x)", func);
+            return;
+    }
+
+    converted_ref = (GLclampf) (ref / 65536.0f);
+
+    _mesa_AlphaFunc(func, converted_ref);
+}
+
+extern void GL_APIENTRY _mesa_BindBufferARB(GLenum target, GLuint buffer);
+static void GL_APIENTRY _es_BindBuffer(GLenum target, GLuint buffer)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindBuffer(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindBufferARB(target, buffer);
+}
+
+extern void GL_APIENTRY _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer);
+static void GL_APIENTRY _es_BindFramebufferOES(GLenum target, GLuint framebuffer)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindFramebufferOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindFramebufferEXT(target, framebuffer);
+}
+
+extern void GL_APIENTRY _mesa_BindRenderbufferEXT(GLenum target, GLuint renderbuffer);
+static void GL_APIENTRY _es_BindRenderbufferOES(GLenum target, GLuint renderbuffer)
+{
+    switch(target) {
+        case GL_RENDERBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindRenderbufferOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindRenderbufferEXT(target, renderbuffer);
+}
+
+extern void GL_APIENTRY _mesa_BindTexture(GLenum target, GLuint texture);
+static void GL_APIENTRY _es_BindTexture(GLenum target, GLuint texture)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindTexture(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindTexture(target, texture);
+}
+
+extern void GL_APIENTRY _mesa_BlendEquation(GLenum mode);
+static void GL_APIENTRY _es_BlendEquationOES(GLenum mode)
+{
+    switch(mode) {
+        case GL_FUNC_ADD_OES:
+        case GL_FUNC_SUBTRACT_OES:
+        case GL_FUNC_REVERSE_SUBTRACT_OES:
+        case GL_MIN_EXT:
+        case GL_MAX_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendEquationOES(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_BlendEquation(mode);
+}
+
+extern void GL_APIENTRY _mesa_BlendEquationSeparateEXT(GLenum modeRGB, GLenum modeAlpha);
+static void GL_APIENTRY _es_BlendEquationSeparateOES(GLenum modeRGB, GLenum modeAlpha)
+{
+    switch(modeRGB) {
+        case GL_FUNC_ADD_OES:
+        case GL_FUNC_SUBTRACT_OES:
+        case GL_FUNC_REVERSE_SUBTRACT_OES:
+        case GL_MIN_EXT:
+        case GL_MAX_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendEquationSeparateOES(modeRGB=0x%x)", modeRGB);
+            return;
+    }
+    switch(modeAlpha) {
+        case GL_FUNC_ADD_OES:
+        case GL_FUNC_SUBTRACT_OES:
+        case GL_FUNC_REVERSE_SUBTRACT_OES:
+        case GL_MIN_EXT:
+        case GL_MAX_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendEquationSeparateOES(modeAlpha=0x%x)", modeAlpha);
+            return;
+    }
+
+    _mesa_BlendEquationSeparateEXT(modeRGB, modeAlpha);
+}
+
+extern void GL_APIENTRY _mesa_BlendFunc(GLenum sfactor, GLenum dfactor);
+static void GL_APIENTRY _es_BlendFunc(GLenum sfactor, GLenum dfactor)
+{
+    switch(sfactor) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_SRC_ALPHA_SATURATE:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFunc(sfactor=0x%x)", sfactor);
+            return;
+    }
+    switch(dfactor) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFunc(dfactor=0x%x)", dfactor);
+            return;
+    }
+
+    _mesa_BlendFunc(sfactor, dfactor);
+}
+
+extern void GL_APIENTRY _mesa_BlendFuncSeparateEXT(GLenum srcRGB, GLenum dstRGB, GLenum srcAlpha, GLenum dstAlpha);
+static void GL_APIENTRY _es_BlendFuncSeparateOES(GLenum srcRGB, GLenum dstRGB, GLenum srcAlpha, GLenum dstAlpha)
+{
+    switch(srcRGB) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_SRC_ALPHA_SATURATE:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparateOES(srcRGB=0x%x)", srcRGB);
+            return;
+    }
+    switch(dstRGB) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparateOES(dstRGB=0x%x)", dstRGB);
+            return;
+    }
+    switch(srcAlpha) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_SRC_ALPHA_SATURATE:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparateOES(srcAlpha=0x%x)", srcAlpha);
+            return;
+    }
+    switch(dstAlpha) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparateOES(dstAlpha=0x%x)", dstAlpha);
+            return;
+    }
+
+    _mesa_BlendFuncSeparateEXT(srcRGB, dstRGB, srcAlpha, dstAlpha);
+}
+
+extern void GL_APIENTRY _mesa_BufferDataARB(GLenum target, GLsizeiptr size, const GLvoid * data, GLenum usage);
+static void GL_APIENTRY _es_BufferData(GLenum target, GLsizeiptr size, const GLvoid *data, GLenum usage)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBufferData(target=0x%x)", target);
+            return;
+    }
+    switch(usage) {
+        case GL_STATIC_DRAW:
+        case GL_DYNAMIC_DRAW:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBufferData(usage=0x%x)", usage);
+            return;
+    }
+
+    _mesa_BufferDataARB(target, size, data, usage);
+}
+
+extern void GL_APIENTRY _mesa_BufferSubDataARB(GLenum target, GLintptr offset, GLsizeiptr size, const GLvoid * data);
+static void GL_APIENTRY _es_BufferSubData(GLenum target, GLintptr offset, GLsizeiptr size, const GLvoid *data)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBufferSubData(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BufferSubDataARB(target, offset, size, data);
+}
+
+extern GLenum GL_APIENTRY _mesa_CheckFramebufferStatusEXT(GLenum target);
+static GLenum GL_APIENTRY _es_CheckFramebufferStatusOES(GLenum target)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCheckFramebufferStatusOES(target=0x%x)", target);
+            return (GLenum) 0;
+    }
+
+    return _mesa_CheckFramebufferStatusEXT(target);
+}
+
+extern void GL_APIENTRY _mesa_Clear(GLbitfield mask);
+static void GL_APIENTRY _es_Clear(GLbitfield mask)
+{
+    switch(mask) {
+        case 0:
+        case (GL_COLOR_BUFFER_BIT):
+        case (GL_DEPTH_BUFFER_BIT):
+        case (GL_STENCIL_BUFFER_BIT):
+        case (GL_COLOR_BUFFER_BIT|GL_DEPTH_BUFFER_BIT):
+        case (GL_COLOR_BUFFER_BIT|GL_STENCIL_BUFFER_BIT):
+        case (GL_DEPTH_BUFFER_BIT|GL_STENCIL_BUFFER_BIT):
+        case (GL_COLOR_BUFFER_BIT|GL_DEPTH_BUFFER_BIT|GL_STENCIL_BUFFER_BIT):
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glClear(mask=0x%x)", mask);
+            return;
+    }
+
+    _mesa_Clear(mask);
+}
+
+extern void GL_APIENTRY _mesa_ClearColor(GLclampf red, GLclampf green, GLclampf blue, GLclampf alpha);
+#define _es_ClearColor _mesa_ClearColor
+
+extern void GL_APIENTRY _mesa_ClearColor(GLclampf  red, GLclampf  green, GLclampf  blue, GLclampf  alpha);
+static void GL_APIENTRY _es_ClearColorx(GLclampx red, GLclampx green, GLclampx blue, GLclampx alpha)
+{
+    GLclampf converted_red;
+    GLclampf converted_green;
+    GLclampf converted_blue;
+    GLclampf converted_alpha;
+
+    converted_red = (GLclampf) (red / 65536.0f);
+    converted_green = (GLclampf) (green / 65536.0f);
+    converted_blue = (GLclampf) (blue / 65536.0f);
+    converted_alpha = (GLclampf) (alpha / 65536.0f);
+
+    _mesa_ClearColor(converted_red, converted_green, converted_blue, converted_alpha);
+}
+
+extern void GL_APIENTRY _mesa_ClearDepthf(GLclampf depth);
+#define _es_ClearDepthf _mesa_ClearDepthf
+
+extern void GL_APIENTRY _mesa_ClearDepthf(GLclampf  depth);
+static void GL_APIENTRY _es_ClearDepthx(GLclampx depth)
+{
+    GLclampf converted_depth;
+
+    converted_depth = (GLclampf) (depth / 65536.0f);
+
+    _mesa_ClearDepthf(converted_depth);
+}
+
+extern void GL_APIENTRY _mesa_ClearStencil(GLint s);
+#define _es_ClearStencil _mesa_ClearStencil
+
+extern void GL_APIENTRY _mesa_ClientActiveTextureARB(GLenum texture);
+static void GL_APIENTRY _es_ClientActiveTexture(GLenum texture)
+{
+    switch(texture) {
+        case GL_TEXTURE0:
+        case GL_TEXTURE1:
+        case GL_TEXTURE2:
+        case GL_TEXTURE3:
+        case GL_TEXTURE4:
+        case GL_TEXTURE5:
+        case GL_TEXTURE6:
+        case GL_TEXTURE7:
+        case GL_TEXTURE8:
+        case GL_TEXTURE9:
+        case GL_TEXTURE10:
+        case GL_TEXTURE11:
+        case GL_TEXTURE12:
+        case GL_TEXTURE13:
+        case GL_TEXTURE14:
+        case GL_TEXTURE15:
+        case GL_TEXTURE16:
+        case GL_TEXTURE17:
+        case GL_TEXTURE18:
+        case GL_TEXTURE19:
+        case GL_TEXTURE20:
+        case GL_TEXTURE21:
+        case GL_TEXTURE22:
+        case GL_TEXTURE23:
+        case GL_TEXTURE24:
+        case GL_TEXTURE25:
+        case GL_TEXTURE26:
+        case GL_TEXTURE27:
+        case GL_TEXTURE28:
+        case GL_TEXTURE29:
+        case GL_TEXTURE30:
+        case GL_TEXTURE31:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glClientActiveTexture(texture=0x%x)", texture);
+            return;
+    }
+
+    _mesa_ClientActiveTextureARB(texture);
+}
+
+extern void GL_APIENTRY _mesa_ClipPlane(GLenum plane, GLdouble const * equation);
+static void GL_APIENTRY _es_ClipPlanef(GLenum plane, const GLfloat *equation)
+{
+    register unsigned int i;
+    unsigned int n_equation = 4;
+    GLdouble converted_equation[4];
+
+    switch(plane) {
+        case GL_CLIP_PLANE0:
+        case GL_CLIP_PLANE1:
+        case GL_CLIP_PLANE2:
+        case GL_CLIP_PLANE3:
+        case GL_CLIP_PLANE4:
+        case GL_CLIP_PLANE5:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glClipPlanef(plane=0x%x)", plane);
+            return;
+    }
+
+    for (i = 0; i < n_equation; i++) {
+        converted_equation[i] = (GLdouble) (equation[i]);
+    }
+
+    _mesa_ClipPlane(plane, converted_equation);
+}
+
+extern void GL_APIENTRY _mesa_ClipPlane(GLenum plane, GLdouble const * equation);
+static void GL_APIENTRY _es_ClipPlanex(GLenum plane, const GLfixed *equation)
+{
+    register unsigned int i;
+    unsigned int n_equation = 4;
+    GLdouble converted_equation[4];
+
+    switch(plane) {
+        case GL_CLIP_PLANE0:
+        case GL_CLIP_PLANE1:
+        case GL_CLIP_PLANE2:
+        case GL_CLIP_PLANE3:
+        case GL_CLIP_PLANE4:
+        case GL_CLIP_PLANE5:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glClipPlanex(plane=0x%x)", plane);
+            return;
+    }
+
+    for (i = 0; i < n_equation; i++) {
+        converted_equation[i] = (GLdouble) (equation[i] / 65536.0);
+    }
+
+    _mesa_ClipPlane(plane, converted_equation);
+}
+
+extern void GL_APIENTRY _vbo_Color4f(GLfloat red, GLfloat green, GLfloat blue, GLfloat alpha);
+#define _es_Color4f _vbo_Color4f
+
+extern void GL_APIENTRY _vbo_Color4f(GLfloat  red, GLfloat  green, GLfloat  blue, GLfloat  alpha);
+static void GL_APIENTRY _es_Color4ub(GLubyte red, GLubyte green, GLubyte blue, GLubyte alpha)
+{
+    GLfloat converted_red;
+    GLfloat converted_green;
+    GLfloat converted_blue;
+    GLfloat converted_alpha;
+
+    converted_red = (GLfloat) (red / 255.0f);
+    converted_green = (GLfloat) (green / 255.0f);
+    converted_blue = (GLfloat) (blue / 255.0f);
+    converted_alpha = (GLfloat) (alpha / 255.0f);
+
+    _vbo_Color4f(converted_red, converted_green, converted_blue, converted_alpha);
+}
+
+extern void GL_APIENTRY _vbo_Color4f(GLfloat  red, GLfloat  green, GLfloat  blue, GLfloat  alpha);
+static void GL_APIENTRY _es_Color4x(GLfixed red, GLfixed green, GLfixed blue, GLfixed alpha)
+{
+    GLfloat converted_red;
+    GLfloat converted_green;
+    GLfloat converted_blue;
+    GLfloat converted_alpha;
+
+    converted_red = (GLfloat) (red / 65536.0f);
+    converted_green = (GLfloat) (green / 65536.0f);
+    converted_blue = (GLfloat) (blue / 65536.0f);
+    converted_alpha = (GLfloat) (alpha / 65536.0f);
+
+    _vbo_Color4f(converted_red, converted_green, converted_blue, converted_alpha);
+}
+
+extern void GL_APIENTRY _mesa_ColorMask(GLboolean red, GLboolean green, GLboolean blue, GLboolean alpha);
+#define _es_ColorMask _mesa_ColorMask
+
+extern void GL_APIENTRY _mesa_ColorPointer(GLint size, GLenum type, GLsizei stride, const GLvoid * pointer);
+static void GL_APIENTRY _es_ColorPointer(GLint size, GLenum type, GLsizei stride, const GLvoid *pointer)
+{
+    switch(size) {
+        case 4:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glColorPointer(size=%d)", size);
+            return;
+    }
+    switch(type) {
+        case GL_UNSIGNED_BYTE:
+        case GL_FLOAT:
+        case GL_FIXED:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glColorPointer(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_ColorPointer(size, type, stride, pointer);
+}
+
+extern void GL_APIENTRY _mesa_CompressedTexImage2DARB(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLint border, GLsizei imageSize, const GLvoid * data);
+static void GL_APIENTRY _es_CompressedTexImage2D(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLint border, GLsizei imageSize, const GLvoid *data)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_PALETTE4_RGB8_OES:
+        case GL_PALETTE4_RGBA8_OES:
+        case GL_PALETTE4_R5_G6_B5_OES:
+        case GL_PALETTE4_RGBA4_OES:
+        case GL_PALETTE4_RGB5_A1_OES:
+        case GL_PALETTE8_RGB8_OES:
+        case GL_PALETTE8_RGBA8_OES:
+        case GL_PALETTE8_R5_G6_B5_OES:
+        case GL_PALETTE8_RGBA4_OES:
+        case GL_PALETTE8_RGB5_A1_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexImage2D(internalFormat=0x%x)", internalFormat);
+            return;
+    }
+    switch(border) {
+        case 0:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glCompressedTexImage2D(border=%d)", border);
+            return;
+    }
+
+    _mesa_CompressedTexImage2DARB(target, level, internalFormat, width, height, border, imageSize, data);
+}
+
+extern void GL_APIENTRY _mesa_CompressedTexSubImage2DARB(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLsizei imageSize, const GLvoid * data);
+static void GL_APIENTRY _es_CompressedTexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLsizei imageSize, const GLvoid *data)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexSubImage2D(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_CompressedTexSubImage2DARB(target, level, xoffset, yoffset, width, height, format, imageSize, data);
+}
+
+extern void GL_APIENTRY _mesa_CopyTexImage2D(GLenum target, GLint level, GLenum internalFormat, GLint x, GLint y, GLsizei width, GLsizei height, GLint border);
+static void GL_APIENTRY _es_CopyTexImage2D(GLenum target, GLint level, GLenum internalFormat, GLint x, GLint y, GLsizei width, GLsizei height, GLint border)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCopyTexImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_ALPHA:
+        case GL_RGB:
+        case GL_RGBA:
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glCopyTexImage2D(internalFormat=0x%x)", internalFormat);
+            return;
+    }
+    switch(border) {
+        case 0:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glCopyTexImage2D(border=%d)", border);
+            return;
+    }
+
+    _mesa_CopyTexImage2D(target, level, internalFormat, x, y, width, height, border);
+}
+
+extern void GL_APIENTRY _mesa_CopyTexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint x, GLint y, GLsizei width, GLsizei height);
+static void GL_APIENTRY _es_CopyTexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint x, GLint y, GLsizei width, GLsizei height)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCopyTexSubImage2D(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_CopyTexSubImage2D(target, level, xoffset, yoffset, x, y, width, height);
+}
+
+extern void GL_APIENTRY _mesa_CullFace(GLenum mode);
+static void GL_APIENTRY _es_CullFace(GLenum mode)
+{
+    switch(mode) {
+        case GL_FRONT:
+        case GL_BACK:
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCullFace(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_CullFace(mode);
+}
+
+extern void GL_APIENTRY _mesa_DeleteBuffersARB(GLsizei n, const GLuint * buffer);
+#define _es_DeleteBuffers _mesa_DeleteBuffersARB
+
+extern void GL_APIENTRY _mesa_DeleteFramebuffersEXT(GLsizei n, const GLuint * framebuffers);
+#define _es_DeleteFramebuffersOES _mesa_DeleteFramebuffersEXT
+
+extern void GL_APIENTRY _mesa_DeleteRenderbuffersEXT(GLsizei n, const GLuint * renderbuffers);
+#define _es_DeleteRenderbuffersOES _mesa_DeleteRenderbuffersEXT
+
+extern void GL_APIENTRY _mesa_DeleteTextures(GLsizei n, const GLuint * textures);
+#define _es_DeleteTextures _mesa_DeleteTextures
+
+extern void GL_APIENTRY _mesa_DepthFunc(GLenum func);
+static void GL_APIENTRY _es_DepthFunc(GLenum func)
+{
+    switch(func) {
+        case GL_NEVER:
+        case GL_LESS:
+        case GL_EQUAL:
+        case GL_LEQUAL:
+        case GL_GREATER:
+        case GL_NOTEQUAL:
+        case GL_GEQUAL:
+        case GL_ALWAYS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDepthFunc(func=0x%x)", func);
+            return;
+    }
+
+    _mesa_DepthFunc(func);
+}
+
+extern void GL_APIENTRY _mesa_DepthMask(GLboolean flag);
+#define _es_DepthMask _mesa_DepthMask
+
+extern void GL_APIENTRY _mesa_DepthRangef(GLclampf zNear, GLclampf zFar);
+#define _es_DepthRangef _mesa_DepthRangef
+
+extern void GL_APIENTRY _mesa_DepthRangef(GLclampf  zNear, GLclampf  zFar);
+static void GL_APIENTRY _es_DepthRangex(GLclampx zNear, GLclampx zFar)
+{
+    GLclampf converted_zNear;
+    GLclampf converted_zFar;
+
+    converted_zNear = (GLclampf) (zNear / 65536.0f);
+    converted_zFar = (GLclampf) (zFar / 65536.0f);
+
+    _mesa_DepthRangef(converted_zNear, converted_zFar);
+}
+
+extern void GL_APIENTRY _mesa_Disable(GLenum cap);
+static void GL_APIENTRY _es_Disable(GLenum cap)
+{
+    switch(cap) {
+        case GL_NORMALIZE:
+        case GL_RESCALE_NORMAL:
+        case GL_CLIP_PLANE0:
+        case GL_CLIP_PLANE1:
+        case GL_CLIP_PLANE2:
+        case GL_CLIP_PLANE3:
+        case GL_CLIP_PLANE4:
+        case GL_CLIP_PLANE5:
+        case GL_FOG:
+        case GL_LIGHTING:
+        case GL_COLOR_MATERIAL:
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+        case GL_POINT_SMOOTH:
+        case GL_LINE_SMOOTH:
+        case GL_CULL_FACE:
+        case GL_POLYGON_OFFSET_FILL:
+        case GL_MULTISAMPLE:
+        case GL_SAMPLE_ALPHA_TO_COVERAGE:
+        case GL_SAMPLE_ALPHA_TO_ONE:
+        case GL_SAMPLE_COVERAGE:
+        case GL_TEXTURE_2D:
+        case GL_SCISSOR_TEST:
+        case GL_ALPHA_TEST:
+        case GL_STENCIL_TEST:
+        case GL_DEPTH_TEST:
+        case GL_BLEND:
+        case GL_DITHER:
+        case GL_COLOR_LOGIC_OP:
+        case GL_POINT_SPRITE_OES:
+        case GL_MATRIX_PALETTE_OES:
+        case GL_TEXTURE_CUBE_MAP_OES:
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDisable(cap=0x%x)", cap);
+            return;
+    }
+
+    _mesa_Disable(cap);
+}
+
+extern void GL_APIENTRY _mesa_DisableClientState(GLenum array);
+static void GL_APIENTRY _es_DisableClientState(GLenum array)
+{
+    switch(array) {
+        case GL_VERTEX_ARRAY:
+        case GL_NORMAL_ARRAY:
+        case GL_COLOR_ARRAY:
+        case GL_TEXTURE_COORD_ARRAY:
+        case GL_MATRIX_INDEX_ARRAY_OES:
+        case GL_WEIGHT_ARRAY_OES:
+        case GL_POINT_SIZE_ARRAY_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDisableClientState(array=0x%x)", array);
+            return;
+    }
+
+    _mesa_DisableClientState(array);
+}
+
+extern void GL_APIENTRY _mesa_DrawArrays(GLenum mode, GLint first, GLsizei count);
+static void GL_APIENTRY _es_DrawArrays(GLenum mode, GLint first, GLsizei count)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDrawArrays(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_DrawArrays(mode, first, count);
+}
+
+extern void GL_APIENTRY _mesa_DrawElements(GLenum mode, GLsizei count, GLenum type, const GLvoid * indices);
+static void GL_APIENTRY _es_DrawElements(GLenum mode, GLsizei count, GLenum type, const GLvoid *indices)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDrawElements(mode=0x%x)", mode);
+            return;
+    }
+    switch(type) {
+        case GL_UNSIGNED_BYTE:
+        case GL_UNSIGNED_SHORT:
+        case GL_UNSIGNED_INT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDrawElements(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_DrawElements(mode, count, type, indices);
+}
+
+extern void GL_APIENTRY _mesa_DrawTexf(GLfloat x, GLfloat y, GLfloat z, GLfloat w, GLfloat h);
+#define _es_DrawTexfOES _mesa_DrawTexf
+
+extern void GL_APIENTRY _mesa_DrawTexfv(const GLfloat * coords);
+#define _es_DrawTexfvOES _mesa_DrawTexfv
+
+extern void GL_APIENTRY _mesa_DrawTexi(GLint x, GLint y, GLint z, GLint w, GLint h);
+#define _es_DrawTexiOES _mesa_DrawTexi
+
+extern void GL_APIENTRY _mesa_DrawTexiv(const GLint * coords);
+#define _es_DrawTexivOES _mesa_DrawTexiv
+
+extern void GL_APIENTRY _mesa_DrawTexs(GLshort x, GLshort y, GLshort z, GLshort w, GLshort h);
+#define _es_DrawTexsOES _mesa_DrawTexs
+
+extern void GL_APIENTRY _mesa_DrawTexsv(const GLshort * coords);
+#define _es_DrawTexsvOES _mesa_DrawTexsv
+
+extern void GL_APIENTRY _mesa_DrawTexf(GLfloat  x, GLfloat  y, GLfloat  z, GLfloat  w, GLfloat  h);
+static void GL_APIENTRY _es_DrawTexxOES(GLfixed x, GLfixed y, GLfixed z, GLfixed w, GLfixed h)
+{
+    GLfloat converted_x;
+    GLfloat converted_y;
+    GLfloat converted_z;
+    GLfloat converted_w;
+    GLfloat converted_h;
+
+    converted_x = (GLfloat) (x / 65536.0f);
+    converted_y = (GLfloat) (y / 65536.0f);
+    converted_z = (GLfloat) (z / 65536.0f);
+    converted_w = (GLfloat) (w / 65536.0f);
+    converted_h = (GLfloat) (h / 65536.0f);
+
+    _mesa_DrawTexf(converted_x, converted_y, converted_z, converted_w, converted_h);
+}
+
+extern void GL_APIENTRY _mesa_DrawTexfv(GLfloat const * coords);
+static void GL_APIENTRY _es_DrawTexxvOES(const GLfixed *coords)
+{
+    register unsigned int i;
+    unsigned int n_coords = 5;
+    GLfloat converted_coords[5];
+
+    for (i = 0; i < n_coords; i++) {
+        converted_coords[i] = (GLfloat) (coords[i] / 65536.0f);
+    }
+
+    _mesa_DrawTexfv(converted_coords);
+}
+
+extern void GL_APIENTRY _mesa_EGLImageTargetRenderbufferStorageOES(GLenum target, GLeglImageOES image);
+static void GL_APIENTRY _es_EGLImageTargetRenderbufferStorageOES(GLenum target, GLeglImageOES image)
+{
+    switch(target) {
+        case GL_RENDERBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glEGLImageTargetRenderbufferStorageOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_EGLImageTargetRenderbufferStorageOES(target, image);
+}
+
+extern void GL_APIENTRY _mesa_EGLImageTargetTexture2DOES(GLenum target, GLeglImageOES image);
+static void GL_APIENTRY _es_EGLImageTargetTexture2DOES(GLenum target, GLeglImageOES image)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glEGLImageTargetTexture2DOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_EGLImageTargetTexture2DOES(target, image);
+}
+
+extern void GL_APIENTRY _mesa_Enable(GLenum cap);
+static void GL_APIENTRY _es_Enable(GLenum cap)
+{
+    switch(cap) {
+        case GL_NORMALIZE:
+        case GL_RESCALE_NORMAL:
+        case GL_CLIP_PLANE0:
+        case GL_CLIP_PLANE1:
+        case GL_CLIP_PLANE2:
+        case GL_CLIP_PLANE3:
+        case GL_CLIP_PLANE4:
+        case GL_CLIP_PLANE5:
+        case GL_FOG:
+        case GL_LIGHTING:
+        case GL_COLOR_MATERIAL:
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+        case GL_POINT_SMOOTH:
+        case GL_LINE_SMOOTH:
+        case GL_CULL_FACE:
+        case GL_POLYGON_OFFSET_FILL:
+        case GL_MULTISAMPLE:
+        case GL_SAMPLE_ALPHA_TO_COVERAGE:
+        case GL_SAMPLE_ALPHA_TO_ONE:
+        case GL_SAMPLE_COVERAGE:
+        case GL_TEXTURE_2D:
+        case GL_SCISSOR_TEST:
+        case GL_ALPHA_TEST:
+        case GL_STENCIL_TEST:
+        case GL_DEPTH_TEST:
+        case GL_BLEND:
+        case GL_DITHER:
+        case GL_COLOR_LOGIC_OP:
+        case GL_POINT_SPRITE_OES:
+        case GL_MATRIX_PALETTE_OES:
+        case GL_TEXTURE_CUBE_MAP_OES:
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glEnable(cap=0x%x)", cap);
+            return;
+    }
+
+    _mesa_Enable(cap);
+}
+
+extern void GL_APIENTRY _mesa_EnableClientState(GLenum array);
+static void GL_APIENTRY _es_EnableClientState(GLenum array)
+{
+    switch(array) {
+        case GL_VERTEX_ARRAY:
+        case GL_NORMAL_ARRAY:
+        case GL_COLOR_ARRAY:
+        case GL_TEXTURE_COORD_ARRAY:
+        case GL_MATRIX_INDEX_ARRAY_OES:
+        case GL_WEIGHT_ARRAY_OES:
+        case GL_POINT_SIZE_ARRAY_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glEnableClientState(array=0x%x)", array);
+            return;
+    }
+
+    _mesa_EnableClientState(array);
+}
+
+extern void GL_APIENTRY _mesa_Finish(void);
+#define _es_Finish _mesa_Finish
+
+extern void GL_APIENTRY _mesa_Flush(void);
+#define _es_Flush _mesa_Flush
+
+extern void GL_APIENTRY _mesa_Fogf(GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_Fogf(GLenum pname, GLfloat param)
+{
+    switch(pname) {
+        case GL_FOG_MODE:
+            if (param != GL_EXP && param != GL_EXP2 && param != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_FOG_DENSITY:
+        case GL_FOG_START:
+        case GL_FOG_END:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogf(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_Fogf(pname, param);
+}
+
+extern void GL_APIENTRY _mesa_Fogfv(GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_Fogfv(GLenum pname, const GLfloat *params)
+{
+    switch(pname) {
+        case GL_FOG_MODE: /* size 1 */
+            if (params[0] != GL_EXP && params[0] != GL_EXP2 && params[0] != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_FOG_COLOR: /* size 4 */
+        case GL_FOG_DENSITY: /* size 1 */
+        case GL_FOG_START: /* size 1 */
+        case GL_FOG_END: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_Fogfv(pname, params);
+}
+
+extern void GL_APIENTRY _mesa_Fogf(GLenum pname, GLfloat  param);
+static void GL_APIENTRY _es_Fogx(GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+    int convert_param_value = 1;
+
+    switch(pname) {
+        case GL_FOG_MODE:
+            if (param != GL_EXP && param != GL_EXP2 && param != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_FOG_DENSITY:
+        case GL_FOG_START:
+        case GL_FOG_END:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogx(pname=0x%x)", pname);
+            return;
+    }
+
+    if (convert_param_value) {
+        converted_param = (GLfloat) (param / 65536.0f);
+    } else {
+        converted_param = (GLfloat) param;
+    }
+
+    _mesa_Fogf(pname, converted_param);
+}
+
+extern void GL_APIENTRY _mesa_Fogfv(GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _es_Fogxv(GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+    int convert_params_value = 1;
+
+    switch(pname) {
+        case GL_FOG_MODE:
+            if (params[0] != GL_EXP && params[0] != GL_EXP2 && params[0] != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_FOG_COLOR:
+            n_params = 4;
+            break;
+        case GL_FOG_DENSITY:
+        case GL_FOG_START:
+        case GL_FOG_END:
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFogxv(pname=0x%x)", pname);
+            return;
+    }
+
+    if (convert_params_value) {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) (params[i] / 65536.0f);
+        }
+    } else {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) params[i];
+        }
+    }
+
+    _mesa_Fogfv(pname, converted_params);
+}
+
+extern void GL_APIENTRY _mesa_FramebufferRenderbufferEXT(GLenum target, GLenum attachment, GLenum renderbuffertarget, GLuint renderbuffer);
+static void GL_APIENTRY _es_FramebufferRenderbufferOES(GLenum target, GLenum attachment, GLenum renderbuffertarget, GLuint renderbuffer)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferRenderbufferOES(target=0x%x)", target);
+            return;
+    }
+    switch(attachment) {
+        case GL_COLOR_ATTACHMENT0_OES:
+        case GL_DEPTH_ATTACHMENT_OES:
+        case GL_STENCIL_ATTACHMENT_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferRenderbufferOES(attachment=0x%x)", attachment);
+            return;
+    }
+    switch(renderbuffertarget) {
+        case GL_RENDERBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferRenderbufferOES(renderbuffertarget=0x%x)", renderbuffertarget);
+            return;
+    }
+
+    _mesa_FramebufferRenderbufferEXT(target, attachment, renderbuffertarget, renderbuffer);
+}
+
+extern void GL_APIENTRY _mesa_FramebufferTexture2DEXT(GLenum target, GLenum attachment, GLenum textarget, GLuint texture, GLint level);
+static void GL_APIENTRY _es_FramebufferTexture2DOES(GLenum target, GLenum attachment, GLenum textarget, GLuint texture, GLint level)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferTexture2DOES(target=0x%x)", target);
+            return;
+    }
+    switch(attachment) {
+        case GL_COLOR_ATTACHMENT0_OES:
+        case GL_DEPTH_ATTACHMENT_OES:
+        case GL_STENCIL_ATTACHMENT_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferTexture2DOES(attachment=0x%x)", attachment);
+            return;
+    }
+    switch(textarget) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glFramebufferTexture2DOES(textarget=0x%x)", textarget);
+            return;
+    }
+
+    _mesa_FramebufferTexture2DEXT(target, attachment, textarget, texture, level);
+}
+
+extern void GL_APIENTRY _mesa_FrontFace(GLenum mode);
+static void GL_APIENTRY _es_FrontFace(GLenum mode)
+{
+    switch(mode) {
+        case GL_CW:
+        case GL_CCW:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFrontFace(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_FrontFace(mode);
+}
+
+extern void GL_APIENTRY _mesa_Frustum(GLdouble  left, GLdouble  right, GLdouble  bottom, GLdouble  top, GLdouble  zNear, GLdouble  zFar);
+static void GL_APIENTRY _es_Frustumf(GLfloat left, GLfloat right, GLfloat bottom, GLfloat top, GLfloat zNear, GLfloat zFar)
+{
+    GLdouble converted_left;
+    GLdouble converted_right;
+    GLdouble converted_bottom;
+    GLdouble converted_top;
+    GLdouble converted_zNear;
+    GLdouble converted_zFar;
+
+    converted_left = (GLdouble) (left);
+    converted_right = (GLdouble) (right);
+    converted_bottom = (GLdouble) (bottom);
+    converted_top = (GLdouble) (top);
+    converted_zNear = (GLdouble) (zNear);
+    converted_zFar = (GLdouble) (zFar);
+
+    _mesa_Frustum(converted_left, converted_right, converted_bottom, converted_top, converted_zNear, converted_zFar);
+}
+
+extern void GL_APIENTRY _mesa_Frustum(GLdouble  left, GLdouble  right, GLdouble  bottom, GLdouble  top, GLdouble  zNear, GLdouble  zFar);
+static void GL_APIENTRY _es_Frustumx(GLfixed left, GLfixed right, GLfixed bottom, GLfixed top, GLfixed zNear, GLfixed zFar)
+{
+    GLdouble converted_left;
+    GLdouble converted_right;
+    GLdouble converted_bottom;
+    GLdouble converted_top;
+    GLdouble converted_zNear;
+    GLdouble converted_zFar;
+
+    converted_left = (GLdouble) (left / 65536.0);
+    converted_right = (GLdouble) (right / 65536.0);
+    converted_bottom = (GLdouble) (bottom / 65536.0);
+    converted_top = (GLdouble) (top / 65536.0);
+    converted_zNear = (GLdouble) (zNear / 65536.0);
+    converted_zFar = (GLdouble) (zFar / 65536.0);
+
+    _mesa_Frustum(converted_left, converted_right, converted_bottom, converted_top, converted_zNear, converted_zFar);
+}
+
+extern void GL_APIENTRY _mesa_GenBuffersARB(GLsizei n, GLuint * buffer);
+#define _es_GenBuffers _mesa_GenBuffersARB
+
+extern void GL_APIENTRY _mesa_GenFramebuffersEXT(GLsizei n, GLuint * ids);
+#define _es_GenFramebuffersOES _mesa_GenFramebuffersEXT
+
+extern void GL_APIENTRY _mesa_GenRenderbuffersEXT(GLsizei n, GLuint * renderbuffers);
+#define _es_GenRenderbuffersOES _mesa_GenRenderbuffersEXT
+
+extern void GL_APIENTRY _mesa_GenTextures(GLsizei n, GLuint * textures);
+#define _es_GenTextures _mesa_GenTextures
+
+extern void GL_APIENTRY _mesa_GenerateMipmapEXT(GLenum target);
+static void GL_APIENTRY _es_GenerateMipmapOES(GLenum target)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGenerateMipmapOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_GenerateMipmapEXT(target);
+}
+
+extern void GL_APIENTRY _mesa_GetBooleanv(GLenum pname, GLboolean * params);
+#define _es_GetBooleanv _mesa_GetBooleanv
+
+extern void GL_APIENTRY _mesa_GetBufferParameterivARB(GLenum target, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetBufferParameteriv(GLenum target, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_BUFFER_SIZE:
+        case GL_BUFFER_USAGE:
+        case GL_BUFFER_ACCESS_OES:
+        case GL_BUFFER_MAPPED_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetBufferParameterivARB(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetBufferPointervARB(GLenum target, GLenum pname, GLvoid ** params);
+static void GL_APIENTRY _es_GetBufferPointervOES(GLenum target, GLenum pname, GLvoid **params)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferPointervOES(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_BUFFER_MAP_POINTER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferPointervOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetBufferPointervARB(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetClipPlane(GLenum plane, GLdouble * equation);
+static void GL_APIENTRY _es_GetClipPlanef(GLenum plane, GLfloat *equation)
+{
+    register unsigned int i;
+    unsigned int n_equation = 4;
+    GLdouble converted_equation[4];
+
+    switch(plane) {
+        case GL_CLIP_PLANE0:
+        case GL_CLIP_PLANE1:
+        case GL_CLIP_PLANE2:
+        case GL_CLIP_PLANE3:
+        case GL_CLIP_PLANE4:
+        case GL_CLIP_PLANE5:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetClipPlanef(plane=0x%x)", plane);
+            return;
+    }
+
+    _mesa_GetClipPlane(plane, converted_equation);
+    for (i = 0; i < n_equation; i++) {
+        equation[i] = (GLfloat) (converted_equation[i]);
+    }
+}
+
+extern void GL_APIENTRY _mesa_GetClipPlane(GLenum plane, GLdouble * equation);
+static void GL_APIENTRY _es_GetClipPlanex(GLenum plane, GLfixed *equation)
+{
+    register unsigned int i;
+    unsigned int n_equation = 4;
+    GLdouble converted_equation[4];
+
+    switch(plane) {
+        case GL_CLIP_PLANE0:
+        case GL_CLIP_PLANE1:
+        case GL_CLIP_PLANE2:
+        case GL_CLIP_PLANE3:
+        case GL_CLIP_PLANE4:
+        case GL_CLIP_PLANE5:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetClipPlanex(plane=0x%x)", plane);
+            return;
+    }
+
+    _mesa_GetClipPlane(plane, converted_equation);
+    for (i = 0; i < n_equation; i++) {
+        equation[i] = (GLfixed) (converted_equation[i] * 65536);
+    }
+}
+
+extern GLenum GL_APIENTRY _mesa_GetError(void);
+#define _es_GetError _mesa_GetError
+
+extern void GL_APIENTRY _mesa_GetFixedv(GLenum pname, GLfixed * params);
+#define _es_GetFixedv _mesa_GetFixedv
+
+extern void GL_APIENTRY _mesa_GetFloatv(GLenum pname, GLfloat * params);
+#define _es_GetFloatv _mesa_GetFloatv
+
+extern void GL_APIENTRY _mesa_GetFramebufferAttachmentParameterivEXT(GLenum target, GLenum attachment, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetFramebufferAttachmentParameterivOES(GLenum target, GLenum attachment, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetFramebufferAttachmentParameterivOES(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_FRAMEBUFFER_ATTACHMENT_OBJECT_TYPE_OES: /* size 1 */
+        case GL_FRAMEBUFFER_ATTACHMENT_OBJECT_NAME_OES: /* size 1 */
+        case GL_FRAMEBUFFER_ATTACHMENT_TEXTURE_LEVEL_OES: /* size 1 */
+        case GL_FRAMEBUFFER_ATTACHMENT_TEXTURE_CUBE_MAP_FACE_OES: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetFramebufferAttachmentParameterivOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetFramebufferAttachmentParameterivEXT(target, attachment, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetIntegerv(GLenum pname, GLint * params);
+#define _es_GetIntegerv _mesa_GetIntegerv
+
+extern void GL_APIENTRY _mesa_GetLightfv(GLenum light, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetLightfv(GLenum light, GLenum pname, GLfloat *params)
+{
+    switch(light) {
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetLightfv(light=0x%x)", light);
+            return;
+    }
+    switch(pname) {
+        case GL_AMBIENT: /* size 4 */
+        case GL_DIFFUSE: /* size 4 */
+        case GL_SPECULAR: /* size 4 */
+        case GL_POSITION: /* size 4 */
+        case GL_SPOT_DIRECTION: /* size 3 */
+        case GL_SPOT_EXPONENT: /* size 1 */
+        case GL_SPOT_CUTOFF: /* size 1 */
+        case GL_CONSTANT_ATTENUATION: /* size 1 */
+        case GL_LINEAR_ATTENUATION: /* size 1 */
+        case GL_QUADRATIC_ATTENUATION: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetLightfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetLightfv(light, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetLightfv(GLenum light, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetLightxv(GLenum light, GLenum pname, GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+
+    switch(light) {
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetLightxv(light=0x%x)", light);
+            return;
+    }
+    switch(pname) {
+        case GL_AMBIENT:
+        case GL_DIFFUSE:
+        case GL_SPECULAR:
+        case GL_POSITION:
+            n_params = 4;
+            break;
+        case GL_SPOT_DIRECTION:
+            n_params = 3;
+            break;
+        case GL_SPOT_EXPONENT:
+        case GL_SPOT_CUTOFF:
+        case GL_CONSTANT_ATTENUATION:
+        case GL_LINEAR_ATTENUATION:
+        case GL_QUADRATIC_ATTENUATION:
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetLightxv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetLightfv(light, pname, converted_params);
+    for (i = 0; i < n_params; i++) {
+        params[i] = (GLint) (converted_params[i] * 65536);
+    }
+}
+
+extern void GL_APIENTRY _mesa_GetMaterialfv(GLenum face, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetMaterialfv(GLenum face, GLenum pname, GLfloat *params)
+{
+    switch(face) {
+        case GL_FRONT:
+        case GL_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetMaterialfv(face=0x%x)", face);
+            return;
+    }
+    switch(pname) {
+        case GL_SHININESS: /* size 1 */
+        case GL_AMBIENT: /* size 4 */
+        case GL_DIFFUSE: /* size 4 */
+        case GL_AMBIENT_AND_DIFFUSE: /* size 4 */
+        case GL_SPECULAR: /* size 4 */
+        case GL_EMISSION: /* size 4 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetMaterialfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetMaterialfv(face, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetMaterialfv(GLenum face, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetMaterialxv(GLenum face, GLenum pname, GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+
+    switch(face) {
+        case GL_FRONT:
+        case GL_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetMaterialxv(face=0x%x)", face);
+            return;
+    }
+    switch(pname) {
+        case GL_SHININESS:
+            n_params = 1;
+            break;
+        case GL_AMBIENT:
+        case GL_DIFFUSE:
+        case GL_AMBIENT_AND_DIFFUSE:
+        case GL_SPECULAR:
+        case GL_EMISSION:
+            n_params = 4;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetMaterialxv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetMaterialfv(face, pname, converted_params);
+    for (i = 0; i < n_params; i++) {
+        params[i] = (GLint) (converted_params[i] * 65536);
+    }
+}
+
+extern void GL_APIENTRY _mesa_GetPointerv(GLenum pname, GLvoid ** params);
+static void GL_APIENTRY _es_GetPointerv(GLenum pname, GLvoid **params)
+{
+    switch(pname) {
+        case GL_VERTEX_ARRAY_POINTER:
+        case GL_NORMAL_ARRAY_POINTER:
+        case GL_COLOR_ARRAY_POINTER:
+        case GL_TEXTURE_COORD_ARRAY_POINTER:
+        case GL_MATRIX_INDEX_ARRAY_POINTER_OES:
+        case GL_WEIGHT_ARRAY_POINTER_OES:
+        case GL_POINT_SIZE_ARRAY_POINTER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetPointerv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetPointerv(pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetRenderbufferParameterivEXT(GLenum target, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetRenderbufferParameterivOES(GLenum target, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_RENDERBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetRenderbufferParameterivOES(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_RENDERBUFFER_WIDTH_OES: /* size 1 */
+        case GL_RENDERBUFFER_HEIGHT_OES: /* size 1 */
+        case GL_RENDERBUFFER_INTERNAL_FORMAT_OES: /* size 1 */
+        case GL_RENDERBUFFER_RED_SIZE_OES: /* size 1 */
+        case GL_RENDERBUFFER_GREEN_SIZE_OES: /* size 1 */
+        case GL_RENDERBUFFER_BLUE_SIZE_OES: /* size 1 */
+        case GL_RENDERBUFFER_ALPHA_SIZE_OES: /* size 1 */
+        case GL_RENDERBUFFER_DEPTH_SIZE_OES: /* size 1 */
+        case GL_RENDERBUFFER_STENCIL_SIZE_OES: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetRenderbufferParameterivOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetRenderbufferParameterivEXT(target, pname, params);
+}
+
+extern const GLubyte * GL_APIENTRY _mesa_GetString(GLenum name);
+static const GLubyte * GL_APIENTRY _es_GetString(GLenum name)
+{
+    switch(name) {
+        case GL_VENDOR:
+        case GL_RENDERER:
+        case GL_VERSION:
+        case GL_EXTENSIONS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetString(name=0x%x)", name);
+            return (const GLubyte *) 0;
+    }
+
+    return _mesa_GetString(name);
+}
+
+extern void GL_APIENTRY _mesa_GetTexEnvfv(GLenum target, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetTexEnvfv(GLenum target, GLenum pname, GLfloat *params)
+{
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvfv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvfv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_COLOR && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvfv(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvfv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES: /* size 1 */
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT: /* size 1 */
+        case GL_TEXTURE_ENV_COLOR: /* size 4 */
+        case GL_RGB_SCALE: /* size 1 */
+        case GL_ALPHA_SCALE: /* size 1 */
+            break;
+        case GL_TEXTURE_ENV_MODE: /* size 1 */
+        case GL_COMBINE_RGB: /* size 1 */
+        case GL_COMBINE_ALPHA: /* size 1 */
+        case GL_SRC0_RGB: /* size 1 */
+        case GL_SRC1_RGB: /* size 1 */
+        case GL_SRC2_RGB: /* size 1 */
+        case GL_SRC0_ALPHA: /* size 1 */
+        case GL_SRC1_ALPHA: /* size 1 */
+        case GL_SRC2_ALPHA: /* size 1 */
+        case GL_OPERAND0_RGB: /* size 1 */
+        case GL_OPERAND1_RGB: /* size 1 */
+        case GL_OPERAND2_RGB: /* size 1 */
+        case GL_OPERAND0_ALPHA: /* size 1 */
+        case GL_OPERAND1_ALPHA: /* size 1 */
+        case GL_OPERAND2_ALPHA: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexEnvfv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetTexEnviv(GLenum target, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetTexEnviv(GLenum target, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnviv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnviv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_COLOR && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnviv(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnviv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES: /* size 1 */
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT: /* size 1 */
+        case GL_TEXTURE_ENV_COLOR: /* size 4 */
+        case GL_RGB_SCALE: /* size 1 */
+        case GL_ALPHA_SCALE: /* size 1 */
+            break;
+        case GL_TEXTURE_ENV_MODE: /* size 1 */
+        case GL_COMBINE_RGB: /* size 1 */
+        case GL_COMBINE_ALPHA: /* size 1 */
+        case GL_SRC0_RGB: /* size 1 */
+        case GL_SRC1_RGB: /* size 1 */
+        case GL_SRC2_RGB: /* size 1 */
+        case GL_SRC0_ALPHA: /* size 1 */
+        case GL_SRC1_ALPHA: /* size 1 */
+        case GL_SRC2_ALPHA: /* size 1 */
+        case GL_OPERAND0_RGB: /* size 1 */
+        case GL_OPERAND1_RGB: /* size 1 */
+        case GL_OPERAND2_RGB: /* size 1 */
+        case GL_OPERAND0_ALPHA: /* size 1 */
+        case GL_OPERAND1_ALPHA: /* size 1 */
+        case GL_OPERAND2_ALPHA: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnviv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexEnviv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetTexEnvfv(GLenum target, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetTexEnvxv(GLenum target, GLenum pname, GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+    int convert_params_value = 1;
+
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvxv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvxv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_COLOR && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvxv(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvxv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES:
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT:
+            n_params = 1;
+            break;
+        case GL_TEXTURE_ENV_COLOR:
+            n_params = 4;
+            break;
+        case GL_RGB_SCALE:
+        case GL_ALPHA_SCALE:
+            n_params = 1;
+            break;
+        case GL_TEXTURE_ENV_MODE:
+        case GL_COMBINE_RGB:
+        case GL_COMBINE_ALPHA:
+        case GL_SRC0_RGB:
+        case GL_SRC1_RGB:
+        case GL_SRC2_RGB:
+        case GL_SRC0_ALPHA:
+        case GL_SRC1_ALPHA:
+        case GL_SRC2_ALPHA:
+        case GL_OPERAND0_RGB:
+        case GL_OPERAND1_RGB:
+        case GL_OPERAND2_RGB:
+        case GL_OPERAND0_ALPHA:
+        case GL_OPERAND1_ALPHA:
+        case GL_OPERAND2_ALPHA:
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexEnvxv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexEnvfv(target, pname, converted_params);
+    if (convert_params_value) {
+        for (i = 0; i < n_params; i++) {
+            params[i] = (GLint) (converted_params[i] * 65536);
+        }
+    } else {
+        for (i = 0; i < n_params; i++) {
+            params[i] = (GLfixed) converted_params[i];
+        }
+    }
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_GetTexGenfv(GLenum coord, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _check_GetTexGenfvOES(GLenum coord, GLenum pname, GLfloat *params)
+{
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexGenfvOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexGenfvOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _es_GetTexGenfv(coord, pname, params);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_GetTexGenfv(GLenum coord, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _check_GetTexGenivOES(GLenum coord, GLenum pname, GLint *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 1;
+    GLfloat converted_params[1];
+
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexGenivOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES:
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexGenivOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _es_GetTexGenfv(coord, pname, converted_params);
+    for (i = 0; i < n_params; i++) {
+        params[i] = (GLfloat) converted_params[i];
+    }
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_GetTexGenfv(GLenum coord, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _check_GetTexGenxvOES(GLenum coord, GLenum pname, GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 1;
+    GLfloat converted_params[1];
+
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexGenxvOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES:
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexGenxvOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _es_GetTexGenfv(coord, pname, converted_params);
+    for (i = 0; i < n_params; i++) {
+        params[i] = (GLfloat) converted_params[i];
+    }
+}
+
+extern void GL_APIENTRY _mesa_GetTexParameterfv(GLenum target, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetTexParameterfv(GLenum target, GLenum pname, GLfloat *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameterfv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+        case GL_GENERATE_MIPMAP: /* size 1 */
+            break;
+        case GL_TEXTURE_CROP_RECT_OES: /* size 4 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameterfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexParameterfv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetTexParameteriv(GLenum target, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetTexParameteriv(GLenum target, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+        case GL_GENERATE_MIPMAP: /* size 1 */
+            break;
+        case GL_TEXTURE_CROP_RECT_OES: /* size 4 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexParameteriv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetTexParameterfv(GLenum target, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetTexParameterxv(GLenum target, GLenum pname, GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+    int convert_params_value = 1;
+
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameterxv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S:
+        case GL_TEXTURE_WRAP_T:
+        case GL_TEXTURE_MIN_FILTER:
+        case GL_TEXTURE_MAG_FILTER:
+        case GL_GENERATE_MIPMAP:
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_TEXTURE_CROP_RECT_OES:
+            n_params = 4;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameterxv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexParameterfv(target, pname, converted_params);
+    if (convert_params_value) {
+        for (i = 0; i < n_params; i++) {
+            params[i] = (GLint) (converted_params[i] * 65536);
+        }
+    } else {
+        for (i = 0; i < n_params; i++) {
+            params[i] = (GLfixed) converted_params[i];
+        }
+    }
+}
+
+extern void GL_APIENTRY _mesa_Hint(GLenum target, GLenum mode);
+static void GL_APIENTRY _es_Hint(GLenum target, GLenum mode)
+{
+    switch(target) {
+        case GL_FOG_HINT:
+        case GL_LINE_SMOOTH_HINT:
+        case GL_PERSPECTIVE_CORRECTION_HINT:
+        case GL_POINT_SMOOTH_HINT:
+        case GL_GENERATE_MIPMAP_HINT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glHint(target=0x%x)", target);
+            return;
+    }
+    switch(mode) {
+        case GL_FASTEST:
+        case GL_NICEST:
+        case GL_DONT_CARE:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glHint(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_Hint(target, mode);
+}
+
+extern GLboolean GL_APIENTRY _mesa_IsBufferARB(GLuint buffer);
+#define _es_IsBuffer _mesa_IsBufferARB
+
+extern GLboolean GL_APIENTRY _mesa_IsEnabled(GLenum cap);
+static GLboolean GL_APIENTRY _es_IsEnabled(GLenum cap)
+{
+    switch(cap) {
+        case GL_NORMALIZE:
+        case GL_RESCALE_NORMAL:
+        case GL_CLIP_PLANE0:
+        case GL_CLIP_PLANE1:
+        case GL_CLIP_PLANE2:
+        case GL_CLIP_PLANE3:
+        case GL_CLIP_PLANE4:
+        case GL_CLIP_PLANE5:
+        case GL_FOG:
+        case GL_LIGHTING:
+        case GL_COLOR_MATERIAL:
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+        case GL_POINT_SMOOTH:
+        case GL_LINE_SMOOTH:
+        case GL_CULL_FACE:
+        case GL_POLYGON_OFFSET_FILL:
+        case GL_MULTISAMPLE:
+        case GL_SAMPLE_ALPHA_TO_COVERAGE:
+        case GL_SAMPLE_ALPHA_TO_ONE:
+        case GL_SAMPLE_COVERAGE:
+        case GL_TEXTURE_2D:
+        case GL_SCISSOR_TEST:
+        case GL_ALPHA_TEST:
+        case GL_STENCIL_TEST:
+        case GL_DEPTH_TEST:
+        case GL_BLEND:
+        case GL_DITHER:
+        case GL_COLOR_LOGIC_OP:
+        case GL_POINT_SPRITE_OES:
+        case GL_TEXTURE_CUBE_MAP_OES:
+        case GL_TEXTURE_GEN_STR_OES:
+        case GL_VERTEX_ARRAY:
+        case GL_NORMAL_ARRAY:
+        case GL_COLOR_ARRAY:
+        case GL_TEXTURE_COORD_ARRAY:
+        case GL_MATRIX_INDEX_ARRAY_OES:
+        case GL_WEIGHT_ARRAY_OES:
+        case GL_POINT_SIZE_ARRAY_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glIsEnabled(cap=0x%x)", cap);
+            return GL_FALSE;
+    }
+
+    return _mesa_IsEnabled(cap);
+}
+
+extern GLboolean GL_APIENTRY _mesa_IsFramebufferEXT(GLuint framebuffer);
+#define _es_IsFramebufferOES _mesa_IsFramebufferEXT
+
+extern GLboolean GL_APIENTRY _mesa_IsRenderbufferEXT(GLuint renderbuffer);
+#define _es_IsRenderbufferOES _mesa_IsRenderbufferEXT
+
+extern GLboolean GL_APIENTRY _mesa_IsTexture(GLuint texture);
+#define _es_IsTexture _mesa_IsTexture
+
+extern void GL_APIENTRY _mesa_LightModelf(GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_LightModelf(GLenum pname, GLfloat param)
+{
+    switch(pname) {
+        case GL_LIGHT_MODEL_TWO_SIDE:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelf(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_LightModelf(pname, param);
+}
+
+extern void GL_APIENTRY _mesa_LightModelfv(GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_LightModelfv(GLenum pname, const GLfloat *params)
+{
+    switch(pname) {
+        case GL_LIGHT_MODEL_AMBIENT: /* size 4 */
+            break;
+        case GL_LIGHT_MODEL_TWO_SIDE: /* size 1 */
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_LightModelfv(pname, params);
+}
+
+extern void GL_APIENTRY _mesa_LightModelf(GLenum pname, GLfloat  param);
+static void GL_APIENTRY _es_LightModelx(GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+
+    switch(pname) {
+        case GL_LIGHT_MODEL_TWO_SIDE:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelx(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelx(pname=0x%x)", pname);
+            return;
+    }
+
+    converted_param = (GLfloat) param;
+
+    _mesa_LightModelf(pname, converted_param);
+}
+
+extern void GL_APIENTRY _mesa_LightModelfv(GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _es_LightModelxv(GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+    int convert_params_value = 1;
+
+    switch(pname) {
+        case GL_LIGHT_MODEL_AMBIENT:
+            n_params = 4;
+            break;
+        case GL_LIGHT_MODEL_TWO_SIDE:
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightModelxv(pname=0x%x)", pname);
+            return;
+    }
+
+    if (convert_params_value) {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) (params[i] / 65536.0f);
+        }
+    } else {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) params[i];
+        }
+    }
+
+    _mesa_LightModelfv(pname, converted_params);
+}
+
+extern void GL_APIENTRY _mesa_Lightf(GLenum light, GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_Lightf(GLenum light, GLenum pname, GLfloat param)
+{
+    switch(light) {
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightf(light=0x%x)", light);
+            return;
+    }
+    switch(pname) {
+        case GL_SPOT_EXPONENT:
+        case GL_SPOT_CUTOFF:
+        case GL_CONSTANT_ATTENUATION:
+        case GL_LINEAR_ATTENUATION:
+        case GL_QUADRATIC_ATTENUATION:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightf(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_Lightf(light, pname, param);
+}
+
+extern void GL_APIENTRY _mesa_Lightfv(GLenum light, GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_Lightfv(GLenum light, GLenum pname, const GLfloat *params)
+{
+    switch(light) {
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightfv(light=0x%x)", light);
+            return;
+    }
+    switch(pname) {
+        case GL_AMBIENT: /* size 4 */
+        case GL_DIFFUSE: /* size 4 */
+        case GL_SPECULAR: /* size 4 */
+        case GL_POSITION: /* size 4 */
+        case GL_SPOT_DIRECTION: /* size 3 */
+        case GL_SPOT_EXPONENT: /* size 1 */
+        case GL_SPOT_CUTOFF: /* size 1 */
+        case GL_CONSTANT_ATTENUATION: /* size 1 */
+        case GL_LINEAR_ATTENUATION: /* size 1 */
+        case GL_QUADRATIC_ATTENUATION: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_Lightfv(light, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_Lightf(GLenum light, GLenum pname, GLfloat  param);
+static void GL_APIENTRY _es_Lightx(GLenum light, GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+
+    switch(light) {
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightx(light=0x%x)", light);
+            return;
+    }
+    switch(pname) {
+        case GL_SPOT_EXPONENT:
+        case GL_SPOT_CUTOFF:
+        case GL_CONSTANT_ATTENUATION:
+        case GL_LINEAR_ATTENUATION:
+        case GL_QUADRATIC_ATTENUATION:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightx(pname=0x%x)", pname);
+            return;
+    }
+
+    converted_param = (GLfloat) (param / 65536.0f);
+
+    _mesa_Lightf(light, pname, converted_param);
+}
+
+extern void GL_APIENTRY _mesa_Lightfv(GLenum light, GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _es_Lightxv(GLenum light, GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+
+    switch(light) {
+        case GL_LIGHT0:
+        case GL_LIGHT1:
+        case GL_LIGHT2:
+        case GL_LIGHT3:
+        case GL_LIGHT4:
+        case GL_LIGHT5:
+        case GL_LIGHT6:
+        case GL_LIGHT7:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightxv(light=0x%x)", light);
+            return;
+    }
+    switch(pname) {
+        case GL_AMBIENT:
+        case GL_DIFFUSE:
+        case GL_SPECULAR:
+        case GL_POSITION:
+            n_params = 4;
+            break;
+        case GL_SPOT_DIRECTION:
+            n_params = 3;
+            break;
+        case GL_SPOT_EXPONENT:
+        case GL_SPOT_CUTOFF:
+        case GL_CONSTANT_ATTENUATION:
+        case GL_LINEAR_ATTENUATION:
+        case GL_QUADRATIC_ATTENUATION:
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLightxv(pname=0x%x)", pname);
+            return;
+    }
+
+    for (i = 0; i < n_params; i++) {
+        converted_params[i] = (GLfloat) (params[i] / 65536.0f);
+    }
+
+    _mesa_Lightfv(light, pname, converted_params);
+}
+
+extern void GL_APIENTRY _mesa_LineWidth(GLfloat width);
+#define _es_LineWidth _mesa_LineWidth
+
+extern void GL_APIENTRY _mesa_LineWidth(GLfloat  width);
+static void GL_APIENTRY _es_LineWidthx(GLfixed width)
+{
+    GLfloat converted_width;
+
+    converted_width = (GLfloat) (width / 65536.0f);
+
+    _mesa_LineWidth(converted_width);
+}
+
+extern void GL_APIENTRY _mesa_LoadIdentity(void);
+#define _es_LoadIdentity _mesa_LoadIdentity
+
+extern void GL_APIENTRY _mesa_LoadMatrixf(const GLfloat * m);
+#define _es_LoadMatrixf _mesa_LoadMatrixf
+
+extern void GL_APIENTRY _mesa_LoadMatrixf(GLfloat const * m);
+static void GL_APIENTRY _es_LoadMatrixx(const GLfixed *m)
+{
+    register unsigned int i;
+    unsigned int n_m = 16;
+    GLfloat converted_m[16];
+
+    for (i = 0; i < n_m; i++) {
+        converted_m[i] = (GLfloat) (m[i] / 65536.0f);
+    }
+
+    _mesa_LoadMatrixf(converted_m);
+}
+
+extern void GL_APIENTRY _mesa_LogicOp(GLenum opcode);
+static void GL_APIENTRY _es_LogicOp(GLenum opcode)
+{
+    switch(opcode) {
+        case GL_CLEAR:
+        case GL_SET:
+        case GL_COPY:
+        case GL_COPY_INVERTED:
+        case GL_NOOP:
+        case GL_INVERT:
+        case GL_AND:
+        case GL_NAND:
+        case GL_OR:
+        case GL_NOR:
+        case GL_XOR:
+        case GL_EQUIV:
+        case GL_AND_REVERSE:
+        case GL_AND_INVERTED:
+        case GL_OR_REVERSE:
+        case GL_OR_INVERTED:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glLogicOp(opcode=0x%x)", opcode);
+            return;
+    }
+
+    _mesa_LogicOp(opcode);
+}
+
+extern void * GL_APIENTRY _mesa_MapBufferARB(GLenum target, GLenum access);
+static void * GL_APIENTRY _es_MapBufferOES(GLenum target, GLenum access)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMapBufferOES(target=0x%x)", target);
+            return (void *) 0;
+    }
+    switch(access) {
+        case GL_WRITE_ONLY_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMapBufferOES(access=0x%x)", access);
+            return (void *) 0;
+    }
+
+    return _mesa_MapBufferARB(target, access);
+}
+
+extern void GL_APIENTRY _vbo_Materialf(GLenum face, GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_Materialf(GLenum face, GLenum pname, GLfloat param)
+{
+    switch(face) {
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialf(face=0x%x)", face);
+            return;
+    }
+    switch(pname) {
+        case GL_SHININESS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialf(pname=0x%x)", pname);
+            return;
+    }
+
+    _vbo_Materialf(face, pname, param);
+}
+
+extern void GL_APIENTRY _vbo_Materialfv(GLenum face, GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_Materialfv(GLenum face, GLenum pname, const GLfloat *params)
+{
+    switch(face) {
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialfv(face=0x%x)", face);
+            return;
+    }
+    switch(pname) {
+        case GL_AMBIENT: /* size 4 */
+        case GL_DIFFUSE: /* size 4 */
+        case GL_AMBIENT_AND_DIFFUSE: /* size 4 */
+        case GL_SPECULAR: /* size 4 */
+        case GL_EMISSION: /* size 4 */
+        case GL_SHININESS: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _vbo_Materialfv(face, pname, params);
+}
+
+extern void GL_APIENTRY _vbo_Materialf(GLenum face, GLenum pname, GLfloat  param);
+static void GL_APIENTRY _es_Materialx(GLenum face, GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+
+    switch(face) {
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialx(face=0x%x)", face);
+            return;
+    }
+    switch(pname) {
+        case GL_SHININESS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialx(pname=0x%x)", pname);
+            return;
+    }
+
+    converted_param = (GLfloat) (param / 65536.0f);
+
+    _vbo_Materialf(face, pname, converted_param);
+}
+
+extern void GL_APIENTRY _vbo_Materialfv(GLenum face, GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _es_Materialxv(GLenum face, GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+
+    switch(face) {
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialxv(face=0x%x)", face);
+            return;
+    }
+    switch(pname) {
+        case GL_AMBIENT:
+        case GL_DIFFUSE:
+        case GL_AMBIENT_AND_DIFFUSE:
+        case GL_SPECULAR:
+        case GL_EMISSION:
+            n_params = 4;
+            break;
+        case GL_SHININESS:
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMaterialxv(pname=0x%x)", pname);
+            return;
+    }
+
+    for (i = 0; i < n_params; i++) {
+        converted_params[i] = (GLfloat) (params[i] / 65536.0f);
+    }
+
+    _vbo_Materialfv(face, pname, converted_params);
+}
+
+extern void GL_APIENTRY _mesa_MatrixMode(GLenum mode);
+static void GL_APIENTRY _es_MatrixMode(GLenum mode)
+{
+    switch(mode) {
+        case GL_MODELVIEW:
+        case GL_PROJECTION:
+        case GL_TEXTURE:
+        case GL_MATRIX_PALETTE_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMatrixMode(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_MatrixMode(mode);
+}
+
+extern void GL_APIENTRY _mesa_MultMatrixf(const GLfloat * m);
+#define _es_MultMatrixf _mesa_MultMatrixf
+
+extern void GL_APIENTRY _mesa_MultMatrixf(GLfloat const * m);
+static void GL_APIENTRY _es_MultMatrixx(const GLfixed *m)
+{
+    register unsigned int i;
+    unsigned int n_m = 16;
+    GLfloat converted_m[16];
+
+    for (i = 0; i < n_m; i++) {
+        converted_m[i] = (GLfloat) (m[i] / 65536.0f);
+    }
+
+    _mesa_MultMatrixf(converted_m);
+}
+
+extern void GL_APIENTRY _mesa_MultiDrawArraysEXT(GLenum mode, const GLint * first, const GLsizei * count, GLsizei primcount);
+static void GL_APIENTRY _es_MultiDrawArraysEXT(GLenum mode, const GLint *first, const GLsizei *count, GLsizei primcount)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiDrawArraysEXT(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_MultiDrawArraysEXT(mode, first, count, primcount);
+}
+
+extern void GL_APIENTRY _mesa_MultiDrawElementsEXT(GLenum mode, const GLsizei * count, GLenum type, const GLvoid ** indices, GLsizei primcount);
+static void GL_APIENTRY _es_MultiDrawElementsEXT(GLenum mode, const GLsizei *count, GLenum type, const GLvoid **indices, GLsizei primcount)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiDrawElementsEXT(mode=0x%x)", mode);
+            return;
+    }
+    switch(type) {
+        case GL_UNSIGNED_BYTE:
+        case GL_UNSIGNED_SHORT:
+        case GL_UNSIGNED_INT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiDrawElementsEXT(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_MultiDrawElementsEXT(mode, count, type, indices, primcount);
+}
+
+extern void GL_APIENTRY _vbo_MultiTexCoord4f(GLenum texture, GLfloat s, GLfloat t, GLfloat r, GLfloat q);
+static void GL_APIENTRY _es_MultiTexCoord4f(GLenum texture, GLfloat s, GLfloat t, GLfloat r, GLfloat q)
+{
+    switch(texture) {
+        case GL_TEXTURE0:
+        case GL_TEXTURE1:
+        case GL_TEXTURE2:
+        case GL_TEXTURE3:
+        case GL_TEXTURE4:
+        case GL_TEXTURE5:
+        case GL_TEXTURE6:
+        case GL_TEXTURE7:
+        case GL_TEXTURE8:
+        case GL_TEXTURE9:
+        case GL_TEXTURE10:
+        case GL_TEXTURE11:
+        case GL_TEXTURE12:
+        case GL_TEXTURE13:
+        case GL_TEXTURE14:
+        case GL_TEXTURE15:
+        case GL_TEXTURE16:
+        case GL_TEXTURE17:
+        case GL_TEXTURE18:
+        case GL_TEXTURE19:
+        case GL_TEXTURE20:
+        case GL_TEXTURE21:
+        case GL_TEXTURE22:
+        case GL_TEXTURE23:
+        case GL_TEXTURE24:
+        case GL_TEXTURE25:
+        case GL_TEXTURE26:
+        case GL_TEXTURE27:
+        case GL_TEXTURE28:
+        case GL_TEXTURE29:
+        case GL_TEXTURE30:
+        case GL_TEXTURE31:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiTexCoord4f(texture=0x%x)", texture);
+            return;
+    }
+
+    _vbo_MultiTexCoord4f(texture, s, t, r, q);
+}
+
+extern void GL_APIENTRY _vbo_MultiTexCoord4f(GLenum texture, GLfloat  s, GLfloat  t, GLfloat  r, GLfloat  q);
+static void GL_APIENTRY _es_MultiTexCoord4x(GLenum texture, GLfixed s, GLfixed t, GLfixed r, GLfixed q)
+{
+    GLfloat converted_s;
+    GLfloat converted_t;
+    GLfloat converted_r;
+    GLfloat converted_q;
+
+    switch(texture) {
+        case GL_TEXTURE0:
+        case GL_TEXTURE1:
+        case GL_TEXTURE2:
+        case GL_TEXTURE3:
+        case GL_TEXTURE4:
+        case GL_TEXTURE5:
+        case GL_TEXTURE6:
+        case GL_TEXTURE7:
+        case GL_TEXTURE8:
+        case GL_TEXTURE9:
+        case GL_TEXTURE10:
+        case GL_TEXTURE11:
+        case GL_TEXTURE12:
+        case GL_TEXTURE13:
+        case GL_TEXTURE14:
+        case GL_TEXTURE15:
+        case GL_TEXTURE16:
+        case GL_TEXTURE17:
+        case GL_TEXTURE18:
+        case GL_TEXTURE19:
+        case GL_TEXTURE20:
+        case GL_TEXTURE21:
+        case GL_TEXTURE22:
+        case GL_TEXTURE23:
+        case GL_TEXTURE24:
+        case GL_TEXTURE25:
+        case GL_TEXTURE26:
+        case GL_TEXTURE27:
+        case GL_TEXTURE28:
+        case GL_TEXTURE29:
+        case GL_TEXTURE30:
+        case GL_TEXTURE31:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiTexCoord4x(texture=0x%x)", texture);
+            return;
+    }
+
+    converted_s = (GLfloat) (s / 65536.0f);
+    converted_t = (GLfloat) (t / 65536.0f);
+    converted_r = (GLfloat) (r / 65536.0f);
+    converted_q = (GLfloat) (q / 65536.0f);
+
+    _vbo_MultiTexCoord4f(texture, converted_s, converted_t, converted_r, converted_q);
+}
+
+extern void GL_APIENTRY _vbo_Normal3f(GLfloat nx, GLfloat ny, GLfloat nz);
+#define _es_Normal3f _vbo_Normal3f
+
+extern void GL_APIENTRY _vbo_Normal3f(GLfloat  nx, GLfloat  ny, GLfloat  nz);
+static void GL_APIENTRY _es_Normal3x(GLfixed nx, GLfixed ny, GLfixed nz)
+{
+    GLfloat converted_nx;
+    GLfloat converted_ny;
+    GLfloat converted_nz;
+
+    converted_nx = (GLfloat) (nx / 65536.0f);
+    converted_ny = (GLfloat) (ny / 65536.0f);
+    converted_nz = (GLfloat) (nz / 65536.0f);
+
+    _vbo_Normal3f(converted_nx, converted_ny, converted_nz);
+}
+
+extern void GL_APIENTRY _mesa_NormalPointer(GLenum type, GLsizei stride, const GLvoid * pointer);
+static void GL_APIENTRY _es_NormalPointer(GLenum type, GLsizei stride, const GLvoid *pointer)
+{
+    switch(type) {
+        case GL_BYTE:
+        case GL_SHORT:
+        case GL_FLOAT:
+        case GL_FIXED:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glNormalPointer(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_NormalPointer(type, stride, pointer);
+}
+
+extern void GL_APIENTRY _mesa_Ortho(GLdouble  left, GLdouble  right, GLdouble  bottom, GLdouble  top, GLdouble  zNear, GLdouble  zFar);
+static void GL_APIENTRY _es_Orthof(GLfloat left, GLfloat right, GLfloat bottom, GLfloat top, GLfloat zNear, GLfloat zFar)
+{
+    GLdouble converted_left;
+    GLdouble converted_right;
+    GLdouble converted_bottom;
+    GLdouble converted_top;
+    GLdouble converted_zNear;
+    GLdouble converted_zFar;
+
+    converted_left = (GLdouble) (left);
+    converted_right = (GLdouble) (right);
+    converted_bottom = (GLdouble) (bottom);
+    converted_top = (GLdouble) (top);
+    converted_zNear = (GLdouble) (zNear);
+    converted_zFar = (GLdouble) (zFar);
+
+    _mesa_Ortho(converted_left, converted_right, converted_bottom, converted_top, converted_zNear, converted_zFar);
+}
+
+extern void GL_APIENTRY _mesa_Ortho(GLdouble  left, GLdouble  right, GLdouble  bottom, GLdouble  top, GLdouble  zNear, GLdouble  zFar);
+static void GL_APIENTRY _es_Orthox(GLfixed left, GLfixed right, GLfixed bottom, GLfixed top, GLfixed zNear, GLfixed zFar)
+{
+    GLdouble converted_left;
+    GLdouble converted_right;
+    GLdouble converted_bottom;
+    GLdouble converted_top;
+    GLdouble converted_zNear;
+    GLdouble converted_zFar;
+
+    converted_left = (GLdouble) (left / 65536.0);
+    converted_right = (GLdouble) (right / 65536.0);
+    converted_bottom = (GLdouble) (bottom / 65536.0);
+    converted_top = (GLdouble) (top / 65536.0);
+    converted_zNear = (GLdouble) (zNear / 65536.0);
+    converted_zFar = (GLdouble) (zFar / 65536.0);
+
+    _mesa_Ortho(converted_left, converted_right, converted_bottom, converted_top, converted_zNear, converted_zFar);
+}
+
+extern void GL_APIENTRY _mesa_PixelStorei(GLenum pname, GLint param);
+static void GL_APIENTRY _es_PixelStorei(GLenum pname, GLint param)
+{
+    switch(pname) {
+        case GL_PACK_ALIGNMENT:
+        case GL_UNPACK_ALIGNMENT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glPixelStorei(pname=0x%x)", pname);
+            return;
+    }
+    switch(param) {
+        case 1:
+        case 2:
+        case 4:
+        case 8:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glPixelStorei(param=%d)", param);
+            return;
+    }
+
+    _mesa_PixelStorei(pname, param);
+}
+
+extern void GL_APIENTRY _mesa_PointParameterf(GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_PointParameterf(GLenum pname, GLfloat param)
+{
+    switch(pname) {
+        case GL_POINT_SIZE_MIN:
+        case GL_POINT_SIZE_MAX:
+        case GL_POINT_FADE_THRESHOLD_SIZE:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glPointParameterf(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_PointParameterf(pname, param);
+}
+
+extern void GL_APIENTRY _mesa_PointParameterfv(GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_PointParameterfv(GLenum pname, const GLfloat *params)
+{
+    switch(pname) {
+        case GL_POINT_SIZE_MIN: /* size 1 */
+        case GL_POINT_SIZE_MAX: /* size 1 */
+        case GL_POINT_FADE_THRESHOLD_SIZE: /* size 1 */
+        case GL_POINT_DISTANCE_ATTENUATION: /* size 3 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glPointParameterfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_PointParameterfv(pname, params);
+}
+
+extern void GL_APIENTRY _mesa_PointParameterf(GLenum pname, GLfloat  param);
+static void GL_APIENTRY _es_PointParameterx(GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+
+    switch(pname) {
+        case GL_POINT_SIZE_MIN:
+        case GL_POINT_SIZE_MAX:
+        case GL_POINT_FADE_THRESHOLD_SIZE:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glPointParameterx(pname=0x%x)", pname);
+            return;
+    }
+
+    converted_param = (GLfloat) (param / 65536.0f);
+
+    _mesa_PointParameterf(pname, converted_param);
+}
+
+extern void GL_APIENTRY _mesa_PointParameterfv(GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _es_PointParameterxv(GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 3;
+    GLfloat converted_params[3];
+
+    switch(pname) {
+        case GL_POINT_SIZE_MIN:
+        case GL_POINT_SIZE_MAX:
+        case GL_POINT_FADE_THRESHOLD_SIZE:
+            n_params = 1;
+            break;
+        case GL_POINT_DISTANCE_ATTENUATION:
+            n_params = 3;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glPointParameterxv(pname=0x%x)", pname);
+            return;
+    }
+
+    for (i = 0; i < n_params; i++) {
+        converted_params[i] = (GLfloat) (params[i] / 65536.0f);
+    }
+
+    _mesa_PointParameterfv(pname, converted_params);
+}
+
+extern void GL_APIENTRY _mesa_PointSize(GLfloat size);
+#define _es_PointSize _mesa_PointSize
+
+extern void GL_APIENTRY _mesa_PointSizePointer(GLenum type, GLsizei stride, const GLvoid * pointer);
+static void GL_APIENTRY _es_PointSizePointerOES(GLenum type, GLsizei stride, const GLvoid *pointer)
+{
+    switch(type) {
+        case GL_FLOAT:
+        case GL_FIXED:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glPointSizePointerOES(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_PointSizePointer(type, stride, pointer);
+}
+
+extern void GL_APIENTRY _mesa_PointSize(GLfloat  size);
+static void GL_APIENTRY _es_PointSizex(GLfixed size)
+{
+    GLfloat converted_size;
+
+    converted_size = (GLfloat) (size / 65536.0f);
+
+    _mesa_PointSize(converted_size);
+}
+
+extern void GL_APIENTRY _mesa_PolygonOffset(GLfloat factor, GLfloat units);
+#define _es_PolygonOffset _mesa_PolygonOffset
+
+extern void GL_APIENTRY _mesa_PolygonOffset(GLfloat  factor, GLfloat  units);
+static void GL_APIENTRY _es_PolygonOffsetx(GLfixed factor, GLfixed units)
+{
+    GLfloat converted_factor;
+    GLfloat converted_units;
+
+    converted_factor = (GLfloat) (factor / 65536.0f);
+    converted_units = (GLfloat) (units / 65536.0f);
+
+    _mesa_PolygonOffset(converted_factor, converted_units);
+}
+
+extern void GL_APIENTRY _mesa_PopMatrix(void);
+#define _es_PopMatrix _mesa_PopMatrix
+
+extern void GL_APIENTRY _mesa_PushMatrix(void);
+#define _es_PushMatrix _mesa_PushMatrix
+
+/* this function is special and is defined elsewhere */
+extern GLbitfield GL_APIENTRY _es_QueryMatrixxOES(GLfixed * mantissa, GLint * exponent);
+#define _check_QueryMatrixxOES _es_QueryMatrixxOES
+
+extern void GL_APIENTRY _mesa_ReadPixels(GLint x, GLint y, GLsizei width, GLsizei height, GLenum format, GLenum type, GLvoid * pixels);
+static void GL_APIENTRY _es_ReadPixels(GLint x, GLint y, GLsizei width, GLsizei height, GLenum format, GLenum type, GLvoid *pixels)
+{
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_ReadPixels(x, y, width, height, format, type, pixels);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_RenderbufferStorageEXT(GLenum target, GLenum internalFormat, GLsizei width, GLsizei height);
+static void GL_APIENTRY _check_RenderbufferStorageOES(GLenum target, GLenum internalFormat, GLsizei width, GLsizei height)
+{
+    switch(target) {
+        case GL_RENDERBUFFER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glRenderbufferStorageOES(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_DEPTH_COMPONENT16_OES:
+        case GL_RGBA4_OES:
+        case GL_RGB5_A1_OES:
+        case GL_RGB565_OES:
+        case GL_STENCIL_INDEX8_OES:
+        case GL_DEPTH_COMPONENT24_OES:
+        case GL_DEPTH_COMPONENT32_OES:
+        case GL_RGB8_OES:
+        case GL_RGBA8_OES:
+        case GL_STENCIL_INDEX1_OES:
+        case GL_STENCIL_INDEX4_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glRenderbufferStorageOES(internalFormat=0x%x)", internalFormat);
+            return;
+    }
+
+    _es_RenderbufferStorageEXT(target, internalFormat, width, height);
+}
+
+extern void GL_APIENTRY _mesa_Rotatef(GLfloat angle, GLfloat x, GLfloat y, GLfloat z);
+#define _es_Rotatef _mesa_Rotatef
+
+extern void GL_APIENTRY _mesa_Rotatef(GLfloat  angle, GLfloat  x, GLfloat  y, GLfloat  z);
+static void GL_APIENTRY _es_Rotatex(GLfixed angle, GLfixed x, GLfixed y, GLfixed z)
+{
+    GLfloat converted_angle;
+    GLfloat converted_x;
+    GLfloat converted_y;
+    GLfloat converted_z;
+
+    converted_angle = (GLfloat) (angle / 65536.0f);
+    converted_x = (GLfloat) (x / 65536.0f);
+    converted_y = (GLfloat) (y / 65536.0f);
+    converted_z = (GLfloat) (z / 65536.0f);
+
+    _mesa_Rotatef(converted_angle, converted_x, converted_y, converted_z);
+}
+
+extern void GL_APIENTRY _mesa_SampleCoverageARB(GLclampf value, GLboolean invert);
+#define _es_SampleCoverage _mesa_SampleCoverageARB
+
+extern void GL_APIENTRY _mesa_SampleCoverageARB(GLclampf  value, GLboolean invert);
+static void GL_APIENTRY _es_SampleCoveragex(GLclampx value, GLboolean invert)
+{
+    GLclampf converted_value;
+
+    converted_value = (GLclampf) (value / 65536.0f);
+
+    _mesa_SampleCoverageARB(converted_value, invert);
+}
+
+extern void GL_APIENTRY _mesa_Scalef(GLfloat x, GLfloat y, GLfloat z);
+#define _es_Scalef _mesa_Scalef
+
+extern void GL_APIENTRY _mesa_Scalef(GLfloat  x, GLfloat  y, GLfloat  z);
+static void GL_APIENTRY _es_Scalex(GLfixed x, GLfixed y, GLfixed z)
+{
+    GLfloat converted_x;
+    GLfloat converted_y;
+    GLfloat converted_z;
+
+    converted_x = (GLfloat) (x / 65536.0f);
+    converted_y = (GLfloat) (y / 65536.0f);
+    converted_z = (GLfloat) (z / 65536.0f);
+
+    _mesa_Scalef(converted_x, converted_y, converted_z);
+}
+
+extern void GL_APIENTRY _mesa_Scissor(GLint x, GLint y, GLsizei width, GLsizei height);
+#define _es_Scissor _mesa_Scissor
+
+extern void GL_APIENTRY _mesa_ShadeModel(GLenum mode);
+static void GL_APIENTRY _es_ShadeModel(GLenum mode)
+{
+    switch(mode) {
+        case GL_FLAT:
+        case GL_SMOOTH:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glShadeModel(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_ShadeModel(mode);
+}
+
+extern void GL_APIENTRY _mesa_StencilFunc(GLenum func, GLint ref, GLuint mask);
+static void GL_APIENTRY _es_StencilFunc(GLenum func, GLint ref, GLuint mask)
+{
+    switch(func) {
+        case GL_NEVER:
+        case GL_LESS:
+        case GL_LEQUAL:
+        case GL_GREATER:
+        case GL_GEQUAL:
+        case GL_EQUAL:
+        case GL_NOTEQUAL:
+        case GL_ALWAYS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilFunc(func=0x%x)", func);
+            return;
+    }
+
+    _mesa_StencilFunc(func, ref, mask);
+}
+
+extern void GL_APIENTRY _mesa_StencilMask(GLuint mask);
+#define _es_StencilMask _mesa_StencilMask
+
+extern void GL_APIENTRY _mesa_StencilOp(GLenum fail, GLenum zfail, GLenum zpass);
+static void GL_APIENTRY _es_StencilOp(GLenum fail, GLenum zfail, GLenum zpass)
+{
+    switch(fail) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP_OES:
+        case GL_DECR_WRAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOp(fail=0x%x)", fail);
+            return;
+    }
+    switch(zfail) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP_OES:
+        case GL_DECR_WRAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOp(zfail=0x%x)", zfail);
+            return;
+    }
+    switch(zpass) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP_OES:
+        case GL_DECR_WRAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOp(zpass=0x%x)", zpass);
+            return;
+    }
+
+    _mesa_StencilOp(fail, zfail, zpass);
+}
+
+extern void GL_APIENTRY _mesa_TexCoordPointer(GLint size, GLenum type, GLsizei stride, const GLvoid * pointer);
+static void GL_APIENTRY _es_TexCoordPointer(GLint size, GLenum type, GLsizei stride, const GLvoid *pointer)
+{
+    switch(size) {
+        case 2:
+        case 3:
+        case 4:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexCoordPointer(size=%d)", size);
+            return;
+    }
+    switch(type) {
+        case GL_BYTE:
+        case GL_SHORT:
+        case GL_FLOAT:
+        case GL_FIXED:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexCoordPointer(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_TexCoordPointer(size, type, stride, pointer);
+}
+
+extern void GL_APIENTRY _mesa_TexEnvf(GLenum target, GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_TexEnvf(GLenum target, GLenum pname, GLfloat param)
+{
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA && pname != GL_TEXTURE_ENV_COLOR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT:
+            break;
+        case GL_TEXTURE_ENV_MODE:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_DECAL && param != GL_BLEND && param != GL_ADD && param != GL_COMBINE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_RGB:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_ADD && param != GL_ADD_SIGNED && param != GL_INTERPOLATE && param != GL_SUBTRACT && param != GL_DOT3_RGB && param != GL_DOT3_RGBA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_ALPHA:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_ADD && param != GL_ADD_SIGNED && param != GL_INTERPOLATE && param != GL_SUBTRACT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_RGB_SCALE:
+        case GL_ALPHA_SCALE:
+            if (param != 1.0 && param != 2.0 && param != 4.0) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_SRC0_RGB:
+        case GL_SRC1_RGB:
+        case GL_SRC2_RGB:
+        case GL_SRC0_ALPHA:
+        case GL_SRC1_ALPHA:
+        case GL_SRC2_ALPHA:
+            if (param != GL_TEXTURE && param != GL_CONSTANT && param != GL_PRIMARY_COLOR && param != GL_PREVIOUS && param != GL_TEXTURE0 && param != GL_TEXTURE1 && param != GL_TEXTURE2 && param != GL_TEXTURE3 && param != GL_TEXTURE4 && param != GL_TEXTURE5 && param != GL_TEXTURE6 && param != GL_TEXTURE7 && param != GL_TEXTURE8 && param != GL_TEXTURE9 && param != GL_TEXTURE10 && param != GL_TEXTURE11 && param != GL_TEXTURE12 && param != GL_TEXTURE13 && param != GL_TEXTURE14 && param != GL_TEXTURE15 && param != GL_TEXTURE16 && param != GL_TEXTURE17 && param != GL_TEXTURE18 && param != GL_TEXTURE19 && param != GL_TEXTURE20 && param != GL_TEXTURE21 && param != GL_TEXTURE22 && param != GL_TEXTURE23 && param != GL_TEXTURE24 && param != GL_TEXTURE25 && param != GL_TEXTURE26 && param != GL_TEXTURE27 && param != GL_TEXTURE28 && param != GL_TEXTURE29 && param != GL_TEXTURE30 && param != GL_TEXTURE31) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_RGB:
+        case GL_OPERAND1_RGB:
+        case GL_OPERAND2_RGB:
+            if (param != GL_SRC_COLOR && param != GL_ONE_MINUS_SRC_COLOR && param != GL_SRC_ALPHA && param != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_ALPHA:
+        case GL_OPERAND1_ALPHA:
+        case GL_OPERAND2_ALPHA:
+            if (param != GL_SRC_ALPHA && param != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvf(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexEnvf(target, pname, param);
+}
+
+extern void GL_APIENTRY _mesa_TexEnvfv(GLenum target, GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_TexEnvfv(GLenum target, GLenum pname, const GLfloat *params)
+{
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA && pname != GL_TEXTURE_ENV_COLOR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES: /* size 1 */
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT: /* size 1 */
+            break;
+        case GL_TEXTURE_ENV_MODE: /* size 1 */
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_DECAL && params[0] != GL_BLEND && params[0] != GL_ADD && params[0] != GL_COMBINE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_RGB: /* size 1 */
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_ADD && params[0] != GL_ADD_SIGNED && params[0] != GL_INTERPOLATE && params[0] != GL_SUBTRACT && params[0] != GL_DOT3_RGB && params[0] != GL_DOT3_RGBA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_ALPHA: /* size 1 */
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_ADD && params[0] != GL_ADD_SIGNED && params[0] != GL_INTERPOLATE && params[0] != GL_SUBTRACT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_RGB_SCALE:
+        case GL_ALPHA_SCALE:
+            if (params[0] != 1.0 && params[0] != 2.0 && params[0] != 4.0) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_SRC0_RGB: /* size 1 */
+        case GL_SRC1_RGB: /* size 1 */
+        case GL_SRC2_RGB: /* size 1 */
+        case GL_SRC0_ALPHA: /* size 1 */
+        case GL_SRC1_ALPHA: /* size 1 */
+        case GL_SRC2_ALPHA: /* size 1 */
+            if (params[0] != GL_TEXTURE && params[0] != GL_CONSTANT && params[0] != GL_PRIMARY_COLOR && params[0] != GL_PREVIOUS && params[0] != GL_TEXTURE0 && params[0] != GL_TEXTURE1 && params[0] != GL_TEXTURE2 && params[0] != GL_TEXTURE3 && params[0] != GL_TEXTURE4 && params[0] != GL_TEXTURE5 && params[0] != GL_TEXTURE6 && params[0] != GL_TEXTURE7 && params[0] != GL_TEXTURE8 && params[0] != GL_TEXTURE9 && params[0] != GL_TEXTURE10 && params[0] != GL_TEXTURE11 && params[0] != GL_TEXTURE12 && params[0] != GL_TEXTURE13 && params[0] != GL_TEXTURE14 && params[0] != GL_TEXTURE15 && params[0] != GL_TEXTURE16 && params[0] != GL_TEXTURE17 && params[0] != GL_TEXTURE18 && params[0] != GL_TEXTURE19 && params[0] != GL_TEXTURE20 && params[0] != GL_TEXTURE21 && params[0] != GL_TEXTURE22 && params[0] != GL_TEXTURE23 && params[0] != GL_TEXTURE24 && params[0] != GL_TEXTURE25 && params[0] != GL_TEXTURE26 && params[0] != GL_TEXTURE27 && params[0] != GL_TEXTURE28 && params[0] != GL_TEXTURE29 && params[0] != GL_TEXTURE30 && params[0] != GL_TEXTURE31) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_RGB: /* size 1 */
+        case GL_OPERAND1_RGB: /* size 1 */
+        case GL_OPERAND2_RGB: /* size 1 */
+            if (params[0] != GL_SRC_COLOR && params[0] != GL_ONE_MINUS_SRC_COLOR && params[0] != GL_SRC_ALPHA && params[0] != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_ALPHA: /* size 1 */
+        case GL_OPERAND1_ALPHA: /* size 1 */
+        case GL_OPERAND2_ALPHA: /* size 1 */
+            if (params[0] != GL_SRC_ALPHA && params[0] != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV_COLOR: /* size 4 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexEnvfv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_TexEnvi(GLenum target, GLenum pname, GLint param);
+static void GL_APIENTRY _es_TexEnvi(GLenum target, GLenum pname, GLint param)
+{
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA && pname != GL_TEXTURE_ENV_COLOR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT:
+            break;
+        case GL_TEXTURE_ENV_MODE:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_DECAL && param != GL_BLEND && param != GL_ADD && param != GL_COMBINE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_RGB:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_ADD && param != GL_ADD_SIGNED && param != GL_INTERPOLATE && param != GL_SUBTRACT && param != GL_DOT3_RGB && param != GL_DOT3_RGBA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_ALPHA:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_ADD && param != GL_ADD_SIGNED && param != GL_INTERPOLATE && param != GL_SUBTRACT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_RGB_SCALE:
+        case GL_ALPHA_SCALE:
+            if (param != 1.0 && param != 2.0 && param != 4.0) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_SRC0_RGB:
+        case GL_SRC1_RGB:
+        case GL_SRC2_RGB:
+        case GL_SRC0_ALPHA:
+        case GL_SRC1_ALPHA:
+        case GL_SRC2_ALPHA:
+            if (param != GL_TEXTURE && param != GL_CONSTANT && param != GL_PRIMARY_COLOR && param != GL_PREVIOUS && param != GL_TEXTURE0 && param != GL_TEXTURE1 && param != GL_TEXTURE2 && param != GL_TEXTURE3 && param != GL_TEXTURE4 && param != GL_TEXTURE5 && param != GL_TEXTURE6 && param != GL_TEXTURE7 && param != GL_TEXTURE8 && param != GL_TEXTURE9 && param != GL_TEXTURE10 && param != GL_TEXTURE11 && param != GL_TEXTURE12 && param != GL_TEXTURE13 && param != GL_TEXTURE14 && param != GL_TEXTURE15 && param != GL_TEXTURE16 && param != GL_TEXTURE17 && param != GL_TEXTURE18 && param != GL_TEXTURE19 && param != GL_TEXTURE20 && param != GL_TEXTURE21 && param != GL_TEXTURE22 && param != GL_TEXTURE23 && param != GL_TEXTURE24 && param != GL_TEXTURE25 && param != GL_TEXTURE26 && param != GL_TEXTURE27 && param != GL_TEXTURE28 && param != GL_TEXTURE29 && param != GL_TEXTURE30 && param != GL_TEXTURE31) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_RGB:
+        case GL_OPERAND1_RGB:
+        case GL_OPERAND2_RGB:
+            if (param != GL_SRC_COLOR && param != GL_ONE_MINUS_SRC_COLOR && param != GL_SRC_ALPHA && param != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_ALPHA:
+        case GL_OPERAND1_ALPHA:
+        case GL_OPERAND2_ALPHA:
+            if (param != GL_SRC_ALPHA && param != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvi(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexEnvi(target, pname, param);
+}
+
+extern void GL_APIENTRY _mesa_TexEnviv(GLenum target, GLenum pname, const GLint * params);
+static void GL_APIENTRY _es_TexEnviv(GLenum target, GLenum pname, const GLint *params)
+{
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA && pname != GL_TEXTURE_ENV_COLOR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES: /* size 1 */
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT: /* size 1 */
+            break;
+        case GL_TEXTURE_ENV_MODE: /* size 1 */
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_DECAL && params[0] != GL_BLEND && params[0] != GL_ADD && params[0] != GL_COMBINE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_RGB: /* size 1 */
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_ADD && params[0] != GL_ADD_SIGNED && params[0] != GL_INTERPOLATE && params[0] != GL_SUBTRACT && params[0] != GL_DOT3_RGB && params[0] != GL_DOT3_RGBA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_COMBINE_ALPHA: /* size 1 */
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_ADD && params[0] != GL_ADD_SIGNED && params[0] != GL_INTERPOLATE && params[0] != GL_SUBTRACT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_RGB_SCALE:
+        case GL_ALPHA_SCALE:
+            if (params[0] != 1.0 && params[0] != 2.0 && params[0] != 4.0) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_SRC0_RGB: /* size 1 */
+        case GL_SRC1_RGB: /* size 1 */
+        case GL_SRC2_RGB: /* size 1 */
+        case GL_SRC0_ALPHA: /* size 1 */
+        case GL_SRC1_ALPHA: /* size 1 */
+        case GL_SRC2_ALPHA: /* size 1 */
+            if (params[0] != GL_TEXTURE && params[0] != GL_CONSTANT && params[0] != GL_PRIMARY_COLOR && params[0] != GL_PREVIOUS && params[0] != GL_TEXTURE0 && params[0] != GL_TEXTURE1 && params[0] != GL_TEXTURE2 && params[0] != GL_TEXTURE3 && params[0] != GL_TEXTURE4 && params[0] != GL_TEXTURE5 && params[0] != GL_TEXTURE6 && params[0] != GL_TEXTURE7 && params[0] != GL_TEXTURE8 && params[0] != GL_TEXTURE9 && params[0] != GL_TEXTURE10 && params[0] != GL_TEXTURE11 && params[0] != GL_TEXTURE12 && params[0] != GL_TEXTURE13 && params[0] != GL_TEXTURE14 && params[0] != GL_TEXTURE15 && params[0] != GL_TEXTURE16 && params[0] != GL_TEXTURE17 && params[0] != GL_TEXTURE18 && params[0] != GL_TEXTURE19 && params[0] != GL_TEXTURE20 && params[0] != GL_TEXTURE21 && params[0] != GL_TEXTURE22 && params[0] != GL_TEXTURE23 && params[0] != GL_TEXTURE24 && params[0] != GL_TEXTURE25 && params[0] != GL_TEXTURE26 && params[0] != GL_TEXTURE27 && params[0] != GL_TEXTURE28 && params[0] != GL_TEXTURE29 && params[0] != GL_TEXTURE30 && params[0] != GL_TEXTURE31) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_RGB: /* size 1 */
+        case GL_OPERAND1_RGB: /* size 1 */
+        case GL_OPERAND2_RGB: /* size 1 */
+            if (params[0] != GL_SRC_COLOR && params[0] != GL_ONE_MINUS_SRC_COLOR && params[0] != GL_SRC_ALPHA && params[0] != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_OPERAND0_ALPHA: /* size 1 */
+        case GL_OPERAND1_ALPHA: /* size 1 */
+        case GL_OPERAND2_ALPHA: /* size 1 */
+            if (params[0] != GL_SRC_ALPHA && params[0] != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV_COLOR: /* size 4 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnviv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexEnviv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_TexEnvf(GLenum target, GLenum pname, GLfloat  param);
+static void GL_APIENTRY _es_TexEnvx(GLenum target, GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+    int convert_param_value = 1;
+
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA && pname != GL_TEXTURE_ENV_COLOR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT:
+            break;
+        case GL_TEXTURE_ENV_MODE:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_DECAL && param != GL_BLEND && param != GL_ADD && param != GL_COMBINE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_COMBINE_RGB:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_ADD && param != GL_ADD_SIGNED && param != GL_INTERPOLATE && param != GL_SUBTRACT && param != GL_DOT3_RGB && param != GL_DOT3_RGBA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_COMBINE_ALPHA:
+            if (param != GL_REPLACE && param != GL_MODULATE && param != GL_ADD && param != GL_ADD_SIGNED && param != GL_INTERPOLATE && param != GL_SUBTRACT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_RGB_SCALE:
+        case GL_ALPHA_SCALE:
+            if (param != 1.0 && param != 2.0 && param != 4.0) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_SRC0_RGB:
+        case GL_SRC1_RGB:
+        case GL_SRC2_RGB:
+        case GL_SRC0_ALPHA:
+        case GL_SRC1_ALPHA:
+        case GL_SRC2_ALPHA:
+            if (param != GL_TEXTURE && param != GL_CONSTANT && param != GL_PRIMARY_COLOR && param != GL_PREVIOUS && param != GL_TEXTURE0 && param != GL_TEXTURE1 && param != GL_TEXTURE2 && param != GL_TEXTURE3 && param != GL_TEXTURE4 && param != GL_TEXTURE5 && param != GL_TEXTURE6 && param != GL_TEXTURE7 && param != GL_TEXTURE8 && param != GL_TEXTURE9 && param != GL_TEXTURE10 && param != GL_TEXTURE11 && param != GL_TEXTURE12 && param != GL_TEXTURE13 && param != GL_TEXTURE14 && param != GL_TEXTURE15 && param != GL_TEXTURE16 && param != GL_TEXTURE17 && param != GL_TEXTURE18 && param != GL_TEXTURE19 && param != GL_TEXTURE20 && param != GL_TEXTURE21 && param != GL_TEXTURE22 && param != GL_TEXTURE23 && param != GL_TEXTURE24 && param != GL_TEXTURE25 && param != GL_TEXTURE26 && param != GL_TEXTURE27 && param != GL_TEXTURE28 && param != GL_TEXTURE29 && param != GL_TEXTURE30 && param != GL_TEXTURE31) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_OPERAND0_RGB:
+        case GL_OPERAND1_RGB:
+        case GL_OPERAND2_RGB:
+            if (param != GL_SRC_COLOR && param != GL_ONE_MINUS_SRC_COLOR && param != GL_SRC_ALPHA && param != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_OPERAND0_ALPHA:
+        case GL_OPERAND1_ALPHA:
+        case GL_OPERAND2_ALPHA:
+            if (param != GL_SRC_ALPHA && param != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvx(pname=0x%x)", pname);
+            return;
+    }
+
+    if (convert_param_value) {
+        converted_param = (GLfloat) (param / 65536.0f);
+    } else {
+        converted_param = (GLfloat) param;
+    }
+
+    _mesa_TexEnvf(target, pname, converted_param);
+}
+
+extern void GL_APIENTRY _mesa_TexEnvfv(GLenum target, GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _es_TexEnvxv(GLenum target, GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+    int convert_params_value = 1;
+
+    switch(target) {
+        case GL_POINT_SPRITE_OES:
+            if (pname != GL_COORD_REPLACE_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_FILTER_CONTROL_EXT:
+            if (pname != GL_TEXTURE_LOD_BIAS_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(target=0x%x)", target);
+                return;
+            }
+            break;
+        case GL_TEXTURE_ENV:
+            if (pname != GL_TEXTURE_ENV_MODE && pname != GL_COMBINE_RGB && pname != GL_COMBINE_ALPHA && pname != GL_RGB_SCALE && pname != GL_ALPHA_SCALE && pname != GL_SRC0_RGB && pname != GL_SRC1_RGB && pname != GL_SRC2_RGB && pname != GL_SRC0_ALPHA && pname != GL_SRC1_ALPHA && pname != GL_SRC2_ALPHA && pname != GL_OPERAND0_RGB && pname != GL_OPERAND1_RGB && pname != GL_OPERAND2_RGB && pname != GL_OPERAND0_ALPHA && pname != GL_OPERAND1_ALPHA && pname != GL_OPERAND2_ALPHA && pname != GL_TEXTURE_ENV_COLOR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(target=0x%x)", target);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_COORD_REPLACE_OES:
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_TEXTURE_LOD_BIAS_EXT:
+            n_params = 1;
+            break;
+        case GL_TEXTURE_ENV_MODE:
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_DECAL && params[0] != GL_BLEND && params[0] != GL_ADD && params[0] != GL_COMBINE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_COMBINE_RGB:
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_ADD && params[0] != GL_ADD_SIGNED && params[0] != GL_INTERPOLATE && params[0] != GL_SUBTRACT && params[0] != GL_DOT3_RGB && params[0] != GL_DOT3_RGBA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_COMBINE_ALPHA:
+            if (params[0] != GL_REPLACE && params[0] != GL_MODULATE && params[0] != GL_ADD && params[0] != GL_ADD_SIGNED && params[0] != GL_INTERPOLATE && params[0] != GL_SUBTRACT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_RGB_SCALE:
+        case GL_ALPHA_SCALE:
+            if (params[0] != 1.0 && params[0] != 2.0 && params[0] != 4.0) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_SRC0_RGB:
+        case GL_SRC1_RGB:
+        case GL_SRC2_RGB:
+        case GL_SRC0_ALPHA:
+        case GL_SRC1_ALPHA:
+        case GL_SRC2_ALPHA:
+            if (params[0] != GL_TEXTURE && params[0] != GL_CONSTANT && params[0] != GL_PRIMARY_COLOR && params[0] != GL_PREVIOUS && params[0] != GL_TEXTURE0 && params[0] != GL_TEXTURE1 && params[0] != GL_TEXTURE2 && params[0] != GL_TEXTURE3 && params[0] != GL_TEXTURE4 && params[0] != GL_TEXTURE5 && params[0] != GL_TEXTURE6 && params[0] != GL_TEXTURE7 && params[0] != GL_TEXTURE8 && params[0] != GL_TEXTURE9 && params[0] != GL_TEXTURE10 && params[0] != GL_TEXTURE11 && params[0] != GL_TEXTURE12 && params[0] != GL_TEXTURE13 && params[0] != GL_TEXTURE14 && params[0] != GL_TEXTURE15 && params[0] != GL_TEXTURE16 && params[0] != GL_TEXTURE17 && params[0] != GL_TEXTURE18 && params[0] != GL_TEXTURE19 && params[0] != GL_TEXTURE20 && params[0] != GL_TEXTURE21 && params[0] != GL_TEXTURE22 && params[0] != GL_TEXTURE23 && params[0] != GL_TEXTURE24 && params[0] != GL_TEXTURE25 && params[0] != GL_TEXTURE26 && params[0] != GL_TEXTURE27 && params[0] != GL_TEXTURE28 && params[0] != GL_TEXTURE29 && params[0] != GL_TEXTURE30 && params[0] != GL_TEXTURE31) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_OPERAND0_RGB:
+        case GL_OPERAND1_RGB:
+        case GL_OPERAND2_RGB:
+            if (params[0] != GL_SRC_COLOR && params[0] != GL_ONE_MINUS_SRC_COLOR && params[0] != GL_SRC_ALPHA && params[0] != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_OPERAND0_ALPHA:
+        case GL_OPERAND1_ALPHA:
+        case GL_OPERAND2_ALPHA:
+            if (params[0] != GL_SRC_ALPHA && params[0] != GL_ONE_MINUS_SRC_ALPHA) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_TEXTURE_ENV_COLOR:
+            n_params = 4;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexEnvxv(pname=0x%x)", pname);
+            return;
+    }
+
+    if (convert_params_value) {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) (params[i] / 65536.0f);
+        }
+    } else {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) params[i];
+        }
+    }
+
+    _mesa_TexEnvfv(target, pname, converted_params);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_TexGenf(GLenum coord, GLenum pname, GLfloat param);
+static void GL_APIENTRY _check_TexGenfOES(GLenum coord, GLenum pname, GLfloat param)
+{
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenfOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES:
+            if (param != GL_NORMAL_MAP_OES && param != GL_REFLECTION_MAP_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenfOES(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenfOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _es_TexGenf(coord, pname, param);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_TexGenfv(GLenum coord, GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _check_TexGenfvOES(GLenum coord, GLenum pname, const GLfloat *params)
+{
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenfvOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES: /* size 1 */
+            if (params[0] != GL_NORMAL_MAP_OES && params[0] != GL_REFLECTION_MAP_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenfvOES(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenfvOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _es_TexGenfv(coord, pname, params);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_TexGenf(GLenum coord, GLenum pname, GLfloat  param);
+static void GL_APIENTRY _check_TexGeniOES(GLenum coord, GLenum pname, GLint param)
+{
+    GLfloat converted_param;
+
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGeniOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES:
+            if (param != GL_NORMAL_MAP_OES && param != GL_REFLECTION_MAP_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGeniOES(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGeniOES(pname=0x%x)", pname);
+            return;
+    }
+
+    converted_param = (GLfloat) param;
+
+    _es_TexGenf(coord, pname, converted_param);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_TexGenfv(GLenum coord, GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _check_TexGenivOES(GLenum coord, GLenum pname, const GLint *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 1;
+    GLfloat converted_params[1];
+
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenivOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES:
+            if (params[0] != GL_NORMAL_MAP_OES && params[0] != GL_REFLECTION_MAP_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenivOES(pname=0x%x)", pname);
+                return;
+            }
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenivOES(pname=0x%x)", pname);
+            return;
+    }
+
+    for (i = 0; i < n_params; i++) {
+        converted_params[i] = (GLfloat) params[i];
+    }
+
+    _es_TexGenfv(coord, pname, converted_params);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_TexGenf(GLenum coord, GLenum pname, GLfloat  param);
+static void GL_APIENTRY _check_TexGenxOES(GLenum coord, GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenxOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES:
+            if (param != GL_NORMAL_MAP_OES && param != GL_REFLECTION_MAP_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenxOES(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenxOES(pname=0x%x)", pname);
+            return;
+    }
+
+    converted_param = (GLfloat) param;
+
+    _es_TexGenf(coord, pname, converted_param);
+}
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_TexGenfv(GLenum coord, GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _check_TexGenxvOES(GLenum coord, GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 1;
+    GLfloat converted_params[1];
+
+    switch(coord) {
+        case GL_TEXTURE_GEN_STR_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenxvOES(coord=0x%x)", coord);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_GEN_MODE_OES:
+            if (params[0] != GL_NORMAL_MAP_OES && params[0] != GL_REFLECTION_MAP_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenxvOES(pname=0x%x)", pname);
+                return;
+            }
+            n_params = 1;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexGenxvOES(pname=0x%x)", pname);
+            return;
+    }
+
+    for (i = 0; i < n_params; i++) {
+        converted_params[i] = (GLfloat) params[i];
+    }
+
+    _es_TexGenfv(coord, pname, converted_params);
+}
+
+extern void GL_APIENTRY _mesa_TexImage2D(GLenum target, GLint level, GLint internalFormat, GLsizei width, GLsizei height, GLint border, GLenum format, GLenum type, const GLvoid * pixels);
+static void GL_APIENTRY _es_TexImage2D(GLenum target, GLint level, GLint internalFormat, GLsizei width, GLsizei height, GLint border, GLenum format, GLenum type, const GLvoid *pixels)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_ALPHA:
+        case GL_RGB:
+        case GL_RGBA:
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexImage2D(internalFormat=%d)", internalFormat);
+            return;
+    }
+    switch(border) {
+        case 0:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexImage2D(border=%d)", border);
+            return;
+    }
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_TexImage2D(target, level, internalFormat, width, height, border, format, type, pixels);
+}
+
+extern void GL_APIENTRY _mesa_TexParameterf(GLenum target, GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_TexParameterf(GLenum target, GLenum pname, GLfloat param)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S:
+        case GL_TEXTURE_WRAP_T:
+            if (param != GL_CLAMP_TO_EDGE && param != GL_REPEAT && param != GL_MIRRORED_REPEAT_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR && param != GL_NEAREST_MIPMAP_NEAREST && param != GL_NEAREST_MIPMAP_LINEAR && param != GL_LINEAR_MIPMAP_NEAREST && param != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_GENERATE_MIPMAP:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameterf(target, pname, param);
+}
+
+extern void GL_APIENTRY _mesa_TexParameterfv(GLenum target, GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_TexParameterfv(GLenum target, GLenum pname, const GLfloat *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+            if (params[0] != GL_CLAMP_TO_EDGE && params[0] != GL_REPEAT && params[0] != GL_MIRRORED_REPEAT_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR && params[0] != GL_NEAREST_MIPMAP_NEAREST && params[0] != GL_NEAREST_MIPMAP_LINEAR && params[0] != GL_LINEAR_MIPMAP_NEAREST && params[0] != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_GENERATE_MIPMAP: /* size 1 */
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT: /* size 1 */
+        case GL_TEXTURE_CROP_RECT_OES: /* size 4 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameterfv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_TexParameteri(GLenum target, GLenum pname, GLint param);
+static void GL_APIENTRY _es_TexParameteri(GLenum target, GLenum pname, GLint param)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S:
+        case GL_TEXTURE_WRAP_T:
+            if (param != GL_CLAMP_TO_EDGE && param != GL_REPEAT && param != GL_MIRRORED_REPEAT_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR && param != GL_NEAREST_MIPMAP_NEAREST && param != GL_NEAREST_MIPMAP_LINEAR && param != GL_LINEAR_MIPMAP_NEAREST && param != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_GENERATE_MIPMAP:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameteri(target, pname, param);
+}
+
+extern void GL_APIENTRY _mesa_TexParameteriv(GLenum target, GLenum pname, const GLint * params);
+static void GL_APIENTRY _es_TexParameteriv(GLenum target, GLenum pname, const GLint *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+            if (params[0] != GL_CLAMP_TO_EDGE && params[0] != GL_REPEAT && params[0] != GL_MIRRORED_REPEAT_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR && params[0] != GL_NEAREST_MIPMAP_NEAREST && params[0] != GL_NEAREST_MIPMAP_LINEAR && params[0] != GL_LINEAR_MIPMAP_NEAREST && params[0] != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_GENERATE_MIPMAP: /* size 1 */
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT: /* size 1 */
+        case GL_TEXTURE_CROP_RECT_OES: /* size 4 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameteriv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_TexParameterf(GLenum target, GLenum pname, GLfloat  param);
+static void GL_APIENTRY _es_TexParameterx(GLenum target, GLenum pname, GLfixed param)
+{
+    GLfloat converted_param;
+    int convert_param_value = 1;
+
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterx(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S:
+        case GL_TEXTURE_WRAP_T:
+            if (param != GL_CLAMP_TO_EDGE && param != GL_REPEAT && param != GL_MIRRORED_REPEAT_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_TEXTURE_MIN_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR && param != GL_NEAREST_MIPMAP_NEAREST && param != GL_NEAREST_MIPMAP_LINEAR && param != GL_LINEAR_MIPMAP_NEAREST && param != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_TEXTURE_MAG_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_GENERATE_MIPMAP:
+            if (param != GL_TRUE && param != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterx(pname=0x%x)", pname);
+                return;
+            }
+            convert_param_value = 0;
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterx(pname=0x%x)", pname);
+            return;
+    }
+
+    if (convert_param_value) {
+        converted_param = (GLfloat) (param / 65536.0f);
+    } else {
+        converted_param = (GLfloat) param;
+    }
+
+    _mesa_TexParameterf(target, pname, converted_param);
+}
+
+extern void GL_APIENTRY _mesa_TexParameterfv(GLenum target, GLenum pname, GLfloat const * params);
+static void GL_APIENTRY _es_TexParameterxv(GLenum target, GLenum pname, const GLfixed *params)
+{
+    register unsigned int i;
+    unsigned int n_params = 4;
+    GLfloat converted_params[4];
+    int convert_params_value = 1;
+
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterxv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S:
+        case GL_TEXTURE_WRAP_T:
+            if (params[0] != GL_CLAMP_TO_EDGE && params[0] != GL_REPEAT && params[0] != GL_MIRRORED_REPEAT_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_TEXTURE_MIN_FILTER:
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR && params[0] != GL_NEAREST_MIPMAP_NEAREST && params[0] != GL_NEAREST_MIPMAP_LINEAR && params[0] != GL_LINEAR_MIPMAP_NEAREST && params[0] != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_TEXTURE_MAG_FILTER:
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_GENERATE_MIPMAP:
+            if (params[0] != GL_TRUE && params[0] != GL_FALSE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterxv(pname=0x%x)", pname);
+                return;
+            }
+            convert_params_value = 0;
+            n_params = 1;
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT:
+            n_params = 1;
+            break;
+        case GL_TEXTURE_CROP_RECT_OES:
+            n_params = 4;
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterxv(pname=0x%x)", pname);
+            return;
+    }
+
+    if (convert_params_value) {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) (params[i] / 65536.0f);
+        }
+    } else {
+        for (i = 0; i < n_params; i++) {
+            converted_params[i] = (GLfloat) params[i];
+        }
+    }
+
+    _mesa_TexParameterfv(target, pname, converted_params);
+}
+
+extern void GL_APIENTRY _mesa_TexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid * pixels);
+static void GL_APIENTRY _es_TexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid *pixels)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y_OES:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexSubImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_TexSubImage2D(target, level, xoffset, yoffset, width, height, format, type, pixels);
+}
+
+extern void GL_APIENTRY _mesa_Translatef(GLfloat x, GLfloat y, GLfloat z);
+#define _es_Translatef _mesa_Translatef
+
+extern void GL_APIENTRY _mesa_Translatef(GLfloat  x, GLfloat  y, GLfloat  z);
+static void GL_APIENTRY _es_Translatex(GLfixed x, GLfixed y, GLfixed z)
+{
+    GLfloat converted_x;
+    GLfloat converted_y;
+    GLfloat converted_z;
+
+    converted_x = (GLfloat) (x / 65536.0f);
+    converted_y = (GLfloat) (y / 65536.0f);
+    converted_z = (GLfloat) (z / 65536.0f);
+
+    _mesa_Translatef(converted_x, converted_y, converted_z);
+}
+
+extern GLboolean GL_APIENTRY _mesa_UnmapBufferARB(GLenum target);
+static GLboolean GL_APIENTRY _es_UnmapBufferOES(GLenum target)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glUnmapBufferOES(target=0x%x)", target);
+            return GL_FALSE;
+    }
+
+    return _mesa_UnmapBufferARB(target);
+}
+
+extern void GL_APIENTRY _mesa_VertexPointer(GLint size, GLenum type, GLsizei stride, const GLvoid * pointer);
+static void GL_APIENTRY _es_VertexPointer(GLint size, GLenum type, GLsizei stride, const GLvoid *pointer)
+{
+    switch(size) {
+        case 2:
+        case 3:
+        case 4:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glVertexPointer(size=%d)", size);
+            return;
+    }
+    switch(type) {
+        case GL_BYTE:
+        case GL_SHORT:
+        case GL_FLOAT:
+        case GL_FIXED:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glVertexPointer(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_VertexPointer(size, type, stride, pointer);
+}
+
+extern void GL_APIENTRY _mesa_Viewport(GLint x, GLint y, GLsizei width, GLsizei height);
+#define _es_Viewport _mesa_Viewport
+
+
+struct _glapi_table *
+_mesa_create_exec_table_es1(void)
+{
+   struct _glapi_table *exec;
+   exec = _mesa_alloc_dispatch_table(sizeof *exec);
+   if (exec == NULL)
+      return NULL;
+
+
+    SET_ActiveTexture(exec, _es_ActiveTexture);
+    SET_AlphaFunc(exec, _es_AlphaFunc);
+    SET_AlphaFuncx(exec, _es_AlphaFuncx);
+    SET_BindBuffer(exec, _es_BindBuffer);
+    SET_BindFramebufferOES(exec, _es_BindFramebufferOES);
+    SET_BindRenderbufferOES(exec, _es_BindRenderbufferOES);
+    SET_BindTexture(exec, _es_BindTexture);
+    SET_BlendEquationOES(exec, _es_BlendEquationOES);
+    SET_BlendEquationSeparateOES(exec, _es_BlendEquationSeparateOES);
+    SET_BlendFunc(exec, _es_BlendFunc);
+    SET_BlendFuncSeparateOES(exec, _es_BlendFuncSeparateOES);
+    SET_BufferData(exec, _es_BufferData);
+    SET_BufferSubData(exec, _es_BufferSubData);
+    SET_CheckFramebufferStatusOES(exec, _es_CheckFramebufferStatusOES);
+    SET_Clear(exec, _es_Clear);
+    SET_ClearColor(exec, _es_ClearColor);
+    SET_ClearColorx(exec, _es_ClearColorx);
+    SET_ClearDepthf(exec, _es_ClearDepthf);
+    SET_ClearDepthx(exec, _es_ClearDepthx);
+    SET_ClearStencil(exec, _es_ClearStencil);
+    SET_ClientActiveTexture(exec, _es_ClientActiveTexture);
+    SET_ClipPlanef(exec, _es_ClipPlanef);
+    SET_ClipPlanex(exec, _es_ClipPlanex);
+    SET_Color4f(exec, _es_Color4f);
+    SET_Color4ub(exec, _es_Color4ub);
+    SET_Color4x(exec, _es_Color4x);
+    SET_ColorMask(exec, _es_ColorMask);
+    SET_ColorPointer(exec, _es_ColorPointer);
+    SET_CompressedTexImage2D(exec, _es_CompressedTexImage2D);
+    SET_CompressedTexSubImage2D(exec, _es_CompressedTexSubImage2D);
+    SET_CopyTexImage2D(exec, _es_CopyTexImage2D);
+    SET_CopyTexSubImage2D(exec, _es_CopyTexSubImage2D);
+    SET_CullFace(exec, _es_CullFace);
+    SET_DeleteBuffers(exec, _es_DeleteBuffers);
+    SET_DeleteFramebuffersOES(exec, _es_DeleteFramebuffersOES);
+    SET_DeleteRenderbuffersOES(exec, _es_DeleteRenderbuffersOES);
+    SET_DeleteTextures(exec, _es_DeleteTextures);
+    SET_DepthFunc(exec, _es_DepthFunc);
+    SET_DepthMask(exec, _es_DepthMask);
+    SET_DepthRangef(exec, _es_DepthRangef);
+    SET_DepthRangex(exec, _es_DepthRangex);
+    SET_Disable(exec, _es_Disable);
+    SET_DisableClientState(exec, _es_DisableClientState);
+    SET_DrawArrays(exec, _es_DrawArrays);
+    SET_DrawElements(exec, _es_DrawElements);
+    SET_DrawTexfOES(exec, _es_DrawTexfOES);
+    SET_DrawTexfvOES(exec, _es_DrawTexfvOES);
+    SET_DrawTexiOES(exec, _es_DrawTexiOES);
+    SET_DrawTexivOES(exec, _es_DrawTexivOES);
+    SET_DrawTexsOES(exec, _es_DrawTexsOES);
+    SET_DrawTexsvOES(exec, _es_DrawTexsvOES);
+    SET_DrawTexxOES(exec, _es_DrawTexxOES);
+    SET_DrawTexxvOES(exec, _es_DrawTexxvOES);
+    SET_EGLImageTargetRenderbufferStorageOES(exec, _es_EGLImageTargetRenderbufferStorageOES);
+    SET_EGLImageTargetTexture2DOES(exec, _es_EGLImageTargetTexture2DOES);
+    SET_Enable(exec, _es_Enable);
+    SET_EnableClientState(exec, _es_EnableClientState);
+    SET_Finish(exec, _es_Finish);
+    SET_Flush(exec, _es_Flush);
+    SET_Fogf(exec, _es_Fogf);
+    SET_Fogfv(exec, _es_Fogfv);
+    SET_Fogx(exec, _es_Fogx);
+    SET_Fogxv(exec, _es_Fogxv);
+    SET_FramebufferRenderbufferOES(exec, _es_FramebufferRenderbufferOES);
+    SET_FramebufferTexture2DOES(exec, _es_FramebufferTexture2DOES);
+    SET_FrontFace(exec, _es_FrontFace);
+    SET_Frustumf(exec, _es_Frustumf);
+    SET_Frustumx(exec, _es_Frustumx);
+    SET_GenBuffers(exec, _es_GenBuffers);
+    SET_GenFramebuffersOES(exec, _es_GenFramebuffersOES);
+    SET_GenRenderbuffersOES(exec, _es_GenRenderbuffersOES);
+    SET_GenTextures(exec, _es_GenTextures);
+    SET_GenerateMipmapOES(exec, _es_GenerateMipmapOES);
+    SET_GetBooleanv(exec, _es_GetBooleanv);
+    SET_GetBufferParameteriv(exec, _es_GetBufferParameteriv);
+    SET_GetBufferPointervOES(exec, _es_GetBufferPointervOES);
+    SET_GetClipPlanef(exec, _es_GetClipPlanef);
+    SET_GetClipPlanex(exec, _es_GetClipPlanex);
+    SET_GetError(exec, _es_GetError);
+    SET_GetFixedv(exec, _es_GetFixedv);
+    SET_GetFloatv(exec, _es_GetFloatv);
+    SET_GetFramebufferAttachmentParameterivOES(exec, _es_GetFramebufferAttachmentParameterivOES);
+    SET_GetIntegerv(exec, _es_GetIntegerv);
+    SET_GetLightfv(exec, _es_GetLightfv);
+    SET_GetLightxv(exec, _es_GetLightxv);
+    SET_GetMaterialfv(exec, _es_GetMaterialfv);
+    SET_GetMaterialxv(exec, _es_GetMaterialxv);
+    SET_GetPointerv(exec, _es_GetPointerv);
+    SET_GetRenderbufferParameterivOES(exec, _es_GetRenderbufferParameterivOES);
+    SET_GetString(exec, _es_GetString);
+    SET_GetTexEnvfv(exec, _es_GetTexEnvfv);
+    SET_GetTexEnviv(exec, _es_GetTexEnviv);
+    SET_GetTexEnvxv(exec, _es_GetTexEnvxv);
+    SET_GetTexGenfvOES(exec, _check_GetTexGenfvOES);
+    SET_GetTexGenivOES(exec, _check_GetTexGenivOES);
+    SET_GetTexGenxvOES(exec, _check_GetTexGenxvOES);
+    SET_GetTexParameterfv(exec, _es_GetTexParameterfv);
+    SET_GetTexParameteriv(exec, _es_GetTexParameteriv);
+    SET_GetTexParameterxv(exec, _es_GetTexParameterxv);
+    SET_Hint(exec, _es_Hint);
+    SET_IsBuffer(exec, _es_IsBuffer);
+    SET_IsEnabled(exec, _es_IsEnabled);
+    SET_IsFramebufferOES(exec, _es_IsFramebufferOES);
+    SET_IsRenderbufferOES(exec, _es_IsRenderbufferOES);
+    SET_IsTexture(exec, _es_IsTexture);
+    SET_LightModelf(exec, _es_LightModelf);
+    SET_LightModelfv(exec, _es_LightModelfv);
+    SET_LightModelx(exec, _es_LightModelx);
+    SET_LightModelxv(exec, _es_LightModelxv);
+    SET_Lightf(exec, _es_Lightf);
+    SET_Lightfv(exec, _es_Lightfv);
+    SET_Lightx(exec, _es_Lightx);
+    SET_Lightxv(exec, _es_Lightxv);
+    SET_LineWidth(exec, _es_LineWidth);
+    SET_LineWidthx(exec, _es_LineWidthx);
+    SET_LoadIdentity(exec, _es_LoadIdentity);
+    SET_LoadMatrixf(exec, _es_LoadMatrixf);
+    SET_LoadMatrixx(exec, _es_LoadMatrixx);
+    SET_LogicOp(exec, _es_LogicOp);
+    SET_MapBufferOES(exec, _es_MapBufferOES);
+    SET_Materialf(exec, _es_Materialf);
+    SET_Materialfv(exec, _es_Materialfv);
+    SET_Materialx(exec, _es_Materialx);
+    SET_Materialxv(exec, _es_Materialxv);
+    SET_MatrixMode(exec, _es_MatrixMode);
+    SET_MultMatrixf(exec, _es_MultMatrixf);
+    SET_MultMatrixx(exec, _es_MultMatrixx);
+    SET_MultiDrawArraysEXT(exec, _es_MultiDrawArraysEXT);
+    SET_MultiDrawElementsEXT(exec, _es_MultiDrawElementsEXT);
+    SET_MultiTexCoord4f(exec, _es_MultiTexCoord4f);
+    SET_MultiTexCoord4x(exec, _es_MultiTexCoord4x);
+    SET_Normal3f(exec, _es_Normal3f);
+    SET_Normal3x(exec, _es_Normal3x);
+    SET_NormalPointer(exec, _es_NormalPointer);
+    SET_Orthof(exec, _es_Orthof);
+    SET_Orthox(exec, _es_Orthox);
+    SET_PixelStorei(exec, _es_PixelStorei);
+    SET_PointParameterf(exec, _es_PointParameterf);
+    SET_PointParameterfv(exec, _es_PointParameterfv);
+    SET_PointParameterx(exec, _es_PointParameterx);
+    SET_PointParameterxv(exec, _es_PointParameterxv);
+    SET_PointSize(exec, _es_PointSize);
+    SET_PointSizePointerOES(exec, _es_PointSizePointerOES);
+    SET_PointSizex(exec, _es_PointSizex);
+    SET_PolygonOffset(exec, _es_PolygonOffset);
+    SET_PolygonOffsetx(exec, _es_PolygonOffsetx);
+    SET_PopMatrix(exec, _es_PopMatrix);
+    SET_PushMatrix(exec, _es_PushMatrix);
+    SET_QueryMatrixxOES(exec, _check_QueryMatrixxOES);
+    SET_ReadPixels(exec, _es_ReadPixels);
+    SET_RenderbufferStorageOES(exec, _check_RenderbufferStorageOES);
+    SET_Rotatef(exec, _es_Rotatef);
+    SET_Rotatex(exec, _es_Rotatex);
+    SET_SampleCoverage(exec, _es_SampleCoverage);
+    SET_SampleCoveragex(exec, _es_SampleCoveragex);
+    SET_Scalef(exec, _es_Scalef);
+    SET_Scalex(exec, _es_Scalex);
+    SET_Scissor(exec, _es_Scissor);
+    SET_ShadeModel(exec, _es_ShadeModel);
+    SET_StencilFunc(exec, _es_StencilFunc);
+    SET_StencilMask(exec, _es_StencilMask);
+    SET_StencilOp(exec, _es_StencilOp);
+    SET_TexCoordPointer(exec, _es_TexCoordPointer);
+    SET_TexEnvf(exec, _es_TexEnvf);
+    SET_TexEnvfv(exec, _es_TexEnvfv);
+    SET_TexEnvi(exec, _es_TexEnvi);
+    SET_TexEnviv(exec, _es_TexEnviv);
+    SET_TexEnvx(exec, _es_TexEnvx);
+    SET_TexEnvxv(exec, _es_TexEnvxv);
+    SET_TexGenfOES(exec, _check_TexGenfOES);
+    SET_TexGenfvOES(exec, _check_TexGenfvOES);
+    SET_TexGeniOES(exec, _check_TexGeniOES);
+    SET_TexGenivOES(exec, _check_TexGenivOES);
+    SET_TexGenxOES(exec, _check_TexGenxOES);
+    SET_TexGenxvOES(exec, _check_TexGenxvOES);
+    SET_TexImage2D(exec, _es_TexImage2D);
+    SET_TexParameterf(exec, _es_TexParameterf);
+    SET_TexParameterfv(exec, _es_TexParameterfv);
+    SET_TexParameteri(exec, _es_TexParameteri);
+    SET_TexParameteriv(exec, _es_TexParameteriv);
+    SET_TexParameterx(exec, _es_TexParameterx);
+    SET_TexParameterxv(exec, _es_TexParameterxv);
+    SET_TexSubImage2D(exec, _es_TexSubImage2D);
+    SET_Translatef(exec, _es_Translatef);
+    SET_Translatex(exec, _es_Translatex);
+    SET_UnmapBufferOES(exec, _es_UnmapBufferOES);
+    SET_VertexPointer(exec, _es_VertexPointer);
+    SET_Viewport(exec, _es_Viewport);
+
+   return exec;
+}
+
+#endif /* FEATURE_ES1 */
diff --git a/lib/libGL/gallium/libmesagallium/generated/api_exec_es2.c b/lib/libGL/gallium/libmesagallium/generated/api_exec_es2.c
new file mode 100644
index 000000000..d28291b0e
--- /dev/null
+++ b/lib/libGL/gallium/libmesagallium/generated/api_exec_es2.c
@@ -0,0 +1,2471 @@
+/* DO NOT EDIT *************************************************
+ * THIS FILE AUTOMATICALLY GENERATED BY THE es_generator.py SCRIPT
+ * API specification file:   /share/OpenBSD/xenocara/lib/libGL/libmesa/generated/../../../../dist/Mesa/src/mesa/main/APIspec.xml
+ * GLES version:             GLES2.0
+ * date:                     2011-05-15 15:41:34
+ */
+
+
+#include "GLES2/gl2.h"
+#include "GLES2/gl2ext.h"
+#include "main/mfeatures.h"
+
+#if FEATURE_ES2
+
+
+/* These types are needed for the Mesa veneer, but are not defined in
+ * the standard GLES headers.
+ */
+typedef double GLdouble;
+typedef double GLclampd;
+
+/* Mesa error handling requires these */
+extern void *_mesa_get_current_context(void);
+extern void _mesa_error(void *ctx, GLenum error, const char *fmtString, ... );
+
+#include "main/compiler.h"
+#include "main/api_exec.h"
+#include "main/remap.h"
+
+/* cannot include main/dispatch.h here */
+#ifdef IN_DRI_DRIVER
+#define _GLAPI_USE_REMAP_TABLE
+#endif
+/* glapi uses GLAPIENTRY while GLES headers define GL_APIENTRY */
+#ifndef GLAPIENTRY
+#define GLAPIENTRY GL_APIENTRY
+#endif
+#include "es2api/glapi/glapitable.h"
+#include "es2api/glapi/glapioffsets.h"
+#include "es2api/glapi/glapidispatch.h"
+
+#if FEATURE_remap_table
+
+#if !FEATURE_GL
+int driDispatchRemapTable[driDispatchRemapTable_size];
+#endif
+
+#define need_MESA_remap_table
+
+#include "es2api/main/remap_helper.h"
+
+void
+_mesa_init_remap_table_es2(void)
+{
+   _mesa_do_init_remap_table(_mesa_function_pool,
+                             driDispatchRemapTable_size,
+                             MESA_remap_table_functions);
+}
+
+void
+_mesa_map_static_functions_es2(void)
+{
+}
+
+#endif
+
+typedef void (*_glapi_proc)(void); /* generic function pointer */
+
+/*************************************************************
+ * Generated functions begin here
+ */
+
+extern void GL_APIENTRY _mesa_ActiveTextureARB(GLenum texture);
+static void GL_APIENTRY _es_ActiveTexture(GLenum texture)
+{
+    switch(texture) {
+        case GL_TEXTURE0:
+        case GL_TEXTURE1:
+        case GL_TEXTURE2:
+        case GL_TEXTURE3:
+        case GL_TEXTURE4:
+        case GL_TEXTURE5:
+        case GL_TEXTURE6:
+        case GL_TEXTURE7:
+        case GL_TEXTURE8:
+        case GL_TEXTURE9:
+        case GL_TEXTURE10:
+        case GL_TEXTURE11:
+        case GL_TEXTURE12:
+        case GL_TEXTURE13:
+        case GL_TEXTURE14:
+        case GL_TEXTURE15:
+        case GL_TEXTURE16:
+        case GL_TEXTURE17:
+        case GL_TEXTURE18:
+        case GL_TEXTURE19:
+        case GL_TEXTURE20:
+        case GL_TEXTURE21:
+        case GL_TEXTURE22:
+        case GL_TEXTURE23:
+        case GL_TEXTURE24:
+        case GL_TEXTURE25:
+        case GL_TEXTURE26:
+        case GL_TEXTURE27:
+        case GL_TEXTURE28:
+        case GL_TEXTURE29:
+        case GL_TEXTURE30:
+        case GL_TEXTURE31:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glActiveTexture(texture=0x%x)", texture);
+            return;
+    }
+
+    _mesa_ActiveTextureARB(texture);
+}
+
+extern void GL_APIENTRY _mesa_AttachShader(GLuint program, GLuint shader);
+#define _es_AttachShader _mesa_AttachShader
+
+extern void GL_APIENTRY _mesa_BindAttribLocationARB(GLuint program, GLuint index, const char * name);
+#define _es_BindAttribLocation _mesa_BindAttribLocationARB
+
+extern void GL_APIENTRY _mesa_BindBufferARB(GLenum target, GLuint buffer);
+static void GL_APIENTRY _es_BindBuffer(GLenum target, GLuint buffer)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindBuffer(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindBufferARB(target, buffer);
+}
+
+extern void GL_APIENTRY _mesa_BindFramebufferEXT(GLenum target, GLuint framebuffer);
+static void GL_APIENTRY _es_BindFramebuffer(GLenum target, GLuint framebuffer)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindFramebuffer(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindFramebufferEXT(target, framebuffer);
+}
+
+extern void GL_APIENTRY _mesa_BindRenderbufferEXT(GLenum target, GLuint renderbuffer);
+static void GL_APIENTRY _es_BindRenderbuffer(GLenum target, GLuint renderbuffer)
+{
+    switch(target) {
+        case GL_RENDERBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindRenderbuffer(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindRenderbufferEXT(target, renderbuffer);
+}
+
+extern void GL_APIENTRY _mesa_BindTexture(GLenum target, GLuint texture);
+static void GL_APIENTRY _es_BindTexture(GLenum target, GLuint texture)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBindTexture(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BindTexture(target, texture);
+}
+
+extern void GL_APIENTRY _mesa_BlendColor(GLclampf red, GLclampf green, GLclampf blue, GLclampf alpha);
+#define _es_BlendColor _mesa_BlendColor
+
+extern void GL_APIENTRY _mesa_BlendEquation(GLenum mode);
+static void GL_APIENTRY _es_BlendEquation(GLenum mode)
+{
+    switch(mode) {
+        case GL_FUNC_ADD:
+        case GL_FUNC_SUBTRACT:
+        case GL_FUNC_REVERSE_SUBTRACT:
+        case GL_MIN_EXT:
+        case GL_MAX_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendEquation(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_BlendEquation(mode);
+}
+
+extern void GL_APIENTRY _mesa_BlendEquationSeparateEXT(GLenum modeRGB, GLenum modeAlpha);
+static void GL_APIENTRY _es_BlendEquationSeparate(GLenum modeRGB, GLenum modeAlpha)
+{
+    switch(modeRGB) {
+        case GL_FUNC_ADD:
+        case GL_FUNC_SUBTRACT:
+        case GL_FUNC_REVERSE_SUBTRACT:
+        case GL_MIN_EXT:
+        case GL_MAX_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendEquationSeparate(modeRGB=0x%x)", modeRGB);
+            return;
+    }
+    switch(modeAlpha) {
+        case GL_FUNC_ADD:
+        case GL_FUNC_SUBTRACT:
+        case GL_FUNC_REVERSE_SUBTRACT:
+        case GL_MIN_EXT:
+        case GL_MAX_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendEquationSeparate(modeAlpha=0x%x)", modeAlpha);
+            return;
+    }
+
+    _mesa_BlendEquationSeparateEXT(modeRGB, modeAlpha);
+}
+
+extern void GL_APIENTRY _mesa_BlendFunc(GLenum sfactor, GLenum dfactor);
+static void GL_APIENTRY _es_BlendFunc(GLenum sfactor, GLenum dfactor)
+{
+    switch(sfactor) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_SRC_ALPHA_SATURATE:
+        case GL_CONSTANT_COLOR:
+        case GL_CONSTANT_ALPHA:
+        case GL_ONE_MINUS_CONSTANT_COLOR:
+        case GL_ONE_MINUS_CONSTANT_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFunc(sfactor=0x%x)", sfactor);
+            return;
+    }
+    switch(dfactor) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_CONSTANT_COLOR:
+        case GL_CONSTANT_ALPHA:
+        case GL_ONE_MINUS_CONSTANT_COLOR:
+        case GL_ONE_MINUS_CONSTANT_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFunc(dfactor=0x%x)", dfactor);
+            return;
+    }
+
+    _mesa_BlendFunc(sfactor, dfactor);
+}
+
+extern void GL_APIENTRY _mesa_BlendFuncSeparateEXT(GLenum srcRGB, GLenum dstRGB, GLenum srcAlpha, GLenum dstAlpha);
+static void GL_APIENTRY _es_BlendFuncSeparate(GLenum srcRGB, GLenum dstRGB, GLenum srcAlpha, GLenum dstAlpha)
+{
+    switch(srcRGB) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_SRC_ALPHA_SATURATE:
+        case GL_CONSTANT_COLOR:
+        case GL_ONE_MINUS_CONSTANT_COLOR:
+        case GL_CONSTANT_ALPHA:
+        case GL_ONE_MINUS_CONSTANT_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparate(srcRGB=0x%x)", srcRGB);
+            return;
+    }
+    switch(dstRGB) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_CONSTANT_COLOR:
+        case GL_ONE_MINUS_CONSTANT_COLOR:
+        case GL_CONSTANT_ALPHA:
+        case GL_ONE_MINUS_CONSTANT_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparate(dstRGB=0x%x)", dstRGB);
+            return;
+    }
+    switch(srcAlpha) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_SRC_ALPHA_SATURATE:
+        case GL_CONSTANT_COLOR:
+        case GL_ONE_MINUS_CONSTANT_COLOR:
+        case GL_CONSTANT_ALPHA:
+        case GL_ONE_MINUS_CONSTANT_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparate(srcAlpha=0x%x)", srcAlpha);
+            return;
+    }
+    switch(dstAlpha) {
+        case GL_ZERO:
+        case GL_ONE:
+        case GL_SRC_COLOR:
+        case GL_ONE_MINUS_SRC_COLOR:
+        case GL_SRC_ALPHA:
+        case GL_ONE_MINUS_SRC_ALPHA:
+        case GL_DST_ALPHA:
+        case GL_ONE_MINUS_DST_ALPHA:
+        case GL_DST_COLOR:
+        case GL_ONE_MINUS_DST_COLOR:
+        case GL_CONSTANT_COLOR:
+        case GL_ONE_MINUS_CONSTANT_COLOR:
+        case GL_CONSTANT_ALPHA:
+        case GL_ONE_MINUS_CONSTANT_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBlendFuncSeparate(dstAlpha=0x%x)", dstAlpha);
+            return;
+    }
+
+    _mesa_BlendFuncSeparateEXT(srcRGB, dstRGB, srcAlpha, dstAlpha);
+}
+
+extern void GL_APIENTRY _mesa_BufferDataARB(GLenum target, GLsizeiptr size, const GLvoid * data, GLenum usage);
+static void GL_APIENTRY _es_BufferData(GLenum target, GLsizeiptr size, const GLvoid *data, GLenum usage)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBufferData(target=0x%x)", target);
+            return;
+    }
+    switch(usage) {
+        case GL_STATIC_DRAW:
+        case GL_DYNAMIC_DRAW:
+        case GL_STREAM_DRAW:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBufferData(usage=0x%x)", usage);
+            return;
+    }
+
+    _mesa_BufferDataARB(target, size, data, usage);
+}
+
+extern void GL_APIENTRY _mesa_BufferSubDataARB(GLenum target, GLintptr offset, GLsizeiptr size, const GLvoid * data);
+static void GL_APIENTRY _es_BufferSubData(GLenum target, GLintptr offset, GLsizeiptr size, const GLvoid *data)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glBufferSubData(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_BufferSubDataARB(target, offset, size, data);
+}
+
+extern GLenum GL_APIENTRY _mesa_CheckFramebufferStatusEXT(GLenum target);
+static GLenum GL_APIENTRY _es_CheckFramebufferStatus(GLenum target)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCheckFramebufferStatus(target=0x%x)", target);
+            return (GLenum) 0;
+    }
+
+    return _mesa_CheckFramebufferStatusEXT(target);
+}
+
+extern void GL_APIENTRY _mesa_Clear(GLbitfield mask);
+static void GL_APIENTRY _es_Clear(GLbitfield mask)
+{
+    switch(mask) {
+        case 0:
+        case (GL_COLOR_BUFFER_BIT):
+        case (GL_DEPTH_BUFFER_BIT):
+        case (GL_STENCIL_BUFFER_BIT):
+        case (GL_COLOR_BUFFER_BIT|GL_DEPTH_BUFFER_BIT):
+        case (GL_COLOR_BUFFER_BIT|GL_STENCIL_BUFFER_BIT):
+        case (GL_DEPTH_BUFFER_BIT|GL_STENCIL_BUFFER_BIT):
+        case (GL_COLOR_BUFFER_BIT|GL_DEPTH_BUFFER_BIT|GL_STENCIL_BUFFER_BIT):
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glClear(mask=0x%x)", mask);
+            return;
+    }
+
+    _mesa_Clear(mask);
+}
+
+extern void GL_APIENTRY _mesa_ClearColor(GLclampf red, GLclampf green, GLclampf blue, GLclampf alpha);
+#define _es_ClearColor _mesa_ClearColor
+
+extern void GL_APIENTRY _mesa_ClearDepthf(GLclampf depth);
+#define _es_ClearDepthf _mesa_ClearDepthf
+
+extern void GL_APIENTRY _mesa_ClearStencil(GLint s);
+#define _es_ClearStencil _mesa_ClearStencil
+
+extern void GL_APIENTRY _mesa_ColorMask(GLboolean red, GLboolean green, GLboolean blue, GLboolean alpha);
+#define _es_ColorMask _mesa_ColorMask
+
+extern void GL_APIENTRY _mesa_CompileShaderARB(GLuint shader);
+#define _es_CompileShader _mesa_CompileShaderARB
+
+extern void GL_APIENTRY _mesa_CompressedTexImage2DARB(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLint border, GLsizei imageSize, const GLvoid * data);
+static void GL_APIENTRY _es_CompressedTexImage2D(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLint border, GLsizei imageSize, const GLvoid *data)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_PALETTE4_RGB8_OES:
+        case GL_PALETTE4_RGBA8_OES:
+        case GL_PALETTE4_R5_G6_B5_OES:
+        case GL_PALETTE4_RGBA4_OES:
+        case GL_PALETTE4_RGB5_A1_OES:
+        case GL_PALETTE8_RGB8_OES:
+        case GL_PALETTE8_RGBA8_OES:
+        case GL_PALETTE8_R5_G6_B5_OES:
+        case GL_PALETTE8_RGBA4_OES:
+        case GL_PALETTE8_RGB5_A1_OES:
+        case GL_COMPRESSED_RGB_S3TC_DXT1_EXT:
+        case GL_COMPRESSED_RGBA_S3TC_DXT1_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexImage2D(internalFormat=0x%x)", internalFormat);
+            return;
+    }
+    switch(border) {
+        case 0:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glCompressedTexImage2D(border=%d)", border);
+            return;
+    }
+
+    _mesa_CompressedTexImage2DARB(target, level, internalFormat, width, height, border, imageSize, data);
+}
+
+extern void GL_APIENTRY _mesa_CompressedTexImage3DARB(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLsizei depth, GLint border, GLsizei imagesize, const GLvoid * data);
+static void GL_APIENTRY _es_CompressedTexImage3DOES(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLsizei depth, GLint border, GLsizei imagesize, const GLvoid *data)
+{
+    switch(target) {
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexImage3DOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_CompressedTexImage3DARB(target, level, internalFormat, width, height, depth, border, imagesize, data);
+}
+
+extern void GL_APIENTRY _mesa_CompressedTexSubImage2DARB(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLsizei imageSize, const GLvoid * data);
+static void GL_APIENTRY _es_CompressedTexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLsizei imageSize, const GLvoid *data)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexSubImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(format) {
+        case GL_COMPRESSED_RGB_S3TC_DXT1_EXT:
+        case GL_COMPRESSED_RGBA_S3TC_DXT1_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexSubImage2D(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_CompressedTexSubImage2DARB(target, level, xoffset, yoffset, width, height, format, imageSize, data);
+}
+
+extern void GL_APIENTRY _mesa_CompressedTexSubImage3DARB(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLsizei width, GLsizei height, GLsizei depth, GLenum format, GLsizei imagesize, const GLvoid * data);
+static void GL_APIENTRY _es_CompressedTexSubImage3DOES(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLsizei width, GLsizei height, GLsizei depth, GLenum format, GLsizei imagesize, const GLvoid *data)
+{
+    switch(target) {
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCompressedTexSubImage3DOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_CompressedTexSubImage3DARB(target, level, xoffset, yoffset, zoffset, width, height, depth, format, imagesize, data);
+}
+
+extern void GL_APIENTRY _mesa_CopyTexImage2D(GLenum target, GLint level, GLenum internalFormat, GLint x, GLint y, GLsizei width, GLsizei height, GLint border);
+static void GL_APIENTRY _es_CopyTexImage2D(GLenum target, GLint level, GLenum internalFormat, GLint x, GLint y, GLsizei width, GLsizei height, GLint border)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCopyTexImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_ALPHA:
+        case GL_RGB:
+        case GL_RGBA:
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glCopyTexImage2D(internalFormat=0x%x)", internalFormat);
+            return;
+    }
+    switch(border) {
+        case 0:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glCopyTexImage2D(border=%d)", border);
+            return;
+    }
+
+    _mesa_CopyTexImage2D(target, level, internalFormat, x, y, width, height, border);
+}
+
+extern void GL_APIENTRY _mesa_CopyTexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint x, GLint y, GLsizei width, GLsizei height);
+static void GL_APIENTRY _es_CopyTexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint x, GLint y, GLsizei width, GLsizei height)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCopyTexSubImage2D(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_CopyTexSubImage2D(target, level, xoffset, yoffset, x, y, width, height);
+}
+
+extern void GL_APIENTRY _mesa_CopyTexSubImage3D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLint x, GLint y, GLsizei width, GLsizei height);
+static void GL_APIENTRY _es_CopyTexSubImage3DOES(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLint x, GLint y, GLsizei width, GLsizei height)
+{
+    switch(target) {
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCopyTexSubImage3DOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_CopyTexSubImage3D(target, level, xoffset, yoffset, zoffset, x, y, width, height);
+}
+
+extern GLuint GL_APIENTRY _mesa_CreateProgram(void);
+#define _es_CreateProgram _mesa_CreateProgram
+
+extern GLuint GL_APIENTRY _mesa_CreateShader(GLenum type);
+static GLuint GL_APIENTRY _es_CreateShader(GLenum type)
+{
+    switch(type) {
+        case GL_VERTEX_SHADER:
+        case GL_FRAGMENT_SHADER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCreateShader(type=0x%x)", type);
+            return (GLuint) 0;
+    }
+
+    return _mesa_CreateShader(type);
+}
+
+extern void GL_APIENTRY _mesa_CullFace(GLenum mode);
+static void GL_APIENTRY _es_CullFace(GLenum mode)
+{
+    switch(mode) {
+        case GL_FRONT:
+        case GL_BACK:
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glCullFace(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_CullFace(mode);
+}
+
+extern void GL_APIENTRY _mesa_DeleteBuffersARB(GLsizei n, const GLuint * buffer);
+#define _es_DeleteBuffers _mesa_DeleteBuffersARB
+
+extern void GL_APIENTRY _mesa_DeleteFramebuffersEXT(GLsizei n, const GLuint * framebuffers);
+#define _es_DeleteFramebuffers _mesa_DeleteFramebuffersEXT
+
+extern void GL_APIENTRY _mesa_DeleteProgram(GLuint program);
+#define _es_DeleteProgram _mesa_DeleteProgram
+
+extern void GL_APIENTRY _mesa_DeleteRenderbuffersEXT(GLsizei n, const GLuint * renderbuffers);
+#define _es_DeleteRenderbuffers _mesa_DeleteRenderbuffersEXT
+
+extern void GL_APIENTRY _mesa_DeleteShader(GLuint shader);
+#define _es_DeleteShader _mesa_DeleteShader
+
+extern void GL_APIENTRY _mesa_DeleteTextures(GLsizei n, const GLuint * textures);
+#define _es_DeleteTextures _mesa_DeleteTextures
+
+extern void GL_APIENTRY _mesa_DepthFunc(GLenum func);
+static void GL_APIENTRY _es_DepthFunc(GLenum func)
+{
+    switch(func) {
+        case GL_NEVER:
+        case GL_LESS:
+        case GL_EQUAL:
+        case GL_LEQUAL:
+        case GL_GREATER:
+        case GL_NOTEQUAL:
+        case GL_GEQUAL:
+        case GL_ALWAYS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDepthFunc(func=0x%x)", func);
+            return;
+    }
+
+    _mesa_DepthFunc(func);
+}
+
+extern void GL_APIENTRY _mesa_DepthMask(GLboolean flag);
+#define _es_DepthMask _mesa_DepthMask
+
+extern void GL_APIENTRY _mesa_DepthRangef(GLclampf zNear, GLclampf zFar);
+#define _es_DepthRangef _mesa_DepthRangef
+
+extern void GL_APIENTRY _mesa_DetachShader(GLuint program, GLuint shader);
+#define _es_DetachShader _mesa_DetachShader
+
+extern void GL_APIENTRY _mesa_Disable(GLenum cap);
+static void GL_APIENTRY _es_Disable(GLenum cap)
+{
+    switch(cap) {
+        case GL_CULL_FACE:
+        case GL_SCISSOR_TEST:
+        case GL_POLYGON_OFFSET_FILL:
+        case GL_SAMPLE_ALPHA_TO_COVERAGE:
+        case GL_SAMPLE_COVERAGE:
+        case GL_STENCIL_TEST:
+        case GL_DEPTH_TEST:
+        case GL_DITHER:
+        case GL_BLEND:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDisable(cap=0x%x)", cap);
+            return;
+    }
+
+    _mesa_Disable(cap);
+}
+
+extern void GL_APIENTRY _mesa_DisableVertexAttribArrayARB(GLuint index);
+#define _es_DisableVertexAttribArray _mesa_DisableVertexAttribArrayARB
+
+extern void GL_APIENTRY _mesa_DrawArrays(GLenum mode, GLint first, GLsizei count);
+static void GL_APIENTRY _es_DrawArrays(GLenum mode, GLint first, GLsizei count)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDrawArrays(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_DrawArrays(mode, first, count);
+}
+
+extern void GL_APIENTRY _mesa_DrawElements(GLenum mode, GLsizei count, GLenum type, const GLvoid * indices);
+static void GL_APIENTRY _es_DrawElements(GLenum mode, GLsizei count, GLenum type, const GLvoid *indices)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDrawElements(mode=0x%x)", mode);
+            return;
+    }
+    switch(type) {
+        case GL_UNSIGNED_BYTE:
+        case GL_UNSIGNED_SHORT:
+        case GL_UNSIGNED_INT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glDrawElements(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_DrawElements(mode, count, type, indices);
+}
+
+extern void GL_APIENTRY _mesa_EGLImageTargetRenderbufferStorageOES(GLenum target, GLeglImageOES image);
+static void GL_APIENTRY _es_EGLImageTargetRenderbufferStorageOES(GLenum target, GLeglImageOES image)
+{
+    switch(target) {
+        case GL_RENDERBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glEGLImageTargetRenderbufferStorageOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_EGLImageTargetRenderbufferStorageOES(target, image);
+}
+
+extern void GL_APIENTRY _mesa_EGLImageTargetTexture2DOES(GLenum target, GLeglImageOES image);
+static void GL_APIENTRY _es_EGLImageTargetTexture2DOES(GLenum target, GLeglImageOES image)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glEGLImageTargetTexture2DOES(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_EGLImageTargetTexture2DOES(target, image);
+}
+
+extern void GL_APIENTRY _mesa_Enable(GLenum cap);
+static void GL_APIENTRY _es_Enable(GLenum cap)
+{
+    switch(cap) {
+        case GL_CULL_FACE:
+        case GL_SCISSOR_TEST:
+        case GL_POLYGON_OFFSET_FILL:
+        case GL_SAMPLE_ALPHA_TO_COVERAGE:
+        case GL_SAMPLE_COVERAGE:
+        case GL_STENCIL_TEST:
+        case GL_DEPTH_TEST:
+        case GL_DITHER:
+        case GL_BLEND:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glEnable(cap=0x%x)", cap);
+            return;
+    }
+
+    _mesa_Enable(cap);
+}
+
+extern void GL_APIENTRY _mesa_EnableVertexAttribArrayARB(GLuint index);
+#define _es_EnableVertexAttribArray _mesa_EnableVertexAttribArrayARB
+
+extern void GL_APIENTRY _mesa_Finish(void);
+#define _es_Finish _mesa_Finish
+
+extern void GL_APIENTRY _mesa_Flush(void);
+#define _es_Flush _mesa_Flush
+
+extern void GL_APIENTRY _mesa_FramebufferRenderbufferEXT(GLenum target, GLenum attachment, GLenum renderbuffertarget, GLuint renderbuffer);
+static void GL_APIENTRY _es_FramebufferRenderbuffer(GLenum target, GLenum attachment, GLenum renderbuffertarget, GLuint renderbuffer)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferRenderbuffer(target=0x%x)", target);
+            return;
+    }
+    switch(attachment) {
+        case GL_COLOR_ATTACHMENT0:
+        case GL_DEPTH_ATTACHMENT:
+        case GL_STENCIL_ATTACHMENT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferRenderbuffer(attachment=0x%x)", attachment);
+            return;
+    }
+    switch(renderbuffertarget) {
+        case GL_RENDERBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferRenderbuffer(renderbuffertarget=0x%x)", renderbuffertarget);
+            return;
+    }
+
+    _mesa_FramebufferRenderbufferEXT(target, attachment, renderbuffertarget, renderbuffer);
+}
+
+extern void GL_APIENTRY _mesa_FramebufferTexture2DEXT(GLenum target, GLenum attachment, GLenum textarget, GLuint texture, GLint level);
+static void GL_APIENTRY _es_FramebufferTexture2D(GLenum target, GLenum attachment, GLenum textarget, GLuint texture, GLint level)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferTexture2D(target=0x%x)", target);
+            return;
+    }
+    switch(attachment) {
+        case GL_COLOR_ATTACHMENT0:
+        case GL_DEPTH_ATTACHMENT:
+        case GL_STENCIL_ATTACHMENT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferTexture2D(attachment=0x%x)", attachment);
+            return;
+    }
+    switch(textarget) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glFramebufferTexture2D(textarget=0x%x)", textarget);
+            return;
+    }
+
+    _mesa_FramebufferTexture2DEXT(target, attachment, textarget, texture, level);
+}
+
+extern void GL_APIENTRY _mesa_FramebufferTexture3DEXT(GLenum target, GLenum attachment, GLenum textarget, GLuint texture, GLint level, GLint zoffset);
+static void GL_APIENTRY _es_FramebufferTexture3DOES(GLenum target, GLenum attachment, GLenum textarget, GLuint texture, GLint level, GLint zoffset)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferTexture3DOES(target=0x%x)", target);
+            return;
+    }
+    switch(attachment) {
+        case GL_COLOR_ATTACHMENT0:
+        case GL_DEPTH_ATTACHMENT:
+        case GL_STENCIL_ATTACHMENT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFramebufferTexture3DOES(attachment=0x%x)", attachment);
+            return;
+    }
+    switch(textarget) {
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glFramebufferTexture3DOES(textarget=0x%x)", textarget);
+            return;
+    }
+
+    _mesa_FramebufferTexture3DEXT(target, attachment, textarget, texture, level, zoffset);
+}
+
+extern void GL_APIENTRY _mesa_FrontFace(GLenum mode);
+static void GL_APIENTRY _es_FrontFace(GLenum mode)
+{
+    switch(mode) {
+        case GL_CW:
+        case GL_CCW:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glFrontFace(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_FrontFace(mode);
+}
+
+extern void GL_APIENTRY _mesa_GenBuffersARB(GLsizei n, GLuint * buffer);
+#define _es_GenBuffers _mesa_GenBuffersARB
+
+extern void GL_APIENTRY _mesa_GenFramebuffersEXT(GLsizei n, GLuint * ids);
+#define _es_GenFramebuffers _mesa_GenFramebuffersEXT
+
+extern void GL_APIENTRY _mesa_GenRenderbuffersEXT(GLsizei n, GLuint * renderbuffers);
+#define _es_GenRenderbuffers _mesa_GenRenderbuffersEXT
+
+extern void GL_APIENTRY _mesa_GenTextures(GLsizei n, GLuint * textures);
+#define _es_GenTextures _mesa_GenTextures
+
+extern void GL_APIENTRY _mesa_GenerateMipmapEXT(GLenum target);
+static void GL_APIENTRY _es_GenerateMipmap(GLenum target)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGenerateMipmap(target=0x%x)", target);
+            return;
+    }
+
+    _mesa_GenerateMipmapEXT(target);
+}
+
+extern void GL_APIENTRY _mesa_GetActiveAttribARB(GLuint program, GLuint index, GLsizei bufSize, GLsizei * length, GLint * size, GLenum * type, GLchar * name);
+#define _es_GetActiveAttrib _mesa_GetActiveAttribARB
+
+extern void GL_APIENTRY _mesa_GetActiveUniformARB(GLuint program, GLuint index, GLsizei bufSize, GLsizei * length, GLint * size, GLenum * type, GLchar * name);
+#define _es_GetActiveUniform _mesa_GetActiveUniformARB
+
+extern void GL_APIENTRY _mesa_GetAttachedShaders(GLuint program, GLsizei maxCount, GLsizei * count, GLuint * shaders);
+#define _es_GetAttachedShaders _mesa_GetAttachedShaders
+
+extern GLint GL_APIENTRY _mesa_GetAttribLocationARB(GLuint program, const char * name);
+#define _es_GetAttribLocation _mesa_GetAttribLocationARB
+
+extern void GL_APIENTRY _mesa_GetBooleanv(GLenum pname, GLboolean * params);
+#define _es_GetBooleanv _mesa_GetBooleanv
+
+extern void GL_APIENTRY _mesa_GetBufferParameterivARB(GLenum target, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetBufferParameteriv(GLenum target, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_BUFFER_SIZE:
+        case GL_BUFFER_USAGE:
+        case GL_BUFFER_ACCESS_OES:
+        case GL_BUFFER_MAPPED_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetBufferParameterivARB(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetBufferPointervARB(GLenum target, GLenum pname, GLvoid ** params);
+static void GL_APIENTRY _es_GetBufferPointervOES(GLenum target, GLenum pname, GLvoid **params)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferPointervOES(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_BUFFER_MAP_POINTER_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetBufferPointervOES(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetBufferPointervARB(target, pname, params);
+}
+
+extern GLenum GL_APIENTRY _mesa_GetError(void);
+#define _es_GetError _mesa_GetError
+
+extern void GL_APIENTRY _mesa_GetFloatv(GLenum pname, GLfloat * params);
+#define _es_GetFloatv _mesa_GetFloatv
+
+extern void GL_APIENTRY _mesa_GetFramebufferAttachmentParameterivEXT(GLenum target, GLenum attachment, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetFramebufferAttachmentParameteriv(GLenum target, GLenum attachment, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_FRAMEBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetFramebufferAttachmentParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_FRAMEBUFFER_ATTACHMENT_OBJECT_TYPE: /* size 1 */
+        case GL_FRAMEBUFFER_ATTACHMENT_OBJECT_NAME: /* size 1 */
+        case GL_FRAMEBUFFER_ATTACHMENT_TEXTURE_LEVEL: /* size 1 */
+        case GL_FRAMEBUFFER_ATTACHMENT_TEXTURE_CUBE_MAP_FACE: /* size 1 */
+        case GL_FRAMEBUFFER_ATTACHMENT_TEXTURE_3D_ZOFFSET_OES: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetFramebufferAttachmentParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetFramebufferAttachmentParameterivEXT(target, attachment, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetIntegerv(GLenum pname, GLint * params);
+#define _es_GetIntegerv _mesa_GetIntegerv
+
+extern void GL_APIENTRY _mesa_GetProgramInfoLog(GLuint program, GLsizei bufSize, GLsizei * length, GLchar * infoLog);
+#define _es_GetProgramInfoLog _mesa_GetProgramInfoLog
+
+extern void GL_APIENTRY _mesa_GetProgramiv(GLuint program, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetProgramiv(GLuint program, GLenum pname, GLint *params)
+{
+    switch(pname) {
+        case GL_DELETE_STATUS:
+        case GL_LINK_STATUS:
+        case GL_VALIDATE_STATUS:
+        case GL_INFO_LOG_LENGTH:
+        case GL_ATTACHED_SHADERS:
+        case GL_ACTIVE_ATTRIBUTES:
+        case GL_ACTIVE_ATTRIBUTE_MAX_LENGTH:
+        case GL_ACTIVE_UNIFORMS:
+        case GL_ACTIVE_UNIFORM_MAX_LENGTH:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetProgramiv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetProgramiv(program, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetRenderbufferParameterivEXT(GLenum target, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetRenderbufferParameteriv(GLenum target, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_RENDERBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetRenderbufferParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_RENDERBUFFER_WIDTH: /* size 1 */
+        case GL_RENDERBUFFER_HEIGHT: /* size 1 */
+        case GL_RENDERBUFFER_INTERNAL_FORMAT: /* size 1 */
+        case GL_RENDERBUFFER_RED_SIZE: /* size 1 */
+        case GL_RENDERBUFFER_GREEN_SIZE: /* size 1 */
+        case GL_RENDERBUFFER_BLUE_SIZE: /* size 1 */
+        case GL_RENDERBUFFER_ALPHA_SIZE: /* size 1 */
+        case GL_RENDERBUFFER_DEPTH_SIZE: /* size 1 */
+        case GL_RENDERBUFFER_STENCIL_SIZE: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetRenderbufferParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetRenderbufferParameterivEXT(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetShaderInfoLog(GLuint shader, GLsizei bufSize, GLsizei * length, GLchar * infoLog);
+#define _es_GetShaderInfoLog _mesa_GetShaderInfoLog
+
+extern void GL_APIENTRY _mesa_GetShaderPrecisionFormat(GLenum shadertype, GLenum precisiontype, GLint * range, GLint * precision);
+static void GL_APIENTRY _es_GetShaderPrecisionFormat(GLenum shadertype, GLenum precisiontype, GLint *range, GLint *precision)
+{
+    switch(shadertype) {
+        case GL_VERTEX_SHADER:
+        case GL_FRAGMENT_SHADER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetShaderPrecisionFormat(shadertype=0x%x)", shadertype);
+            return;
+    }
+    switch(precisiontype) {
+        case GL_LOW_FLOAT:
+        case GL_MEDIUM_FLOAT:
+        case GL_HIGH_FLOAT:
+        case GL_LOW_INT:
+        case GL_MEDIUM_INT:
+        case GL_HIGH_INT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetShaderPrecisionFormat(precisiontype=0x%x)", precisiontype);
+            return;
+    }
+
+    _mesa_GetShaderPrecisionFormat(shadertype, precisiontype, range, precision);
+}
+
+extern void GL_APIENTRY _mesa_GetShaderSourceARB(GLuint shader, GLsizei bufSize, GLsizei * length, GLchar * source);
+#define _es_GetShaderSource _mesa_GetShaderSourceARB
+
+extern void GL_APIENTRY _mesa_GetShaderiv(GLuint shader, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetShaderiv(GLuint shader, GLenum pname, GLint *params)
+{
+    switch(pname) {
+        case GL_SHADER_TYPE:
+        case GL_COMPILE_STATUS:
+        case GL_DELETE_STATUS:
+        case GL_INFO_LOG_LENGTH:
+        case GL_SHADER_SOURCE_LENGTH:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetShaderiv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetShaderiv(shader, pname, params);
+}
+
+extern const GLubyte * GL_APIENTRY _mesa_GetString(GLenum name);
+static const GLubyte * GL_APIENTRY _es_GetString(GLenum name)
+{
+    switch(name) {
+        case GL_VENDOR:
+        case GL_RENDERER:
+        case GL_VERSION:
+        case GL_EXTENSIONS:
+        case GL_SHADING_LANGUAGE_VERSION:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetString(name=0x%x)", name);
+            return (const GLubyte *) 0;
+    }
+
+    return _mesa_GetString(name);
+}
+
+extern void GL_APIENTRY _mesa_GetTexParameterfv(GLenum target, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetTexParameterfv(GLenum target, GLenum pname, GLfloat *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameterfv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+        case GL_TEXTURE_WRAP_R_OES: /* size 1 */
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameterfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexParameterfv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetTexParameteriv(GLenum target, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetTexParameteriv(GLenum target, GLenum pname, GLint *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+        case GL_TEXTURE_WRAP_R_OES: /* size 1 */
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetTexParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetTexParameteriv(target, pname, params);
+}
+
+extern GLint GL_APIENTRY _mesa_GetUniformLocationARB(GLuint program, const char * name);
+#define _es_GetUniformLocation _mesa_GetUniformLocationARB
+
+extern void GL_APIENTRY _mesa_GetUniformfvARB(GLuint program, GLint location, GLfloat * params);
+#define _es_GetUniformfv _mesa_GetUniformfvARB
+
+extern void GL_APIENTRY _mesa_GetUniformivARB(GLuint program, GLint location, GLint * params);
+#define _es_GetUniformiv _mesa_GetUniformivARB
+
+extern void GL_APIENTRY _mesa_GetVertexAttribPointervARB(GLuint index, GLenum pname, GLvoid ** pointer);
+static void GL_APIENTRY _es_GetVertexAttribPointerv(GLuint index, GLenum pname, GLvoid **pointer)
+{
+    switch(pname) {
+        case GL_VERTEX_ATTRIB_ARRAY_POINTER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetVertexAttribPointerv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetVertexAttribPointervARB(index, pname, pointer);
+}
+
+extern void GL_APIENTRY _mesa_GetVertexAttribfvARB(GLuint index, GLenum pname, GLfloat * params);
+static void GL_APIENTRY _es_GetVertexAttribfv(GLuint index, GLenum pname, GLfloat *params)
+{
+    switch(pname) {
+        case GL_VERTEX_ATTRIB_ARRAY_ENABLED: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_SIZE: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_STRIDE: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_TYPE: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_NORMALIZED: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_BUFFER_BINDING: /* size 1 */
+        case GL_CURRENT_VERTEX_ATTRIB: /* size 16? */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetVertexAttribfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetVertexAttribfvARB(index, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_GetVertexAttribivARB(GLuint index, GLenum pname, GLint * params);
+static void GL_APIENTRY _es_GetVertexAttribiv(GLuint index, GLenum pname, GLint *params)
+{
+    switch(pname) {
+        case GL_VERTEX_ATTRIB_ARRAY_ENABLED: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_SIZE: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_STRIDE: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_TYPE: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_NORMALIZED: /* size 1 */
+        case GL_VERTEX_ATTRIB_ARRAY_BUFFER_BINDING: /* size 1 */
+        case GL_CURRENT_VERTEX_ATTRIB: /* size 16? */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glGetVertexAttribiv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_GetVertexAttribivARB(index, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_Hint(GLenum target, GLenum mode);
+static void GL_APIENTRY _es_Hint(GLenum target, GLenum mode)
+{
+    switch(target) {
+        case GL_FRAGMENT_SHADER_DERIVATIVE_HINT_OES:
+        case GL_GENERATE_MIPMAP_HINT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glHint(target=0x%x)", target);
+            return;
+    }
+    switch(mode) {
+        case GL_FASTEST:
+        case GL_NICEST:
+        case GL_DONT_CARE:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glHint(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_Hint(target, mode);
+}
+
+extern GLboolean GL_APIENTRY _mesa_IsBufferARB(GLuint buffer);
+#define _es_IsBuffer _mesa_IsBufferARB
+
+extern GLboolean GL_APIENTRY _mesa_IsEnabled(GLenum cap);
+static GLboolean GL_APIENTRY _es_IsEnabled(GLenum cap)
+{
+    switch(cap) {
+        case GL_CULL_FACE:
+        case GL_SCISSOR_TEST:
+        case GL_POLYGON_OFFSET_FILL:
+        case GL_SAMPLE_ALPHA_TO_COVERAGE:
+        case GL_SAMPLE_COVERAGE:
+        case GL_STENCIL_TEST:
+        case GL_DEPTH_TEST:
+        case GL_DITHER:
+        case GL_BLEND:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glIsEnabled(cap=0x%x)", cap);
+            return GL_FALSE;
+    }
+
+    return _mesa_IsEnabled(cap);
+}
+
+extern GLboolean GL_APIENTRY _mesa_IsFramebufferEXT(GLuint framebuffer);
+#define _es_IsFramebuffer _mesa_IsFramebufferEXT
+
+extern GLboolean GL_APIENTRY _mesa_IsProgram(GLuint program);
+#define _es_IsProgram _mesa_IsProgram
+
+extern GLboolean GL_APIENTRY _mesa_IsRenderbufferEXT(GLuint renderbuffer);
+#define _es_IsRenderbuffer _mesa_IsRenderbufferEXT
+
+extern GLboolean GL_APIENTRY _mesa_IsShader(GLuint shader);
+#define _es_IsShader _mesa_IsShader
+
+extern GLboolean GL_APIENTRY _mesa_IsTexture(GLuint texture);
+#define _es_IsTexture _mesa_IsTexture
+
+extern void GL_APIENTRY _mesa_LineWidth(GLfloat width);
+#define _es_LineWidth _mesa_LineWidth
+
+extern void GL_APIENTRY _mesa_LinkProgramARB(GLuint program);
+#define _es_LinkProgram _mesa_LinkProgramARB
+
+extern void * GL_APIENTRY _mesa_MapBufferARB(GLenum target, GLenum access);
+static void * GL_APIENTRY _es_MapBufferOES(GLenum target, GLenum access)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMapBufferOES(target=0x%x)", target);
+            return (void *) 0;
+    }
+    switch(access) {
+        case GL_WRITE_ONLY_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMapBufferOES(access=0x%x)", access);
+            return (void *) 0;
+    }
+
+    return _mesa_MapBufferARB(target, access);
+}
+
+extern void GL_APIENTRY _mesa_MultiDrawArraysEXT(GLenum mode, const GLint * first, const GLsizei * count, GLsizei primcount);
+static void GL_APIENTRY _es_MultiDrawArraysEXT(GLenum mode, const GLint *first, const GLsizei *count, GLsizei primcount)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiDrawArraysEXT(mode=0x%x)", mode);
+            return;
+    }
+
+    _mesa_MultiDrawArraysEXT(mode, first, count, primcount);
+}
+
+extern void GL_APIENTRY _mesa_MultiDrawElementsEXT(GLenum mode, const GLsizei * count, GLenum type, const GLvoid ** indices, GLsizei primcount);
+static void GL_APIENTRY _es_MultiDrawElementsEXT(GLenum mode, const GLsizei *count, GLenum type, const GLvoid **indices, GLsizei primcount)
+{
+    switch(mode) {
+        case GL_POINTS:
+        case GL_LINES:
+        case GL_LINE_LOOP:
+        case GL_LINE_STRIP:
+        case GL_TRIANGLES:
+        case GL_TRIANGLE_STRIP:
+        case GL_TRIANGLE_FAN:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiDrawElementsEXT(mode=0x%x)", mode);
+            return;
+    }
+    switch(type) {
+        case GL_UNSIGNED_BYTE:
+        case GL_UNSIGNED_SHORT:
+        case GL_UNSIGNED_INT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glMultiDrawElementsEXT(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_MultiDrawElementsEXT(mode, count, type, indices, primcount);
+}
+
+extern void GL_APIENTRY _mesa_PixelStorei(GLenum pname, GLint param);
+static void GL_APIENTRY _es_PixelStorei(GLenum pname, GLint param)
+{
+    switch(pname) {
+        case GL_PACK_ALIGNMENT:
+        case GL_UNPACK_ALIGNMENT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glPixelStorei(pname=0x%x)", pname);
+            return;
+    }
+    switch(param) {
+        case 1:
+        case 2:
+        case 4:
+        case 8:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glPixelStorei(param=%d)", param);
+            return;
+    }
+
+    _mesa_PixelStorei(pname, param);
+}
+
+extern void GL_APIENTRY _mesa_PolygonOffset(GLfloat factor, GLfloat units);
+#define _es_PolygonOffset _mesa_PolygonOffset
+
+extern void GL_APIENTRY _mesa_ReadPixels(GLint x, GLint y, GLsizei width, GLsizei height, GLenum format, GLenum type, GLvoid * pixels);
+static void GL_APIENTRY _es_ReadPixels(GLint x, GLint y, GLsizei width, GLsizei height, GLenum format, GLenum type, GLvoid *pixels)
+{
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glReadPixels(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_ReadPixels(x, y, width, height, format, type, pixels);
+}
+
+extern void GL_APIENTRY _mesa_ReleaseShaderCompiler(void);
+#define _es_ReleaseShaderCompiler _mesa_ReleaseShaderCompiler
+
+/* this function is special and is defined elsewhere */
+extern void GL_APIENTRY _es_RenderbufferStorageEXT(GLenum target, GLenum internalFormat, GLsizei width, GLsizei height);
+static void GL_APIENTRY _check_RenderbufferStorage(GLenum target, GLenum internalFormat, GLsizei width, GLsizei height)
+{
+    switch(target) {
+        case GL_RENDERBUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glRenderbufferStorage(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_DEPTH_COMPONENT16:
+        case GL_RGBA4:
+        case GL_RGB5_A1:
+        case GL_RGB565:
+        case GL_STENCIL_INDEX8:
+        case GL_DEPTH_COMPONENT24_OES:
+        case GL_DEPTH_COMPONENT32_OES:
+        case GL_RGB8_OES:
+        case GL_RGBA8_OES:
+        case GL_STENCIL_INDEX1_OES:
+        case GL_STENCIL_INDEX4_OES:
+        case GL_DEPTH24_STENCIL8_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glRenderbufferStorage(internalFormat=0x%x)", internalFormat);
+            return;
+    }
+
+    _es_RenderbufferStorageEXT(target, internalFormat, width, height);
+}
+
+extern void GL_APIENTRY _mesa_SampleCoverageARB(GLclampf value, GLboolean invert);
+#define _es_SampleCoverage _mesa_SampleCoverageARB
+
+extern void GL_APIENTRY _mesa_Scissor(GLint x, GLint y, GLsizei width, GLsizei height);
+#define _es_Scissor _mesa_Scissor
+
+extern void GL_APIENTRY _mesa_ShaderBinary(GLsizei n, const GLuint * shaders, GLenum binaryformat, const GLvoid * binary, GLsizei length);
+#define _es_ShaderBinary _mesa_ShaderBinary
+
+extern void GL_APIENTRY _mesa_ShaderSourceARB(GLuint shader, GLsizei count, const GLchar ** string, const int * length);
+#define _es_ShaderSource _mesa_ShaderSourceARB
+
+extern void GL_APIENTRY _mesa_StencilFunc(GLenum func, GLint ref, GLuint mask);
+static void GL_APIENTRY _es_StencilFunc(GLenum func, GLint ref, GLuint mask)
+{
+    switch(func) {
+        case GL_NEVER:
+        case GL_LESS:
+        case GL_LEQUAL:
+        case GL_GREATER:
+        case GL_GEQUAL:
+        case GL_EQUAL:
+        case GL_NOTEQUAL:
+        case GL_ALWAYS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilFunc(func=0x%x)", func);
+            return;
+    }
+
+    _mesa_StencilFunc(func, ref, mask);
+}
+
+extern void GL_APIENTRY _mesa_StencilFuncSeparate(GLenum face, GLenum func, GLint ref, GLuint mask);
+static void GL_APIENTRY _es_StencilFuncSeparate(GLenum face, GLenum func, GLint ref, GLuint mask)
+{
+    switch(face) {
+        case GL_FRONT:
+        case GL_BACK:
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilFuncSeparate(face=0x%x)", face);
+            return;
+    }
+    switch(func) {
+        case GL_NEVER:
+        case GL_LESS:
+        case GL_LEQUAL:
+        case GL_GREATER:
+        case GL_GEQUAL:
+        case GL_EQUAL:
+        case GL_NOTEQUAL:
+        case GL_ALWAYS:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilFuncSeparate(func=0x%x)", func);
+            return;
+    }
+
+    _mesa_StencilFuncSeparate(face, func, ref, mask);
+}
+
+extern void GL_APIENTRY _mesa_StencilMask(GLuint mask);
+#define _es_StencilMask _mesa_StencilMask
+
+extern void GL_APIENTRY _mesa_StencilMaskSeparate(GLenum face, GLuint mask);
+static void GL_APIENTRY _es_StencilMaskSeparate(GLenum face, GLuint mask)
+{
+    switch(face) {
+        case GL_FRONT:
+        case GL_BACK:
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilMaskSeparate(face=0x%x)", face);
+            return;
+    }
+
+    _mesa_StencilMaskSeparate(face, mask);
+}
+
+extern void GL_APIENTRY _mesa_StencilOp(GLenum fail, GLenum zfail, GLenum zpass);
+static void GL_APIENTRY _es_StencilOp(GLenum fail, GLenum zfail, GLenum zpass)
+{
+    switch(fail) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP:
+        case GL_DECR_WRAP:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOp(fail=0x%x)", fail);
+            return;
+    }
+    switch(zfail) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP:
+        case GL_DECR_WRAP:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOp(zfail=0x%x)", zfail);
+            return;
+    }
+    switch(zpass) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP:
+        case GL_DECR_WRAP:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOp(zpass=0x%x)", zpass);
+            return;
+    }
+
+    _mesa_StencilOp(fail, zfail, zpass);
+}
+
+extern void GL_APIENTRY _mesa_StencilOpSeparate(GLenum face, GLenum fail, GLenum zfail, GLenum zpass);
+static void GL_APIENTRY _es_StencilOpSeparate(GLenum face, GLenum fail, GLenum zfail, GLenum zpass)
+{
+    switch(face) {
+        case GL_FRONT:
+        case GL_BACK:
+        case GL_FRONT_AND_BACK:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOpSeparate(face=0x%x)", face);
+            return;
+    }
+    switch(fail) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP:
+        case GL_DECR_WRAP:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOpSeparate(fail=0x%x)", fail);
+            return;
+    }
+    switch(zfail) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP:
+        case GL_DECR_WRAP:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOpSeparate(zfail=0x%x)", zfail);
+            return;
+    }
+    switch(zpass) {
+        case GL_KEEP:
+        case GL_ZERO:
+        case GL_REPLACE:
+        case GL_INCR:
+        case GL_DECR:
+        case GL_INVERT:
+        case GL_INCR_WRAP:
+        case GL_DECR_WRAP:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glStencilOpSeparate(zpass=0x%x)", zpass);
+            return;
+    }
+
+    _mesa_StencilOpSeparate(face, fail, zfail, zpass);
+}
+
+extern void GL_APIENTRY _mesa_TexImage2D(GLenum target, GLint level, GLint internalFormat, GLsizei width, GLsizei height, GLint border, GLenum format, GLenum type, const GLvoid * pixels);
+static void GL_APIENTRY _es_TexImage2D(GLenum target, GLint level, GLint internalFormat, GLsizei width, GLsizei height, GLint border, GLenum format, GLenum type, const GLvoid *pixels)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_ALPHA:
+        case GL_RGB:
+        case GL_RGBA:
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+        case GL_DEPTH_COMPONENT:
+        case GL_DEPTH_STENCIL_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexImage2D(internalFormat=%d)", internalFormat);
+            return;
+    }
+    switch(border) {
+        case 0:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glTexImage2D(border=%d)", border);
+            return;
+    }
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1 && type != GL_UNSIGNED_INT_2_10_10_10_REV_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_DEPTH_COMPONENT:
+            if (type != GL_UNSIGNED_SHORT && type != GL_UNSIGNED_INT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_DEPTH_STENCIL_OES:
+            if (type != GL_UNSIGNED_INT_24_8_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage2D(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_TexImage2D(target, level, internalFormat, width, height, border, format, type, pixels);
+}
+
+extern void GL_APIENTRY _mesa_TexImage3D(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLsizei depth, GLint border, GLenum format, GLenum type, const GLvoid * pixels);
+static void GL_APIENTRY _es_TexImage3DOES(GLenum target, GLint level, GLenum internalFormat, GLsizei width, GLsizei height, GLsizei depth, GLint border, GLenum format, GLenum type, const GLvoid *pixels)
+{
+    switch(target) {
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexImage3DOES(target=0x%x)", target);
+            return;
+    }
+    switch(internalFormat) {
+        case GL_ALPHA:
+        case GL_RGB:
+        case GL_RGBA:
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexImage3DOES(internalFormat=0x%x)", internalFormat);
+            return;
+    }
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1 && type != GL_UNSIGNED_INT_2_10_10_10_REV_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexImage3DOES(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_TexImage3D(target, level, internalFormat, width, height, depth, border, format, type, pixels);
+}
+
+extern void GL_APIENTRY _mesa_TexParameterf(GLenum target, GLenum pname, GLfloat param);
+static void GL_APIENTRY _es_TexParameterf(GLenum target, GLenum pname, GLfloat param)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S:
+        case GL_TEXTURE_WRAP_T:
+        case GL_TEXTURE_WRAP_R_OES:
+            if (param != GL_CLAMP_TO_EDGE && param != GL_REPEAT && param != GL_MIRRORED_REPEAT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR && param != GL_NEAREST_MIPMAP_NEAREST && param != GL_NEAREST_MIPMAP_LINEAR && param != GL_LINEAR_MIPMAP_NEAREST && param != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterf(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameterf(target, pname, param);
+}
+
+extern void GL_APIENTRY _mesa_TexParameterfv(GLenum target, GLenum pname, const GLfloat * params);
+static void GL_APIENTRY _es_TexParameterfv(GLenum target, GLenum pname, const GLfloat *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+        case GL_TEXTURE_WRAP_R_OES: /* size 1 */
+            if (params[0] != GL_CLAMP_TO_EDGE && params[0] != GL_REPEAT && params[0] != GL_MIRRORED_REPEAT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR && params[0] != GL_NEAREST_MIPMAP_NEAREST && params[0] != GL_NEAREST_MIPMAP_LINEAR && params[0] != GL_LINEAR_MIPMAP_NEAREST && params[0] != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameterfv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameterfv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_TexParameteri(GLenum target, GLenum pname, GLint param);
+static void GL_APIENTRY _es_TexParameteri(GLenum target, GLenum pname, GLint param)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S:
+        case GL_TEXTURE_WRAP_T:
+        case GL_TEXTURE_WRAP_R_OES:
+            if (param != GL_CLAMP_TO_EDGE && param != GL_REPEAT && param != GL_MIRRORED_REPEAT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR && param != GL_NEAREST_MIPMAP_NEAREST && param != GL_NEAREST_MIPMAP_LINEAR && param != GL_LINEAR_MIPMAP_NEAREST && param != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER:
+            if (param != GL_NEAREST && param != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteri(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameteri(target, pname, param);
+}
+
+extern void GL_APIENTRY _mesa_TexParameteriv(GLenum target, GLenum pname, const GLint * params);
+static void GL_APIENTRY _es_TexParameteriv(GLenum target, GLenum pname, const GLint *params)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP:
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(target=0x%x)", target);
+            return;
+    }
+    switch(pname) {
+        case GL_TEXTURE_WRAP_S: /* size 1 */
+        case GL_TEXTURE_WRAP_T: /* size 1 */
+        case GL_TEXTURE_WRAP_R_OES: /* size 1 */
+            if (params[0] != GL_CLAMP_TO_EDGE && params[0] != GL_REPEAT && params[0] != GL_MIRRORED_REPEAT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MIN_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR && params[0] != GL_NEAREST_MIPMAP_NEAREST && params[0] != GL_NEAREST_MIPMAP_LINEAR && params[0] != GL_LINEAR_MIPMAP_NEAREST && params[0] != GL_LINEAR_MIPMAP_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAG_FILTER: /* size 1 */
+            if (params[0] != GL_NEAREST && params[0] != GL_LINEAR) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+                return;
+            }
+            break;
+        case GL_TEXTURE_MAX_ANISOTROPY_EXT: /* size 1 */
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexParameteriv(pname=0x%x)", pname);
+            return;
+    }
+
+    _mesa_TexParameteriv(target, pname, params);
+}
+
+extern void GL_APIENTRY _mesa_TexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid * pixels);
+static void GL_APIENTRY _es_TexSubImage2D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLsizei width, GLsizei height, GLenum format, GLenum type, const GLvoid *pixels)
+{
+    switch(target) {
+        case GL_TEXTURE_2D:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_X:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_POSITIVE_Z:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_X:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Y:
+        case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexSubImage2D(target=0x%x)", target);
+            return;
+    }
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1 && type != GL_UNSIGNED_INT_2_10_10_10_REV_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_DEPTH_COMPONENT:
+            if (type != GL_UNSIGNED_SHORT && type != GL_UNSIGNED_INT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_DEPTH_STENCIL_OES:
+            if (type != GL_UNSIGNED_INT_24_8_OES) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage2D(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_TexSubImage2D(target, level, xoffset, yoffset, width, height, format, type, pixels);
+}
+
+extern void GL_APIENTRY _mesa_TexSubImage3D(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLsizei width, GLsizei height, GLsizei depth, GLenum format, GLenum type, const GLvoid * pixels);
+static void GL_APIENTRY _es_TexSubImage3DOES(GLenum target, GLint level, GLint xoffset, GLint yoffset, GLint zoffset, GLsizei width, GLsizei height, GLsizei depth, GLenum format, GLenum type, const GLvoid *pixels)
+{
+    switch(target) {
+        case GL_TEXTURE_3D_OES:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glTexSubImage3DOES(target=0x%x)", target);
+            return;
+    }
+    switch(format) {
+        case GL_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGB:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_5_6_5) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_RGBA:
+            if (type != GL_UNSIGNED_BYTE && type != GL_UNSIGNED_SHORT_4_4_4_4 && type != GL_UNSIGNED_SHORT_5_5_5_1 && type != GL_UNSIGNED_INT_2_10_10_10_REV_EXT) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        case GL_LUMINANCE:
+        case GL_LUMINANCE_ALPHA:
+            if (type != GL_UNSIGNED_BYTE) {
+                _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage3DOES(format=0x%x)", format);
+                return;
+            }
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_OPERATION, "glTexSubImage3DOES(format=0x%x)", format);
+            return;
+    }
+
+    _mesa_TexSubImage3D(target, level, xoffset, yoffset, zoffset, width, height, depth, format, type, pixels);
+}
+
+extern void GL_APIENTRY _mesa_Uniform1fARB(GLint location, GLfloat v0);
+#define _es_Uniform1f _mesa_Uniform1fARB
+
+extern void GL_APIENTRY _mesa_Uniform1fvARB(GLint location, GLsizei count, const GLfloat * values);
+#define _es_Uniform1fv _mesa_Uniform1fvARB
+
+extern void GL_APIENTRY _mesa_Uniform1iARB(GLint location, GLint v0);
+#define _es_Uniform1i _mesa_Uniform1iARB
+
+extern void GL_APIENTRY _mesa_Uniform1ivARB(GLint location, GLsizei count, const GLint * values);
+#define _es_Uniform1iv _mesa_Uniform1ivARB
+
+extern void GL_APIENTRY _mesa_Uniform2fARB(GLint location, GLfloat v0, GLfloat v1);
+#define _es_Uniform2f _mesa_Uniform2fARB
+
+extern void GL_APIENTRY _mesa_Uniform2fvARB(GLint location, GLsizei count, const GLfloat * values);
+#define _es_Uniform2fv _mesa_Uniform2fvARB
+
+extern void GL_APIENTRY _mesa_Uniform2iARB(GLint location, GLint v0, GLint v1);
+#define _es_Uniform2i _mesa_Uniform2iARB
+
+extern void GL_APIENTRY _mesa_Uniform2ivARB(GLint location, GLsizei count, const GLint * values);
+#define _es_Uniform2iv _mesa_Uniform2ivARB
+
+extern void GL_APIENTRY _mesa_Uniform3fARB(GLint location, GLfloat v0, GLfloat v1, GLfloat v2);
+#define _es_Uniform3f _mesa_Uniform3fARB
+
+extern void GL_APIENTRY _mesa_Uniform3fvARB(GLint location, GLsizei count, const GLfloat * values);
+#define _es_Uniform3fv _mesa_Uniform3fvARB
+
+extern void GL_APIENTRY _mesa_Uniform3iARB(GLint location, GLint v0, GLint v1, GLint v2);
+#define _es_Uniform3i _mesa_Uniform3iARB
+
+extern void GL_APIENTRY _mesa_Uniform3ivARB(GLint location, GLsizei count, const GLint * values);
+#define _es_Uniform3iv _mesa_Uniform3ivARB
+
+extern void GL_APIENTRY _mesa_Uniform4fARB(GLint location, GLfloat v0, GLfloat v1, GLfloat v2, GLfloat v3);
+#define _es_Uniform4f _mesa_Uniform4fARB
+
+extern void GL_APIENTRY _mesa_Uniform4fvARB(GLint location, GLsizei count, const GLfloat * values);
+#define _es_Uniform4fv _mesa_Uniform4fvARB
+
+extern void GL_APIENTRY _mesa_Uniform4iARB(GLint location, GLint v0, GLint v1, GLint v2, GLint v3);
+#define _es_Uniform4i _mesa_Uniform4iARB
+
+extern void GL_APIENTRY _mesa_Uniform4ivARB(GLint location, GLsizei count, const GLint * values);
+#define _es_Uniform4iv _mesa_Uniform4ivARB
+
+extern void GL_APIENTRY _mesa_UniformMatrix2fvARB(GLint location, GLsizei count, GLboolean transpose, const GLfloat * value);
+#define _es_UniformMatrix2fv _mesa_UniformMatrix2fvARB
+
+extern void GL_APIENTRY _mesa_UniformMatrix3fvARB(GLint location, GLsizei count, GLboolean transpose, const GLfloat * value);
+#define _es_UniformMatrix3fv _mesa_UniformMatrix3fvARB
+
+extern void GL_APIENTRY _mesa_UniformMatrix4fvARB(GLint location, GLsizei count, GLboolean transpose, const GLfloat * value);
+#define _es_UniformMatrix4fv _mesa_UniformMatrix4fvARB
+
+extern GLboolean GL_APIENTRY _mesa_UnmapBufferARB(GLenum target);
+static GLboolean GL_APIENTRY _es_UnmapBufferOES(GLenum target)
+{
+    switch(target) {
+        case GL_ARRAY_BUFFER:
+        case GL_ELEMENT_ARRAY_BUFFER:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_ENUM, "glUnmapBufferOES(target=0x%x)", target);
+            return GL_FALSE;
+    }
+
+    return _mesa_UnmapBufferARB(target);
+}
+
+extern void GL_APIENTRY _mesa_UseProgramObjectARB(GLuint program);
+#define _es_UseProgram _mesa_UseProgramObjectARB
+
+extern void GL_APIENTRY _mesa_ValidateProgramARB(GLuint program);
+#define _es_ValidateProgram _mesa_ValidateProgramARB
+
+extern void GL_APIENTRY _vbo_VertexAttrib1f(GLuint index, GLfloat x);
+#define _es_VertexAttrib1f _vbo_VertexAttrib1f
+
+extern void GL_APIENTRY _vbo_VertexAttrib1fv(GLuint index, const GLfloat * v);
+#define _es_VertexAttrib1fv _vbo_VertexAttrib1fv
+
+extern void GL_APIENTRY _vbo_VertexAttrib2f(GLuint index, GLfloat x, GLfloat y);
+#define _es_VertexAttrib2f _vbo_VertexAttrib2f
+
+extern void GL_APIENTRY _vbo_VertexAttrib2fv(GLuint index, const GLfloat * v);
+#define _es_VertexAttrib2fv _vbo_VertexAttrib2fv
+
+extern void GL_APIENTRY _vbo_VertexAttrib3f(GLuint index, GLfloat x, GLfloat y, GLfloat z);
+#define _es_VertexAttrib3f _vbo_VertexAttrib3f
+
+extern void GL_APIENTRY _vbo_VertexAttrib3fv(GLuint index, const GLfloat * v);
+#define _es_VertexAttrib3fv _vbo_VertexAttrib3fv
+
+extern void GL_APIENTRY _vbo_VertexAttrib4f(GLuint index, GLfloat x, GLfloat y, GLfloat z, GLfloat w);
+#define _es_VertexAttrib4f _vbo_VertexAttrib4f
+
+extern void GL_APIENTRY _vbo_VertexAttrib4fv(GLuint index, const GLfloat * v);
+#define _es_VertexAttrib4fv _vbo_VertexAttrib4fv
+
+extern void GL_APIENTRY _mesa_VertexAttribPointerARB(GLuint index, GLint size, GLenum type, GLboolean normalized, GLsizei stride, const GLvoid * pointer);
+static void GL_APIENTRY _es_VertexAttribPointer(GLuint index, GLint size, GLenum type, GLboolean normalized, GLsizei stride, const GLvoid *pointer)
+{
+    switch(size) {
+        case 1:
+        case 2:
+        case 3:
+        case 4:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glVertexAttribPointer(size=%d)", size);
+            return;
+    }
+    switch(type) {
+        case GL_BYTE:
+        case GL_UNSIGNED_BYTE:
+        case GL_SHORT:
+        case GL_UNSIGNED_SHORT:
+        case GL_FLOAT:
+        case GL_FIXED:
+            break;
+        default:
+            _mesa_error(_mesa_get_current_context(), GL_INVALID_VALUE, "glVertexAttribPointer(type=0x%x)", type);
+            return;
+    }
+
+    _mesa_VertexAttribPointerARB(index, size, type, normalized, stride, pointer);
+}
+
+extern void GL_APIENTRY _mesa_Viewport(GLint x, GLint y, GLsizei width, GLsizei height);
+#define _es_Viewport _mesa_Viewport
+
+
+struct _glapi_table *
+_mesa_create_exec_table_es2(void)
+{
+   struct _glapi_table *exec;
+   exec = _mesa_alloc_dispatch_table(sizeof *exec);
+   if (exec == NULL)
+      return NULL;
+
+
+    SET_ActiveTexture(exec, _es_ActiveTexture);
+    SET_AttachShader(exec, _es_AttachShader);
+    SET_BindAttribLocation(exec, _es_BindAttribLocation);
+    SET_BindBuffer(exec, _es_BindBuffer);
+    SET_BindFramebuffer(exec, _es_BindFramebuffer);
+    SET_BindRenderbuffer(exec, _es_BindRenderbuffer);
+    SET_BindTexture(exec, _es_BindTexture);
+    SET_BlendColor(exec, _es_BlendColor);
+    SET_BlendEquation(exec, _es_BlendEquation);
+    SET_BlendEquationSeparate(exec, _es_BlendEquationSeparate);
+    SET_BlendFunc(exec, _es_BlendFunc);
+    SET_BlendFuncSeparate(exec, _es_BlendFuncSeparate);
+    SET_BufferData(exec, _es_BufferData);
+    SET_BufferSubData(exec, _es_BufferSubData);
+    SET_CheckFramebufferStatus(exec, _es_CheckFramebufferStatus);
+    SET_Clear(exec, _es_Clear);
+    SET_ClearColor(exec, _es_ClearColor);
+    SET_ClearDepthf(exec, _es_ClearDepthf);
+    SET_ClearStencil(exec, _es_ClearStencil);
+    SET_ColorMask(exec, _es_ColorMask);
+    SET_CompileShader(exec, _es_CompileShader);
+    SET_CompressedTexImage2D(exec, _es_CompressedTexImage2D);
+    SET_CompressedTexImage3DOES(exec, _es_CompressedTexImage3DOES);
+    SET_CompressedTexSubImage2D(exec, _es_CompressedTexSubImage2D);
+    SET_CompressedTexSubImage3DOES(exec, _es_CompressedTexSubImage3DOES);
+    SET_CopyTexImage2D(exec, _es_CopyTexImage2D);
+    SET_CopyTexSubImage2D(exec, _es_CopyTexSubImage2D);
+    SET_CopyTexSubImage3DOES(exec, _es_CopyTexSubImage3DOES);
+    SET_CreateProgram(exec, _es_CreateProgram);
+    SET_CreateShader(exec, _es_CreateShader);
+    SET_CullFace(exec, _es_CullFace);
+    SET_DeleteBuffers(exec, _es_DeleteBuffers);
+    SET_DeleteFramebuffers(exec, _es_DeleteFramebuffers);
+    SET_DeleteProgram(exec, _es_DeleteProgram);
+    SET_DeleteRenderbuffers(exec, _es_DeleteRenderbuffers);
+    SET_DeleteShader(exec, _es_DeleteShader);
+    SET_DeleteTextures(exec, _es_DeleteTextures);
+    SET_DepthFunc(exec, _es_DepthFunc);
+    SET_DepthMask(exec, _es_DepthMask);
+    SET_DepthRangef(exec, _es_DepthRangef);
+    SET_DetachShader(exec, _es_DetachShader);
+    SET_Disable(exec, _es_Disable);
+    SET_DisableVertexAttribArray(exec, _es_DisableVertexAttribArray);
+    SET_DrawArrays(exec, _es_DrawArrays);
+    SET_DrawElements(exec, _es_DrawElements);
+    SET_EGLImageTargetRenderbufferStorageOES(exec, _es_EGLImageTargetRenderbufferStorageOES);
+    SET_EGLImageTargetTexture2DOES(exec, _es_EGLImageTargetTexture2DOES);
+    SET_Enable(exec, _es_Enable);
+    SET_EnableVertexAttribArray(exec, _es_EnableVertexAttribArray);
+    SET_Finish(exec, _es_Finish);
+    SET_Flush(exec, _es_Flush);
+    SET_FramebufferRenderbuffer(exec, _es_FramebufferRenderbuffer);
+    SET_FramebufferTexture2D(exec, _es_FramebufferTexture2D);
+    SET_FramebufferTexture3DOES(exec, _es_FramebufferTexture3DOES);
+    SET_FrontFace(exec, _es_FrontFace);
+    SET_GenBuffers(exec, _es_GenBuffers);
+    SET_GenFramebuffers(exec, _es_GenFramebuffers);
+    SET_GenRenderbuffers(exec, _es_GenRenderbuffers);
+    SET_GenTextures(exec, _es_GenTextures);
+    SET_GenerateMipmap(exec, _es_GenerateMipmap);
+    SET_GetActiveAttrib(exec, _es_GetActiveAttrib);
+    SET_GetActiveUniform(exec, _es_GetActiveUniform);
+    SET_GetAttachedShaders(exec, _es_GetAttachedShaders);
+    SET_GetAttribLocation(exec, _es_GetAttribLocation);
+    SET_GetBooleanv(exec, _es_GetBooleanv);
+    SET_GetBufferParameteriv(exec, _es_GetBufferParameteriv);
+    SET_GetBufferPointervOES(exec, _es_GetBufferPointervOES);
+    SET_GetError(exec, _es_GetError);
+    SET_GetFloatv(exec, _es_GetFloatv);
+    SET_GetFramebufferAttachmentParameteriv(exec, _es_GetFramebufferAttachmentParameteriv);
+    SET_GetIntegerv(exec, _es_GetIntegerv);
+    SET_GetProgramInfoLog(exec, _es_GetProgramInfoLog);
+    SET_GetProgramiv(exec, _es_GetProgramiv);
+    SET_GetRenderbufferParameteriv(exec, _es_GetRenderbufferParameteriv);
+    SET_GetShaderInfoLog(exec, _es_GetShaderInfoLog);
+    SET_GetShaderPrecisionFormat(exec, _es_GetShaderPrecisionFormat);
+    SET_GetShaderSource(exec, _es_GetShaderSource);
+    SET_GetShaderiv(exec, _es_GetShaderiv);
+    SET_GetString(exec, _es_GetString);
+    SET_GetTexParameterfv(exec, _es_GetTexParameterfv);
+    SET_GetTexParameteriv(exec, _es_GetTexParameteriv);
+    SET_GetUniformLocation(exec, _es_GetUniformLocation);
+    SET_GetUniformfv(exec, _es_GetUniformfv);
+    SET_GetUniformiv(exec, _es_GetUniformiv);
+    SET_GetVertexAttribPointerv(exec, _es_GetVertexAttribPointerv);
+    SET_GetVertexAttribfv(exec, _es_GetVertexAttribfv);
+    SET_GetVertexAttribiv(exec, _es_GetVertexAttribiv);
+    SET_Hint(exec, _es_Hint);
+    SET_IsBuffer(exec, _es_IsBuffer);
+    SET_IsEnabled(exec, _es_IsEnabled);
+    SET_IsFramebuffer(exec, _es_IsFramebuffer);
+    SET_IsProgram(exec, _es_IsProgram);
+    SET_IsRenderbuffer(exec, _es_IsRenderbuffer);
+    SET_IsShader(exec, _es_IsShader);
+    SET_IsTexture(exec, _es_IsTexture);
+    SET_LineWidth(exec, _es_LineWidth);
+    SET_LinkProgram(exec, _es_LinkProgram);
+    SET_MapBufferOES(exec, _es_MapBufferOES);
+    SET_MultiDrawArraysEXT(exec, _es_MultiDrawArraysEXT);
+    SET_MultiDrawElementsEXT(exec, _es_MultiDrawElementsEXT);
+    SET_PixelStorei(exec, _es_PixelStorei);
+    SET_PolygonOffset(exec, _es_PolygonOffset);
+    SET_ReadPixels(exec, _es_ReadPixels);
+    SET_ReleaseShaderCompiler(exec, _es_ReleaseShaderCompiler);
+    SET_RenderbufferStorage(exec, _check_RenderbufferStorage);
+    SET_SampleCoverage(exec, _es_SampleCoverage);
+    SET_Scissor(exec, _es_Scissor);
+    SET_ShaderBinary(exec, _es_ShaderBinary);
+    SET_ShaderSource(exec, _es_ShaderSource);
+    SET_StencilFunc(exec, _es_StencilFunc);
+    SET_StencilFuncSeparate(exec, _es_StencilFuncSeparate);
+    SET_StencilMask(exec, _es_StencilMask);
+    SET_StencilMaskSeparate(exec, _es_StencilMaskSeparate);
+    SET_StencilOp(exec, _es_StencilOp);
+    SET_StencilOpSeparate(exec, _es_StencilOpSeparate);
+    SET_TexImage2D(exec, _es_TexImage2D);
+    SET_TexImage3DOES(exec, _es_TexImage3DOES);
+    SET_TexParameterf(exec, _es_TexParameterf);
+    SET_TexParameterfv(exec, _es_TexParameterfv);
+    SET_TexParameteri(exec, _es_TexParameteri);
+    SET_TexParameteriv(exec, _es_TexParameteriv);
+    SET_TexSubImage2D(exec, _es_TexSubImage2D);
+    SET_TexSubImage3DOES(exec, _es_TexSubImage3DOES);
+    SET_Uniform1f(exec, _es_Uniform1f);
+    SET_Uniform1fv(exec, _es_Uniform1fv);
+    SET_Uniform1i(exec, _es_Uniform1i);
+    SET_Uniform1iv(exec, _es_Uniform1iv);
+    SET_Uniform2f(exec, _es_Uniform2f);
+    SET_Uniform2fv(exec, _es_Uniform2fv);
+    SET_Uniform2i(exec, _es_Uniform2i);
+    SET_Uniform2iv(exec, _es_Uniform2iv);
+    SET_Uniform3f(exec, _es_Uniform3f);
+    SET_Uniform3fv(exec, _es_Uniform3fv);
+    SET_Uniform3i(exec, _es_Uniform3i);
+    SET_Uniform3iv(exec, _es_Uniform3iv);
+    SET_Uniform4f(exec, _es_Uniform4f);
+    SET_Uniform4fv(exec, _es_Uniform4fv);
+    SET_Uniform4i(exec, _es_Uniform4i);
+    SET_Uniform4iv(exec, _es_Uniform4iv);
+    SET_UniformMatrix2fv(exec, _es_UniformMatrix2fv);
+    SET_UniformMatrix3fv(exec, _es_UniformMatrix3fv);
+    SET_UniformMatrix4fv(exec, _es_UniformMatrix4fv);
+    SET_UnmapBufferOES(exec, _es_UnmapBufferOES);
+    SET_UseProgram(exec, _es_UseProgram);
+    SET_ValidateProgram(exec, _es_ValidateProgram);
+    SET_VertexAttrib1f(exec, _es_VertexAttrib1f);
+    SET_VertexAttrib1fv(exec, _es_VertexAttrib1fv);
+    SET_VertexAttrib2f(exec, _es_VertexAttrib2f);
+    SET_VertexAttrib2fv(exec, _es_VertexAttrib2fv);
+    SET_VertexAttrib3f(exec, _es_VertexAttrib3f);
+    SET_VertexAttrib3fv(exec, _es_VertexAttrib3fv);
+    SET_VertexAttrib4f(exec, _es_VertexAttrib4f);
+    SET_VertexAttrib4fv(exec, _es_VertexAttrib4fv);
+    SET_VertexAttribPointer(exec, _es_VertexAttribPointer);
+    SET_Viewport(exec, _es_Viewport);
+
+   return exec;
+}
+
+#endif /* FEATURE_ES2 */