fontconfig mesa git update 5 Mar 2015

314 files changed, 2630 insertions, 4900 deletions

diff --git a/mesalib/src/SConscript b/mesalib/src/SConscript
index c25adc728..188ab08b6 100644
--- a/mesalib/src/SConscript
+++ b/mesalib/src/SConscript
@@ -27,8 +27,6 @@ if env['platform'] in ['haiku']:
 
 SConscript('mesa/SConscript')
 
-SConscript('mapi/vgapi/SConscript')
-
 if not env['embedded']:
     if env['platform'] not in ('cygwin', 'darwin', 'freebsd', 'haiku', 'windows'):
         SConscript('glx/SConscript')
diff --git a/mesalib/src/gallium/Android.mk b/mesalib/src/gallium/Android.mk
index bf99e4fb6..b2662ffca 100644
--- a/mesalib/src/gallium/Android.mk
+++ b/mesalib/src/gallium/Android.mk
@@ -79,10 +79,5 @@ ifneq ($(filter vmwgfx, $(MESA_GPU_DRIVERS)),)
 SUBDIRS += winsys/svga/drm drivers/svga
 endif
 
-#
-# Gallium state trackers and their users (targets)
-#
-SUBDIRS += state_trackers/egl targets/egl-static
-
 mkfiles := $(patsubst %,$(GALLIUM_TOP)/%/Android.mk,$(SUBDIRS))
 include $(mkfiles)
diff --git a/mesalib/src/gallium/Makefile.am b/mesalib/src/gallium/Makefile.am
index 6a2a86654..652cb1310 100644
--- a/mesalib/src/gallium/Makefile.am
+++ b/mesalib/src/gallium/Makefile.am
@@ -12,7 +12,6 @@ SUBDIRS += auxiliary
 
 SUBDIRS += \
 	drivers/galahad \
-	drivers/identity \
 	drivers/noop \
 	drivers/trace \
 	drivers/rbug
@@ -115,9 +114,7 @@ EXTRA_DIST = \
 	docs \
 	README.portability \
 	SConscript \
-	state_trackers/egl \
 	state_trackers/gbm \
-	targets/egl-static \
 	targets/gbm \
 	winsys/sw/gdi \
 	winsys/sw/hgl
@@ -139,11 +136,6 @@ if HAVE_DRICOMMON
 SUBDIRS += state_trackers/dri targets/dri
 endif
 
-## the egl target depends on vega
-if HAVE_OPENVG
-SUBDIRS += state_trackers/vega
-endif
-
 if HAVE_X11_DRIVER
 SUBDIRS += state_trackers/glx/xlib targets/libgl-xlib
 endif
diff --git a/mesalib/src/gallium/SConscript b/mesalib/src/gallium/SConscript
index 25d0af6f9..9b9011f7f 100644
--- a/mesalib/src/gallium/SConscript
+++ b/mesalib/src/gallium/SConscript
@@ -13,7 +13,6 @@ SConscript('auxiliary/SConscript')
 # These are common and work across all platforms
 SConscript([
     'drivers/galahad/SConscript',
-    'drivers/identity/SConscript',
     'drivers/llvmpipe/SConscript',
     'drivers/rbug/SConscript',
     'drivers/softpipe/SConscript',
@@ -62,13 +61,6 @@ SConscript([
 ])
 
 if not env['embedded']:
-    SConscript('state_trackers/vega/SConscript')
-    if env['platform'] not in ('cygwin', 'darwin', 'freebsd', 'sunos'):
-        SConscript([
-            'state_trackers/egl/SConscript',
-            'targets/egl-static/SConscript',
-        ])
-
     if env['x11']:
         SConscript([
             'state_trackers/glx/xlib/SConscript',
diff --git a/mesalib/src/gallium/auxiliary/Makefile.am b/mesalib/src/gallium/auxiliary/Makefile.am
index 4b6205797..27a8b3fe4 100644
--- a/mesalib/src/gallium/auxiliary/Makefile.am
+++ b/mesalib/src/gallium/auxiliary/Makefile.am
@@ -12,9 +12,12 @@ noinst_LTLIBRARIES = libgallium.la
 AM_CFLAGS = \
 	-I$(top_srcdir)/src/gallium/auxiliary/util \
 	$(GALLIUM_CFLAGS) \
-	$(VISIBILITY_CFLAGS)
+	$(VISIBILITY_CFLAGS) \
+	$(MSVC2008_COMPAT_CXXFLAGS)
 
-AM_CXXFLAGS = $(VISIBILITY_CXXFLAGS)
+AM_CXXFLAGS = \
+	$(VISIBILITY_CXXFLAGS) \
+	$(MSVC2008_COMPAT_CXXFLAGS)
 
 libgallium_la_SOURCES = \
 	$(C_SOURCES) \
diff --git a/mesalib/src/gallium/auxiliary/SConscript b/mesalib/src/gallium/auxiliary/SConscript
index 6cb6b8c65..d5fa880c7 100644
--- a/mesalib/src/gallium/auxiliary/SConscript
+++ b/mesalib/src/gallium/auxiliary/SConscript
@@ -8,6 +8,10 @@ env.Append(CPPPATH = [
     'util',
 ])
 
+env = env.Clone()
+
+env.MSVC2008Compat()
+
 env.CodeGenerate(
     target = 'indices/u_indices_gen.c', 
     script = 'indices/u_indices_gen.py', 
diff --git a/mesalib/src/gallium/auxiliary/util/u_debug.c b/mesalib/src/gallium/auxiliary/util/u_debug.c
index d79f31ea9..2d2d049b2 100644
--- a/mesalib/src/gallium/auxiliary/util/u_debug.c
+++ b/mesalib/src/gallium/auxiliary/util/u_debug.c
@@ -402,31 +402,28 @@ void debug_print_format(const char *msg, unsigned fmt )
 #endif
 
 
-
-static const struct debug_named_value pipe_prim_names[] = {
-#ifdef DEBUG
-   DEBUG_NAMED_VALUE(PIPE_PRIM_POINTS),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_LINES),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_LOOP),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_STRIP),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLES),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_STRIP),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_FAN),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_QUADS),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_QUAD_STRIP),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_POLYGON),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_LINES_ADJACENCY),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_STRIP_ADJACENCY),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLES_ADJACENCY),
-   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY),
-#endif
-   DEBUG_NAMED_VALUE_END
-};
-
-
-const char *u_prim_name( unsigned prim )
+/** Return string name of given primitive type */
+const char *
+u_prim_name(unsigned prim)
 {
-   return debug_dump_enum(pipe_prim_names, prim);
+   static const struct debug_named_value names[] = {
+      DEBUG_NAMED_VALUE(PIPE_PRIM_POINTS),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_LINES),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_LOOP),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_STRIP),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLES),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_STRIP),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_FAN),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_QUADS),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_QUAD_STRIP),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_POLYGON),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_LINES_ADJACENCY),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_STRIP_ADJACENCY),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLES_ADJACENCY),
+      DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY),
+      DEBUG_NAMED_VALUE_END
+   };
+   return debug_dump_enum(names, prim);
 }
 
 
@@ -722,38 +719,75 @@ error1:
 void
 debug_print_transfer_flags(const char *msg, unsigned usage)
 {
-#define FLAG(x)  { x, #x }
-   static const struct {
-      unsigned bit;
-      const char *name;
-   } flags[] = {
-      FLAG(PIPE_TRANSFER_READ),
-      FLAG(PIPE_TRANSFER_WRITE),
-      FLAG(PIPE_TRANSFER_MAP_DIRECTLY),
-      FLAG(PIPE_TRANSFER_DISCARD_RANGE),
-      FLAG(PIPE_TRANSFER_DONTBLOCK),
-      FLAG(PIPE_TRANSFER_UNSYNCHRONIZED),
-      FLAG(PIPE_TRANSFER_FLUSH_EXPLICIT),
-      FLAG(PIPE_TRANSFER_DISCARD_WHOLE_RESOURCE)
+   static const struct debug_named_value names[] = {
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_READ),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_WRITE),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_MAP_DIRECTLY),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_DISCARD_RANGE),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_DONTBLOCK),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_UNSYNCHRONIZED),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_FLUSH_EXPLICIT),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_DISCARD_WHOLE_RESOURCE),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_PERSISTENT),
+      DEBUG_NAMED_VALUE(PIPE_TRANSFER_COHERENT),
+      DEBUG_NAMED_VALUE_END
    };
-   unsigned i;
 
-   debug_printf("%s ", msg);
+   debug_printf("%s: %s\n", msg, debug_dump_flags(names, usage));
+}
 
-   for (i = 0; i < Elements(flags); i++) {
-      if (usage & flags[i].bit) {
-         debug_printf("%s", flags[i].name);
-         usage &= ~flags[i].bit;
-         if (usage) {
-            debug_printf(" | ");
-         }
-      }
-   }
 
-   debug_printf("\n");
-#undef FLAG
+/**
+ * Print PIPE_BIND_x flags with a message.
+ */
+void
+debug_print_bind_flags(const char *msg, unsigned usage)
+{
+   static const struct debug_named_value names[] = {
+      DEBUG_NAMED_VALUE(PIPE_BIND_DEPTH_STENCIL),
+      DEBUG_NAMED_VALUE(PIPE_BIND_RENDER_TARGET),
+      DEBUG_NAMED_VALUE(PIPE_BIND_BLENDABLE),
+      DEBUG_NAMED_VALUE(PIPE_BIND_SAMPLER_VIEW),
+      DEBUG_NAMED_VALUE(PIPE_BIND_VERTEX_BUFFER),
+      DEBUG_NAMED_VALUE(PIPE_BIND_INDEX_BUFFER),
+      DEBUG_NAMED_VALUE(PIPE_BIND_CONSTANT_BUFFER),
+      DEBUG_NAMED_VALUE(PIPE_BIND_DISPLAY_TARGET),
+      DEBUG_NAMED_VALUE(PIPE_BIND_TRANSFER_WRITE),
+      DEBUG_NAMED_VALUE(PIPE_BIND_TRANSFER_READ),
+      DEBUG_NAMED_VALUE(PIPE_BIND_STREAM_OUTPUT),
+      DEBUG_NAMED_VALUE(PIPE_BIND_CURSOR),
+      DEBUG_NAMED_VALUE(PIPE_BIND_CUSTOM),
+      DEBUG_NAMED_VALUE(PIPE_BIND_GLOBAL),
+      DEBUG_NAMED_VALUE(PIPE_BIND_SHADER_RESOURCE),
+      DEBUG_NAMED_VALUE(PIPE_BIND_COMPUTE_RESOURCE),
+      DEBUG_NAMED_VALUE(PIPE_BIND_COMMAND_ARGS_BUFFER),
+      DEBUG_NAMED_VALUE(PIPE_BIND_SCANOUT),
+      DEBUG_NAMED_VALUE(PIPE_BIND_SHARED),
+      DEBUG_NAMED_VALUE(PIPE_BIND_LINEAR),
+      DEBUG_NAMED_VALUE_END
+   };
+
+   debug_printf("%s: %s\n", msg, debug_dump_flags(names, usage));
 }
 
 
+/**
+ * Print PIPE_USAGE_x enum values with a message.
+ */
+void
+debug_print_usage_enum(const char *msg, unsigned usage)
+{
+   static const struct debug_named_value names[] = {
+      DEBUG_NAMED_VALUE(PIPE_USAGE_DEFAULT),
+      DEBUG_NAMED_VALUE(PIPE_USAGE_IMMUTABLE),
+      DEBUG_NAMED_VALUE(PIPE_USAGE_DYNAMIC),
+      DEBUG_NAMED_VALUE(PIPE_USAGE_STREAM),
+      DEBUG_NAMED_VALUE(PIPE_USAGE_STAGING),
+      DEBUG_NAMED_VALUE_END
+   };
+
+   debug_printf("%s: %s\n", msg, debug_dump_enum(names, usage));
+}
+
 
 #endif
diff --git a/mesalib/src/gallium/auxiliary/util/u_debug.h b/mesalib/src/gallium/auxiliary/util/u_debug.h
index 4c22fdfb6..3b2255244 100644
--- a/mesalib/src/gallium/auxiliary/util/u_debug.h
+++ b/mesalib/src/gallium/auxiliary/util/u_debug.h
@@ -467,6 +467,12 @@ void debug_dump_float_rgba_bmp(const char *filename,
 void
 debug_print_transfer_flags(const char *msg, unsigned usage);
 
+void
+debug_print_bind_flags(const char *msg, unsigned usage);
+
+void
+debug_print_usage_enum(const char *msg, unsigned usage);
+
 
 #ifdef	__cplusplus
 }
diff --git a/mesalib/src/gallium/auxiliary/util/u_format_other.c b/mesalib/src/gallium/auxiliary/util/u_format_other.c
index 85001c14f..20752d469 100644
--- a/mesalib/src/gallium/auxiliary/util/u_format_other.c
+++ b/mesalib/src/gallium/auxiliary/util/u_format_other.c
@@ -42,10 +42,7 @@ util_format_r9g9b9e5_float_unpack_rgba_float(float *dst_row, unsigned dst_stride
       float *dst = dst_row;
       const uint8_t *src = src_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          rgb9e5_to_float3(value, dst);
          dst[3] = 1; /* a */
          src += 4;
@@ -66,10 +63,7 @@ util_format_r9g9b9e5_float_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride
       const float *src = src_row;
       uint8_t *dst = dst_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = float3_to_rgb9e5(src);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(float3_to_rgb9e5(src));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -83,10 +77,7 @@ void
 util_format_r9g9b9e5_float_fetch_rgba_float(float *dst, const uint8_t *src,
                                        unsigned i, unsigned j)
 {
-   uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-   value = util_bswap32(value);
-#endif
+   uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
    rgb9e5_to_float3(value, dst);
    dst[3] = 1; /* a */
 }
@@ -103,10 +94,7 @@ util_format_r9g9b9e5_float_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_str
       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;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          rgb9e5_to_float3(value, p);
          dst[0] = float_to_ubyte(p[0]); /* r */
          dst[1] = float_to_ubyte(p[1]); /* g */
@@ -136,10 +124,7 @@ util_format_r9g9b9e5_float_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_strid
          p[0] = ubyte_to_float(src[0]);
          p[1] = ubyte_to_float(src[1]);
          p[2] = ubyte_to_float(src[2]);
-         value = float3_to_rgb9e5(p);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(float3_to_rgb9e5(p));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -160,10 +145,7 @@ util_format_r11g11b10_float_unpack_rgba_float(float *dst_row, unsigned dst_strid
       float *dst = dst_row;
       const uint8_t *src = src_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          r11g11b10f_to_float3(value, dst);
          dst[3] = 1; /* a */
          src += 4;
@@ -184,10 +166,7 @@ util_format_r11g11b10_float_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
       const float *src = src_row;
       uint8_t *dst = dst_row;
       for(x = 0; x < width; x += 1) {
-         uint32_t value = float3_to_r11g11b10f(src);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(float3_to_r11g11b10f(src));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -201,10 +180,7 @@ void
 util_format_r11g11b10_float_fetch_rgba_float(float *dst, const uint8_t *src,
                                        unsigned i, unsigned j)
 {
-   uint32_t value = *(const uint32_t *)src;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-   value = util_bswap32(value);
-#endif
+   uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
    r11g11b10f_to_float3(value, dst);
    dst[3] = 1; /* a */
 }
@@ -221,10 +197,7 @@ util_format_r11g11b10_float_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       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;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*(const uint32_t *)src);
          r11g11b10f_to_float3(value, p);
          dst[0] = float_to_ubyte(p[0]); /* r */
          dst[1] = float_to_ubyte(p[1]); /* g */
@@ -254,10 +227,7 @@ util_format_r11g11b10_float_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          p[0] = ubyte_to_float(src[0]);
          p[1] = ubyte_to_float(src[1]);
          p[2] = ubyte_to_float(src[2]);
-         value = float3_to_r11g11b10f(p);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(float3_to_r11g11b10f(p));
          *(uint32_t *)dst = value;
          src += 4;
          dst += 4;
@@ -338,13 +308,9 @@ util_format_r8g8bx_snorm_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; x += 1) {
-         uint16_t value = *src++;
+         uint16_t value = util_cpu_to_le16(*src++);
          int16_t r, g;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
          r = ((int16_t)(value << 8)) >> 8;
          g = ((int16_t)(value << 0)) >> 8;
 
@@ -370,13 +336,9 @@ util_format_r8g8bx_snorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_strid
       uint8_t *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; x += 1) {
-         uint16_t value = *src++;
+         uint16_t value = util_cpu_to_le16(*src++);
          int16_t r, g;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
          r = ((int16_t)(value << 8)) >> 8;
          g = ((int16_t)(value << 0)) >> 8;
 
@@ -407,11 +369,7 @@ util_format_r8g8bx_snorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= (uint16_t)(((int8_t)(CLAMP(src[0], -1, 1) * 0x7f)) & 0xff) ;
          value |= (uint16_t)((((int8_t)(CLAMP(src[1], -1, 1) * 0x7f)) & 0xff) << 8) ;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
 
          src += 4;
       }
@@ -437,11 +395,7 @@ util_format_r8g8bx_snorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= src[0] >> 1;
          value |= (src[1] >> 1) << 8;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
 
          src += 4;
       }
@@ -455,13 +409,9 @@ void
 util_format_r8g8bx_snorm_fetch_rgba_float(float *dst, const uint8_t *src,
                                      unsigned i, unsigned j)
 {
-   uint16_t value = *(const uint16_t *)src;
+   uint16_t value = util_cpu_to_le16(*(const uint16_t *)src);
    int16_t r, g;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-   value = util_bswap16(value);
-#endif
-
    r = ((int16_t)(value << 8)) >> 8;
    g = ((int16_t)(value << 0)) >> 8;
 
diff --git a/mesalib/src/gallium/auxiliary/util/u_format_yuv.c b/mesalib/src/gallium/auxiliary/util/u_format_yuv.c
index 891d99cf6..16ed8af7b 100644
--- a/mesalib/src/gallium/auxiliary/util/u_format_yuv.c
+++ b/mesalib/src/gallium/auxiliary/util/u_format_yuv.c
@@ -52,11 +52,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       float r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          r  = ubyte_to_float((value >>  0) & 0xff);
          g0 = ubyte_to_float((value >>  8) & 0xff);
@@ -77,11 +73,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          r  = ubyte_to_float((value >>  0) & 0xff);
          g0 = ubyte_to_float((value >>  8) & 0xff);
@@ -114,11 +106,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       uint8_t r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          r  = (value >>  0) & 0xff;
          g0 = (value >>  8) & 0xff;
@@ -139,11 +127,7 @@ util_format_r8g8_b8g8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          r  = (value >>  0) & 0xff;
          g0 = (value >>  8) & 0xff;
@@ -186,11 +170,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(b)  << 16;
          value |= float_to_ubyte(g1) << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -206,11 +186,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(b)  << 16;
          value |= float_to_ubyte(g1) << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -243,11 +219,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= b  << 16;
          value |= g1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -263,11 +235,7 @@ util_format_r8g8_b8g8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= b  << 16;
          value |= g1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -304,11 +272,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       float r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          g0 = ubyte_to_float((value >>  0) & 0xff);
          r  = ubyte_to_float((value >>  8) & 0xff);
@@ -329,11 +293,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_float(float *dst_row, unsigned dst_strid
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          g0 = ubyte_to_float((value >>  0) & 0xff);
          r  = ubyte_to_float((value >>  8) & 0xff);
@@ -366,11 +326,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       uint8_t r, g0, g1, b;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          g0 = (value >>  0) & 0xff;
          r  = (value >>  8) & 0xff;
@@ -391,11 +347,7 @@ util_format_g8r8_g8b8_unorm_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_st
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          g0 = (value >>  0) & 0xff;
          r  = (value >>  8) & 0xff;
@@ -438,11 +390,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(g1) << 16;
          value |= float_to_ubyte(b)  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -458,11 +406,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_float(uint8_t *dst_row, unsigned dst_strid
          value |= float_to_ubyte(g1) << 16;
          value |= float_to_ubyte(b)  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -495,11 +439,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= g1 << 16;
          value |= b  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -515,11 +455,7 @@ util_format_g8r8_g8b8_unorm_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stri
          value |= g1 << 16;
          value |= b  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -556,11 +492,7 @@ util_format_uyvy_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -577,11 +509,7 @@ util_format_uyvy_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -612,11 +540,7 @@ util_format_uyvy_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -633,11 +557,7 @@ util_format_uyvy_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          u  = (value >>  0) & 0xff;
          y0 = (value >>  8) & 0xff;
@@ -683,11 +603,7 @@ util_format_uyvy_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -702,11 +618,7 @@ util_format_uyvy_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -744,11 +656,7 @@ util_format_uyvy_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -763,11 +671,7 @@ util_format_uyvy_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= v  << 16;
          value |= y1 << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -809,11 +713,7 @@ util_format_yuyv_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -830,11 +730,7 @@ util_format_yuyv_unpack_rgba_float(float *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -865,11 +761,7 @@ util_format_yuyv_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       uint8_t y0, y1, u, v;
 
       for (x = 0; x + 1 < width; x += 2) {
-         value = *src++;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src++);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -886,11 +778,7 @@ util_format_yuyv_unpack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
       }
 
       if (x < width) {
-         value = *src;
-
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         value = util_cpu_to_le32(*src);
 
          y0 = (value >>  0) & 0xff;
          u  = (value >>  8) & 0xff;
@@ -936,11 +824,7 @@ util_format_yuyv_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -955,11 +839,7 @@ util_format_yuyv_pack_rgba_float(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
@@ -997,11 +877,7 @@ util_format_yuyv_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
 
          src += 8;
       }
@@ -1016,11 +892,7 @@ util_format_yuyv_pack_rgba_8unorm(uint8_t *dst_row, unsigned dst_stride,
          value |= y1 << 16;
          value |= v  << 24;
 
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-
-         *dst = value;
+         *dst = util_le32_to_cpu(value);
       }
 
       dst_row += dst_stride/sizeof(*dst_row);
diff --git a/mesalib/src/gallium/auxiliary/util/u_format_zs.c b/mesalib/src/gallium/auxiliary/util/u_format_zs.c
index ed45c52f9..f1ed32f1d 100644
--- a/mesalib/src/gallium/auxiliary/util/u_format_zs.c
+++ b/mesalib/src/gallium/auxiliary/util/u_format_zs.c
@@ -147,10 +147,7 @@ util_format_z16_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint16_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
+         uint16_t value = util_cpu_to_le16(*src++);
          *dst++ = z16_unorm_to_z32_float(value);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -170,10 +167,7 @@ util_format_z16_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint16_t value;
          value = z32_float_to_z16_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -190,10 +184,7 @@ util_format_z16_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
       uint32_t *dst = dst_row;
       const uint16_t *src = (const uint16_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint16_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
+         uint16_t value = util_cpu_to_le16(*src++);
          *dst++ = z16_unorm_to_z32_unorm(value);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -213,10 +204,7 @@ util_format_z16_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint16_t value;
          value = z32_unorm_to_z16_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap16(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le16_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -233,10 +221,7 @@ util_format_z32_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z32_unorm_to_z32_float(value);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -256,10 +241,7 @@ util_format_z32_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_float_to_z32_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -362,10 +344,7 @@ util_format_z24_unorm_s8_uint_unpack_z_float(float *dst_row, unsigned dst_stride
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value =  util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -383,16 +362,10 @@ util_format_z24_unorm_s8_uint_pack_z_float(uint8_t *dst_row, unsigned dst_stride
       const float *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0xff000000;
          value |= z32_float_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -409,10 +382,7 @@ util_format_z24_unorm_s8_uint_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_s
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -430,16 +400,10 @@ util_format_z24_unorm_s8_uint_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stri
       const uint32_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value= *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0xff000000;
          value |= z32_unorm_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -456,10 +420,7 @@ util_format_z24_unorm_s8_uint_unpack_s_8uint(uint8_t *dst_row, unsigned dst_stri
       uint8_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = value >> 24;
       }
       src_row += src_stride/sizeof(*src_row);
@@ -477,16 +438,10 @@ util_format_z24_unorm_s8_uint_pack_s_8uint(uint8_t *dst_row, unsigned dst_stride
       const uint8_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0x00ffffff;
          value |= *src++ << 24;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -503,10 +458,7 @@ util_format_s8_uint_z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -524,16 +476,10 @@ util_format_s8_uint_z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride
       const float *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0x000000ff;
          value |= z32_float_to_z24_unorm(*src++) << 8;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -550,10 +496,7 @@ util_format_s8_uint_z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_s
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -571,16 +514,10 @@ util_format_s8_uint_z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stri
       const uint32_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0x000000ff;
          value |= *src++ & 0xffffff00;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -597,10 +534,7 @@ util_format_s8_uint_z24_unorm_unpack_s_8uint(uint8_t *dst_row, unsigned dst_stri
       uint8_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = value & 0xff;
       }
       src_row += src_stride/sizeof(*src_row);
@@ -618,16 +552,10 @@ util_format_s8_uint_z24_unorm_pack_s_8uint(uint8_t *dst_row, unsigned dst_stride
       const uint8_t *src = src_row;
       uint32_t *dst = (uint32_t *)dst_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *dst;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_le32_to_cpu(*dst);
          value &= 0xffffff00;
          value |= *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -644,10 +572,7 @@ util_format_z24x8_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -667,10 +592,7 @@ util_format_z24x8_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_float_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_le32_to_cpu(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -687,10 +609,7 @@ util_format_z24x8_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value & 0xffffff);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -710,10 +629,7 @@ util_format_z24x8_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_unorm_to_z24_unorm(*src++);
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -730,10 +646,7 @@ util_format_x8z24_unorm_unpack_z_float(float *dst_row, unsigned dst_stride,
       float *dst = dst_row;
       const uint32_t *src = (uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_float(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -753,10 +666,7 @@ util_format_x8z24_unorm_pack_z_float(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_float_to_z24_unorm(*src++) << 8;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
@@ -773,10 +683,7 @@ util_format_x8z24_unorm_unpack_z_32unorm(uint32_t *dst_row, unsigned dst_stride,
       uint32_t *dst = dst_row;
       const uint32_t *src = (const uint32_t *)src_row;
       for(x = 0; x < width; ++x) {
-         uint32_t value = *src++;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
+         uint32_t value = util_cpu_to_le32(*src++);
          *dst++ = z24_unorm_to_z32_unorm(value >> 8);
       }
       src_row += src_stride/sizeof(*src_row);
@@ -796,10 +703,7 @@ util_format_x8z24_unorm_pack_z_32unorm(uint8_t *dst_row, unsigned dst_stride,
       for(x = 0; x < width; ++x) {
          uint32_t value;
          value = z32_unorm_to_z24_unorm(*src++) << 8;
-#ifdef PIPE_ARCH_BIG_ENDIAN
-         value = util_bswap32(value);
-#endif
-         *dst++ = value;
+         *dst++ = util_cpu_to_le32(value);
       }
       dst_row += dst_stride/sizeof(*dst_row);
       src_row += src_stride/sizeof(*src_row);
diff --git a/mesalib/src/gallium/auxiliary/util/u_math.h b/mesalib/src/gallium/auxiliary/util/u_math.h
index d6e83f962..8f62cac66 100644
--- a/mesalib/src/gallium/auxiliary/util/u_math.h
+++ b/mesalib/src/gallium/auxiliary/util/u_math.h
@@ -41,13 +41,7 @@
 
 #include "pipe/p_compiler.h"
 
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-
-#include <math.h>
+#include "c99_math.h"
 #include <float.h>
 #include <stdarg.h>
 
@@ -56,153 +50,15 @@ extern "C" {
 #endif
 
 
-#ifndef M_SQRT2
-#define M_SQRT2 1.41421356237309504880
-#endif
-
-
-#if defined(_MSC_VER)
-
-#if _MSC_VER < 1400 && !defined(__cplusplus)
-
-static INLINE float cosf( float f )
-{
-   return (float) cos( (double) f );
-}
-
-static INLINE float sinf( float f )
-{
-   return (float) sin( (double) f );
-}
-
-static INLINE float ceilf( float f )
-{
-   return (float) ceil( (double) f );
-}
-
-static INLINE float floorf( float f )
-{
-   return (float) floor( (double) f );
-}
-
-static INLINE float powf( float f, float g )
-{
-   return (float) pow( (double) f, (double) g );
-}
-
-static INLINE float sqrtf( float f )
-{
-   return (float) sqrt( (double) f );
-}
-
-static INLINE float fabsf( float f )
-{
-   return (float) fabs( (double) f );
-}
-
-static INLINE float logf( float f )
-{
-   return (float) log( (double) f );
-}
-
-#else
-/* Work-around an extra semi-colon in VS 2005 logf definition */
-#ifdef logf
-#undef logf
-#define logf(x) ((float)log((double)(x)))
-#endif /* logf */
-
-#if _MSC_VER < 1800
-#define isfinite(x) _finite((double)(x))
-#define isnan(x) _isnan((double)(x))
-#endif /* _MSC_VER < 1800 */
-#endif /* _MSC_VER < 1400 && !defined(__cplusplus) */
-
-#if _MSC_VER < 1800
-static INLINE double log2( double x )
-{
-   const double invln2 = 1.442695041;
-   return log( x ) * invln2;
-}
-
-static INLINE double
-round(double x)
-{
-   return x >= 0.0 ? floor(x + 0.5) : ceil(x - 0.5);
-}
-
-static INLINE float
-roundf(float x)
-{
-   return x >= 0.0f ? floorf(x + 0.5f) : ceilf(x - 0.5f);
-}
+#ifdef __cplusplus
+extern "C" {
 #endif
 
-#ifndef INFINITY
-#define INFINITY (DBL_MAX + DBL_MAX)
-#endif
 
-#ifndef NAN
-#define NAN (INFINITY - INFINITY)
+#ifndef M_SQRT2
+#define M_SQRT2 1.41421356237309504880
 #endif
 
-#endif /* _MSC_VER */
-
-
-#if __STDC_VERSION__ < 199901L && (!defined(__cplusplus) || defined(_MSC_VER))
-static INLINE long int
-lrint(double d)
-{
-   long int rounded = (long int)(d + 0.5);
-
-   if (d - floor(d) == 0.5) {
-      if (rounded % 2 != 0)
-         rounded += (d > 0) ? -1 : 1;
-   }
-
-   return rounded;
-}
-
-static INLINE long int
-lrintf(float f)
-{
-   long int rounded = (long int)(f + 0.5f);
-
-   if (f - floorf(f) == 0.5f) {
-      if (rounded % 2 != 0)
-         rounded += (f > 0) ? -1 : 1;
-   }
-
-   return rounded;
-}
-
-static INLINE long long int
-llrint(double d)
-{
-   long long int rounded = (long long int)(d + 0.5);
-
-   if (d - floor(d) == 0.5) {
-      if (rounded % 2 != 0)
-         rounded += (d > 0) ? -1 : 1;
-   }
-
-   return rounded;
-}
-
-static INLINE long long int
-llrintf(float f)
-{
-   long long int rounded = (long long int)(f + 0.5f);
-
-   if (f - floorf(f) == 0.5f) {
-      if (rounded % 2 != 0)
-         rounded += (f > 0) ? -1 : 1;
-   }
-
-   return rounded;
-}
-#endif /* C99 */
-
 #define POW2_TABLE_SIZE_LOG2 9
 #define POW2_TABLE_SIZE (1 << POW2_TABLE_SIZE_LOG2)
 #define POW2_TABLE_OFFSET (POW2_TABLE_SIZE/2)
@@ -499,7 +355,7 @@ util_half_inf_sign(int16_t x)
 #ifndef FFS_DEFINED
 #define FFS_DEFINED 1
 
-#if defined(_MSC_VER) && _MSC_VER >= 1300 && (_M_IX86 || _M_AMD64 || _M_IA64)
+#if defined(_MSC_VER) && (_M_IX86 || _M_AMD64 || _M_IA64)
 unsigned char _BitScanForward(unsigned long* Index, unsigned long Mask);
 #pragma intrinsic(_BitScanForward)
 static INLINE
@@ -541,7 +397,7 @@ unsigned ffs( unsigned u )
 static INLINE unsigned
 util_last_bit(unsigned u)
 {
-#if defined(__GNUC__)
+#if defined(HAVE___BUILTIN_CLZ)
    return u == 0 ? 0 : 32 - __builtin_clz(u);
 #else
    unsigned r = 0;
@@ -663,7 +519,7 @@ float_to_byte_tex(float f)
 static INLINE unsigned
 util_logbase2(unsigned n)
 {
-#if defined(PIPE_CC_GCC)
+#if defined(HAVE___BUILTIN_CLZ)
    return ((sizeof(unsigned) * 8 - 1) - __builtin_clz(n | 1));
 #else
    unsigned pos = 0;
@@ -683,7 +539,7 @@ util_logbase2(unsigned n)
 static INLINE unsigned
 util_next_power_of_two(unsigned x)
 {
-#if defined(PIPE_CC_GCC)
+#if defined(HAVE___BUILTIN_CLZ)
    if (x <= 1)
        return 1;
 
@@ -715,7 +571,7 @@ util_next_power_of_two(unsigned x)
 static INLINE unsigned
 util_bitcount(unsigned n)
 {
-#if defined(PIPE_CC_GCC)
+#if defined(HAVE___BUILTIN_POPCOUNT)
    return __builtin_popcount(n);
 #else
    /* K&R classic bitcount.
@@ -724,8 +580,8 @@ util_bitcount(unsigned n)
     * Requires only one iteration per set bit, instead of
     * one iteration per bit less than highest set bit.
     */
-   unsigned bits = 0;
-   for (bits; n; bits++) {
+   unsigned bits;
+   for (bits = 0; n; bits++) {
       n &= n - 1;
    }
    return bits;
@@ -784,8 +640,7 @@ util_bitreverse(unsigned n)
 static INLINE uint32_t
 util_bswap32(uint32_t n)
 {
-/* We need the gcc version checks for non-autoconf build system */
-#if defined(HAVE___BUILTIN_BSWAP32) || (defined(PIPE_CC_GCC) && (PIPE_CC_GCC_VERSION >= 403))
+#if defined(HAVE___BUILTIN_BSWAP32)
    return __builtin_bswap32(n);
 #else
    return (n >> 24) |
diff --git a/mesalib/src/gallium/auxiliary/util/u_sse.h b/mesalib/src/gallium/auxiliary/util/u_sse.h
index 642f96eb7..d4f51912a 100644
--- a/mesalib/src/gallium/auxiliary/util/u_sse.h
+++ b/mesalib/src/gallium/auxiliary/util/u_sse.h
@@ -44,33 +44,6 @@
 #include <emmintrin.h>
 
 
-/* MSVC before VC8 does not support the _mm_castxxx_yyy */
-#if defined(_MSC_VER) && _MSC_VER < 1500
-
-union __declspec(align(16)) m128_types {
-   __m128 m128;
-   __m128i m128i;
-   __m128d m128d;
-};
-
-static __inline __m128
-_mm_castsi128_ps(__m128i a)
-{
-   union m128_types u;
-   u.m128i = a;
-   return u.m128;
-}
-
-static __inline __m128i
-_mm_castps_si128(__m128 a)
-{
-   union m128_types u;
-   u.m128 = a;
-   return u.m128i;
-}
-
-#endif /* defined(_MSC_VER) && _MSC_VER < 1500 */
-
 union m128i {
    __m128i m;
    ubyte ub[16];
diff --git a/mesalib/src/glsl/Makefile.am b/mesalib/src/glsl/Makefile.am
index 5a0a643da..b466a3b5c 100644
--- a/mesalib/src/glsl/Makefile.am
+++ b/mesalib/src/glsl/Makefile.am
@@ -33,8 +33,12 @@ AM_CPPFLAGS = \
 	-I$(top_srcdir)/src/gtest/include \
 	-I$(top_builddir)/src/glsl/nir \
 	$(DEFINES)
-AM_CFLAGS = $(VISIBILITY_CFLAGS)
-AM_CXXFLAGS = $(VISIBILITY_CXXFLAGS)
+AM_CFLAGS = \
+	$(VISIBILITY_CFLAGS) \
+	$(MSVC2013_COMPAT_CFLAGS)
+AM_CXXFLAGS = \
+	$(VISIBILITY_CXXFLAGS) \
+	$(MSVC2013_COMPAT_CXXFLAGS)
 
 EXTRA_DIST = tests glcpp/tests README TODO glcpp/README	\
 	glsl_lexer.ll					\
diff --git a/mesalib/src/glsl/Makefile.sources b/mesalib/src/glsl/Makefile.sources
index d0210d170..b876642e8 100644
--- a/mesalib/src/glsl/Makefile.sources
+++ b/mesalib/src/glsl/Makefile.sources
@@ -157,6 +157,7 @@ LIBGLSL_FILES = \
 	lower_ubo_reference.cpp \
 	opt_algebraic.cpp \
 	opt_array_splitting.cpp \
+	opt_conditional_discard.cpp \
 	opt_constant_folding.cpp \
 	opt_constant_propagation.cpp \
 	opt_constant_variable.cpp \
diff --git a/mesalib/src/glsl/SConscript b/mesalib/src/glsl/SConscript
index 21c8266a6..26de455f0 100644
--- a/mesalib/src/glsl/SConscript
+++ b/mesalib/src/glsl/SConscript
@@ -6,6 +6,8 @@ from sys import executable as python_cmd
 
 env = env.Clone()
 
+env.MSVC2013Compat()
+
 env.Prepend(CPPPATH = [
     '#include',
     '#src',
diff --git a/mesalib/src/glsl/ast_array_index.cpp b/mesalib/src/glsl/ast_array_index.cpp
index ff0c7576d..ecef651f7 100644
--- a/mesalib/src/glsl/ast_array_index.cpp
+++ b/mesalib/src/glsl/ast_array_index.cpp
@@ -83,11 +83,9 @@ update_max_array_access(ir_rvalue *ir, int idx, YYLTYPE *loc,
 
       if (deref_var != NULL) {
          if (deref_var->var->is_interface_instance()) {
-            const glsl_type *interface_type =
-               deref_var->var->get_interface_type();
             unsigned field_index =
                deref_record->record->type->field_index(deref_record->field);
-            assert(field_index < interface_type->length);
+            assert(field_index < deref_var->var->get_interface_type()->length);
 
             unsigned *const max_ifc_array_access =
                deref_var->var->get_max_ifc_array_access();
diff --git a/mesalib/src/glsl/builtin_functions.cpp b/mesalib/src/glsl/builtin_functions.cpp
index b643927d0..84bbdc2a4 100644
--- a/mesalib/src/glsl/builtin_functions.cpp
+++ b/mesalib/src/glsl/builtin_functions.cpp
@@ -2549,7 +2549,7 @@ builtin_builder::add_image_function(const char *name,
    };
    ir_function *f = new(mem_ctx) ir_function(name);
 
-   for (unsigned i = 0; i < Elements(types); ++i) {
+   for (unsigned i = 0; i < ARRAY_SIZE(types); ++i) {
       if (types[i]->sampler_type != GLSL_TYPE_FLOAT ||
           (flags & IMAGE_FUNCTION_SUPPORTS_FLOAT_DATA_TYPE))
          f->add_signature(_image(types[i], intrinsic_name,
diff --git a/mesalib/src/glsl/builtin_variables.cpp b/mesalib/src/glsl/builtin_variables.cpp
index 65e32ad73..21e7331c7 100644
--- a/mesalib/src/glsl/builtin_variables.cpp
+++ b/mesalib/src/glsl/builtin_variables.cpp
@@ -227,7 +227,7 @@ static const struct gl_builtin_uniform_element gl_NormalMatrix_elements[] = {
 
 #undef MATRIX
 
-#define STATEVAR(name) {#name, name ## _elements, Elements(name ## _elements)}
+#define STATEVAR(name) {#name, name ## _elements, ARRAY_SIZE(name ## _elements)}
 
 static const struct gl_builtin_uniform_desc _mesa_builtin_uniform_desc[] = {
    STATEVAR(gl_NumSamples),
diff --git a/mesalib/src/glsl/glcpp/pp.c b/mesalib/src/glsl/glcpp/pp.c
index a54bcbe16..160c6662f 100644
--- a/mesalib/src/glsl/glcpp/pp.c
+++ b/mesalib/src/glsl/glcpp/pp.c
@@ -25,7 +25,6 @@
 #include <string.h>
 #include <ctype.h>
 #include "glcpp.h"
-#include "main/core.h" /* for isblank() on MSVC */
 
 void
 glcpp_error (YYLTYPE *locp, glcpp_parser_t *parser, const char *fmt, ...)
diff --git a/mesalib/src/glsl/glsl_parser.yy b/mesalib/src/glsl/glsl_parser.yy
index ea3bd8a24..90c216ed9 100644
--- a/mesalib/src/glsl/glsl_parser.yy
+++ b/mesalib/src/glsl/glsl_parser.yy
@@ -1280,7 +1280,7 @@ layout_qualifier_id:
                  { "triangles_adjacency", GL_TRIANGLES_ADJACENCY },
                  { "triangle_strip", GL_TRIANGLE_STRIP },
          };
-         for (unsigned i = 0; i < Elements(map); i++) {
+         for (unsigned i = 0; i < ARRAY_SIZE(map); i++) {
             if (match_layout_qualifier($1, map[i].s, state) == 0) {
                $$.flags.q.prim_type = 1;
                $$.prim_type = map[i].e;
@@ -1344,7 +1344,7 @@ layout_qualifier_id:
                { "r8_snorm", GL_R8_SNORM, GLSL_TYPE_FLOAT }
             };
 
-            for (unsigned i = 0; i < Elements(map); i++) {
+            for (unsigned i = 0; i < ARRAY_SIZE(map); i++) {
                if (match_layout_qualifier($1, map[i].name, state) == 0) {
                   $$.flags.q.explicit_image_format = 1;
                   $$.image_format = map[i].format;
diff --git a/mesalib/src/glsl/glsl_parser_extras.cpp b/mesalib/src/glsl/glsl_parser_extras.cpp
index 9f7931380..79624bc26 100644
--- a/mesalib/src/glsl/glsl_parser_extras.cpp
+++ b/mesalib/src/glsl/glsl_parser_extras.cpp
@@ -119,9 +119,9 @@ _mesa_glsl_parse_state::_mesa_glsl_parse_state(struct gl_context *_ctx,
    this->Const.MaxAtomicBufferBindings = ctx->Const.MaxAtomicBufferBindings;
 
    /* Compute shader constants */
-   for (unsigned i = 0; i < Elements(this->Const.MaxComputeWorkGroupCount); i++)
+   for (unsigned i = 0; i < ARRAY_SIZE(this->Const.MaxComputeWorkGroupCount); i++)
       this->Const.MaxComputeWorkGroupCount[i] = ctx->Const.MaxComputeWorkGroupCount[i];
-   for (unsigned i = 0; i < Elements(this->Const.MaxComputeWorkGroupSize); i++)
+   for (unsigned i = 0; i < ARRAY_SIZE(this->Const.MaxComputeWorkGroupSize); i++)
       this->Const.MaxComputeWorkGroupSize[i] = ctx->Const.MaxComputeWorkGroupSize[i];
 
    this->Const.MaxImageUnits = ctx->Const.MaxImageUnits;
@@ -636,7 +636,7 @@ void _mesa_glsl_extension::set_flags(_mesa_glsl_parse_state *state,
  */
 static const _mesa_glsl_extension *find_extension(const char *name)
 {
-   for (unsigned i = 0; i < Elements(_mesa_glsl_supported_extensions); ++i) {
+   for (unsigned i = 0; i < ARRAY_SIZE(_mesa_glsl_supported_extensions); ++i) {
       if (strcmp(name, _mesa_glsl_supported_extensions[i].name) == 0) {
          return &_mesa_glsl_supported_extensions[i];
       }
@@ -674,7 +674,7 @@ _mesa_glsl_process_extension(const char *name, YYLTYPE *name_locp,
 	 return false;
       } else {
          for (unsigned i = 0;
-              i < Elements(_mesa_glsl_supported_extensions); ++i) {
+              i < ARRAY_SIZE(_mesa_glsl_supported_extensions); ++i) {
             const _mesa_glsl_extension *extension
                = &_mesa_glsl_supported_extensions[i];
             if (extension->compatible_with_state(state)) {
@@ -1627,6 +1627,7 @@ do_common_optimization(exec_list *ir, bool linked,
    }
    progress = do_if_simplification(ir) || progress;
    progress = opt_flatten_nested_if_blocks(ir) || progress;
+   progress = opt_conditional_discard(ir) || progress;
    progress = do_copy_propagation(ir) || progress;
    progress = do_copy_propagation_elements(ir) || progress;
 
diff --git a/mesalib/src/glsl/ir.cpp b/mesalib/src/glsl/ir.cpp
index f4f92e9df..54656f899 100644
--- a/mesalib/src/glsl/ir.cpp
+++ b/mesalib/src/glsl/ir.cpp
@@ -618,8 +618,8 @@ static const char *const operator_strs[] = {
 
 const char *ir_expression::operator_string(ir_expression_operation op)
 {
-   assert((unsigned int) op < Elements(operator_strs));
-   assert(Elements(operator_strs) == (ir_quadop_vector + 1));
+   assert((unsigned int) op < ARRAY_SIZE(operator_strs));
+   assert(ARRAY_SIZE(operator_strs) == (ir_quadop_vector + 1));
    return operator_strs[op];
 }
 
@@ -1707,7 +1707,7 @@ const char *const ir_variable::warn_extension_table[] = {
 void
 ir_variable::enable_extension_warning(const char *extension)
 {
-   for (unsigned i = 0; i < Elements(warn_extension_table); i++) {
+   for (unsigned i = 0; i < ARRAY_SIZE(warn_extension_table); i++) {
       if (strcmp(warn_extension_table[i], extension) == 0) {
          this->data.warn_extension_index = i;
          return;
diff --git a/mesalib/src/glsl/ir_clone.cpp b/mesalib/src/glsl/ir_clone.cpp
index 5c7279ca3..914e0e4d5 100644
--- a/mesalib/src/glsl/ir_clone.cpp
+++ b/mesalib/src/glsl/ir_clone.cpp
@@ -158,7 +158,7 @@ ir_call::clone(void *mem_ctx, struct hash_table *ht) const
 ir_expression *
 ir_expression::clone(void *mem_ctx, struct hash_table *ht) const
 {
-   ir_rvalue *op[Elements(this->operands)] = { NULL, };
+   ir_rvalue *op[ARRAY_SIZE(this->operands)] = { NULL, };
    unsigned int i;
 
    for (i = 0; i < get_num_operands(); i++) {
diff --git a/mesalib/src/glsl/ir_constant_expression.cpp b/mesalib/src/glsl/ir_constant_expression.cpp
index 07dd439d5..388c4c286 100644
--- a/mesalib/src/glsl/ir_constant_expression.cpp
+++ b/mesalib/src/glsl/ir_constant_expression.cpp
@@ -506,7 +506,7 @@ ir_expression::constant_expression_value(struct hash_table *variable_context)
    if (this->type->is_error())
       return NULL;
 
-   ir_constant *op[Elements(this->operands)] = { NULL, };
+   ir_constant *op[ARRAY_SIZE(this->operands)] = { NULL, };
    ir_constant_data data;
 
    memset(&data, 0, sizeof(data));
diff --git a/mesalib/src/glsl/ir_optimization.h b/mesalib/src/glsl/ir_optimization.h
index 7eb861ae5..e6939f3fe 100644
--- a/mesalib/src/glsl/ir_optimization.h
+++ b/mesalib/src/glsl/ir_optimization.h
@@ -77,6 +77,7 @@ bool do_common_optimization(exec_list *ir, bool linked,
 bool do_rebalance_tree(exec_list *instructions);
 bool do_algebraic(exec_list *instructions, bool native_integers,
                   const struct gl_shader_compiler_options *options);
+bool opt_conditional_discard(exec_list *instructions);
 bool do_constant_folding(exec_list *instructions);
 bool do_constant_variable(exec_list *instructions);
 bool do_constant_variable_unlinked(exec_list *instructions);
diff --git a/mesalib/src/glsl/ir_rvalue_visitor.cpp b/mesalib/src/glsl/ir_rvalue_visitor.cpp
index 34cdb1c98..2eee3da7b 100644
--- a/mesalib/src/glsl/ir_rvalue_visitor.cpp
+++ b/mesalib/src/glsl/ir_rvalue_visitor.cpp
@@ -135,6 +135,13 @@ ir_rvalue_base_visitor::rvalue_visit(ir_call *ir)
 }
 
 ir_visitor_status
+ir_rvalue_base_visitor::rvalue_visit(ir_discard *ir)
+{
+   handle_rvalue(&ir->condition);
+   return visit_continue;
+}
+
+ir_visitor_status
 ir_rvalue_base_visitor::rvalue_visit(ir_return *ir)
 {
    handle_rvalue(&ir->value);;
@@ -205,6 +212,12 @@ ir_rvalue_visitor::visit_leave(ir_call *ir)
 }
 
 ir_visitor_status
+ir_rvalue_visitor::visit_leave(ir_discard *ir)
+{
+   return rvalue_visit(ir);
+}
+
+ir_visitor_status
 ir_rvalue_visitor::visit_leave(ir_return *ir)
 {
    return rvalue_visit(ir);
@@ -271,6 +284,12 @@ ir_rvalue_enter_visitor::visit_enter(ir_call *ir)
 }
 
 ir_visitor_status
+ir_rvalue_enter_visitor::visit_enter(ir_discard *ir)
+{
+   return rvalue_visit(ir);
+}
+
+ir_visitor_status
 ir_rvalue_enter_visitor::visit_enter(ir_return *ir)
 {
    return rvalue_visit(ir);
diff --git a/mesalib/src/glsl/ir_rvalue_visitor.h b/mesalib/src/glsl/ir_rvalue_visitor.h
index 04ec0fa39..185c72a5b 100644
--- a/mesalib/src/glsl/ir_rvalue_visitor.h
+++ b/mesalib/src/glsl/ir_rvalue_visitor.h
@@ -36,6 +36,7 @@ public:
    ir_visitor_status rvalue_visit(ir_call *);
    ir_visitor_status rvalue_visit(ir_dereference_array *);
    ir_visitor_status rvalue_visit(ir_dereference_record *);
+   ir_visitor_status rvalue_visit(ir_discard *);
    ir_visitor_status rvalue_visit(ir_expression *);
    ir_visitor_status rvalue_visit(ir_if *);
    ir_visitor_status rvalue_visit(ir_return *);
@@ -54,6 +55,7 @@ public:
    virtual ir_visitor_status visit_leave(ir_call *);
    virtual ir_visitor_status visit_leave(ir_dereference_array *);
    virtual ir_visitor_status visit_leave(ir_dereference_record *);
+   virtual ir_visitor_status visit_leave(ir_discard *);
    virtual ir_visitor_status visit_leave(ir_expression *);
    virtual ir_visitor_status visit_leave(ir_if *);
    virtual ir_visitor_status visit_leave(ir_return *);
@@ -70,6 +72,7 @@ public:
    virtual ir_visitor_status visit_enter(ir_call *);
    virtual ir_visitor_status visit_enter(ir_dereference_array *);
    virtual ir_visitor_status visit_enter(ir_dereference_record *);
+   virtual ir_visitor_status visit_enter(ir_discard *);
    virtual ir_visitor_status visit_enter(ir_expression *);
    virtual ir_visitor_status visit_enter(ir_if *);
    virtual ir_visitor_status visit_enter(ir_return *);
diff --git a/mesalib/src/glsl/ir_validate.cpp b/mesalib/src/glsl/ir_validate.cpp
index 667889480..7a7688cb2 100644
--- a/mesalib/src/glsl/ir_validate.cpp
+++ b/mesalib/src/glsl/ir_validate.cpp
@@ -61,6 +61,7 @@ public:
    virtual ir_visitor_status visit(ir_variable *v);
    virtual ir_visitor_status visit(ir_dereference_variable *ir);
 
+   virtual ir_visitor_status visit_enter(ir_discard *ir);
    virtual ir_visitor_status visit_enter(ir_if *ir);
 
    virtual ir_visitor_status visit_enter(ir_function *ir);
@@ -133,6 +134,20 @@ ir_validate::visit_enter(class ir_dereference_array *ir)
 }
 
 ir_visitor_status
+ir_validate::visit_enter(ir_discard *ir)
+{
+   if (ir->condition && ir->condition->type != glsl_type::bool_type) {
+      printf("ir_discard condition %s type instead of bool.\n",
+	     ir->condition->type->name);
+      ir->print();
+      printf("\n");
+      abort();
+   }
+
+   return visit_continue;
+}
+
+ir_visitor_status
 ir_validate::visit_enter(ir_if *ir)
 {
    if (ir->condition->type != glsl_type::bool_type) {
diff --git a/mesalib/src/glsl/link_interface_blocks.cpp b/mesalib/src/glsl/link_interface_blocks.cpp
index 0ce502d4f..07f5b4223 100644
--- a/mesalib/src/glsl/link_interface_blocks.cpp
+++ b/mesalib/src/glsl/link_interface_blocks.cpp
@@ -50,18 +50,20 @@ struct interface_block_definition
     * represents either the interface instance (for named interfaces), or a
     * member of the interface (for unnamed interfaces).
     */
-   explicit interface_block_definition(const ir_variable *var)
-      : type(var->get_interface_type()),
-        instance_name(NULL),
-        array_size(-1)
+   explicit interface_block_definition(ir_variable *var)
+      : var(var),
+        type(var->get_interface_type()),
+        instance_name(NULL)
    {
       if (var->is_interface_instance()) {
          instance_name = var->name;
-         if (var->type->is_array())
-            array_size = var->type->length;
       }
       explicitly_declared = (var->data.how_declared != ir_var_declared_implicitly);
    }
+   /**
+    * Interface block ir_variable
+    */
+   ir_variable *var;
 
    /**
     * Interface block type
@@ -74,12 +76,6 @@ struct interface_block_definition
    const char *instance_name;
 
    /**
-    * For an interface block array, the array size (or 0 if unsized).
-    * Otherwise -1.
-    */
-   int array_size;
-
-   /**
     * True if this interface block was explicitly declared in the shader;
     * false if it was an implicitly declared built-in interface block.
     */
@@ -95,7 +91,8 @@ struct interface_block_definition
 bool
 intrastage_match(interface_block_definition *a,
                  const interface_block_definition *b,
-                 ir_variable_mode mode)
+                 ir_variable_mode mode,
+                 struct gl_shader_program *prog)
 {
    /* Types must match. */
    if (a->type != b->type) {
@@ -120,18 +117,13 @@ intrastage_match(interface_block_definition *a,
       return false;
    }
 
-   /* Array vs. nonarray must be consistent, and sizes must be
-    * consistent, with the exception that unsized arrays match sized
-    * arrays.
+   /* If a block is an array then it must match across the shader.
+    * Unsized arrays are also processed and matched agaist sized arrays.
     */
-   if ((a->array_size == -1) != (b->array_size == -1))
+   if (b->var->type != a->var->type &&
+       (b->instance_name != NULL || a->instance_name != NULL) &&
+       !validate_intrastage_arrays(prog, b->var, a->var))
       return false;
-   if (b->array_size != 0) {
-      if (a->array_size == 0)
-         a->array_size = b->array_size;
-      else if (a->array_size != b->array_size)
-         return false;
-   }
 
    return true;
 }
@@ -153,8 +145,8 @@ interstage_match(const interface_block_definition *producer,
    /* Unsized arrays should not occur during interstage linking.  They
     * should have all been assigned a size by link_intrastage_shaders.
     */
-   assert(consumer->array_size != 0);
-   assert(producer->array_size != 0);
+   assert(!consumer->var->type->is_unsized_array());
+   assert(!producer->var->type->is_unsized_array());
 
    /* Types must match. */
    if (consumer->type != producer->type) {
@@ -165,20 +157,27 @@ interstage_match(const interface_block_definition *producer,
       if (consumer->explicitly_declared || producer->explicitly_declared)
          return false;
    }
+
+   /* Ignore outermost array if geom shader */
+   const glsl_type *consumer_instance_type;
    if (extra_array_level) {
-      /* Consumer must be an array, and producer must not. */
-      if (consumer->array_size == -1)
-         return false;
-      if (producer->array_size != -1)
-         return false;
+      consumer_instance_type = consumer->var->type->fields.array;
    } else {
-      /* Array vs. nonarray must be consistent, and sizes must be consistent.
-       * Since unsized arrays have been ruled out, we can check this by just
-       * making sure the sizes are equal.
-       */
-      if (consumer->array_size != producer->array_size)
+      consumer_instance_type = consumer->var->type;
+   }
+
+   /* If a block is an array then it must match across shaders.
+    * Since unsized arrays have been ruled out, we can check this by just
+    * making sure the types are equal.
+    */
+   if ((consumer->instance_name != NULL &&
+        consumer_instance_type->is_array()) ||
+       (producer->instance_name != NULL &&
+        producer->var->type->is_array())) {
+      if (consumer_instance_type != producer->var->type)
          return false;
    }
+
    return true;
 }
 
@@ -298,7 +297,8 @@ validate_intrastage_interface_blocks(struct gl_shader_program *prog,
              */
             definitions->store(def);
          } else if (!intrastage_match(prev_def, &def,
-                                      (ir_variable_mode) var->data.mode)) {
+                                      (ir_variable_mode) var->data.mode,
+                                      prog)) {
             linker_error(prog, "definitions of interface block `%s' do not"
                          " match\n", iface_type->name);
             return;
@@ -374,7 +374,7 @@ validate_interstage_uniform_blocks(struct gl_shader_program *prog,
              * uniform matchin rules (for uniforms, it is as though all
              * shaders are in the same shader stage).
              */
-            if (!intrastage_match(old_def, &new_def, ir_var_uniform)) {
+            if (!intrastage_match(old_def, &new_def, ir_var_uniform, prog)) {
                linker_error(prog, "definitions of interface block `%s' do not "
                             "match\n", var->get_interface_type()->name);
                return;
diff --git a/mesalib/src/glsl/linker.cpp b/mesalib/src/glsl/linker.cpp
index 3f5eac1e2..0c4467779 100644
--- a/mesalib/src/glsl/linker.cpp
+++ b/mesalib/src/glsl/linker.cpp
@@ -64,6 +64,7 @@
  * \author Ian Romanick <ian.d.romanick@intel.com>
  */
 
+#include <ctype.h>
 #include "main/core.h"
 #include "glsl_symbol_table.h"
 #include "glsl_parser_extras.h"
@@ -680,6 +681,45 @@ validate_geometry_shader_emissions(struct gl_context *ctx,
    }
 }
 
+bool
+validate_intrastage_arrays(struct gl_shader_program *prog,
+                           ir_variable *const var,
+		           ir_variable *const existing)
+{
+   /* Consider the types to be "the same" if both types are arrays
+    * of the same type and one of the arrays is implicitly sized.
+    * In addition, set the type of the linked variable to the
+    * explicitly sized array.
+    */
+   if (var->type->is_array() && existing->type->is_array() &&
+       (var->type->fields.array == existing->type->fields.array) &&
+       ((var->type->length == 0)|| (existing->type->length == 0))) {
+      if (var->type->length != 0) {
+         if (var->type->length <= existing->data.max_array_access) {
+            linker_error(prog, "%s `%s' declared as type "
+                         "`%s' but outermost dimension has an index"
+                         " of `%i'\n",
+                         mode_string(var),
+                         var->name, var->type->name,
+                         existing->data.max_array_access);
+         }
+         existing->type = var->type;
+         return true;
+      } else if (existing->type->length != 0) {
+         if(existing->type->length <= var->data.max_array_access) {
+            linker_error(prog, "%s `%s' declared as type "
+                         "`%s' but outermost dimension has an index"
+                         " of `%i'\n",
+                         mode_string(var),
+                         var->name, existing->type->name,
+                         var->data.max_array_access);
+         }
+         return true;
+      }
+   }
+   return false;
+}
+
 
 /**
  * Perform validation of global variables used across multiple shaders
@@ -719,50 +759,23 @@ cross_validate_globals(struct gl_shader_program *prog,
 	  */
 	 ir_variable *const existing = variables.get_variable(var->name);
 	 if (existing != NULL) {
-	    if (var->type != existing->type) {
-	       /* Consider the types to be "the same" if both types are arrays
-		* of the same type and one of the arrays is implicitly sized.
-		* In addition, set the type of the linked variable to the
-		* explicitly sized array.
-		*/
-	       if (var->type->is_array()
-		   && existing->type->is_array()
-		   && (var->type->fields.array == existing->type->fields.array)
-		   && ((var->type->length == 0)
-		       || (existing->type->length == 0))) {
-		  if (var->type->length != 0) {
-                     if (var->type->length <= existing->data.max_array_access) {
-                        linker_error(prog, "%s `%s' declared as type "
-                                     "`%s' but outermost dimension has an index"
-                                     " of `%i'\n",
-                                     mode_string(var),
-                                     var->name, var->type->name,
-                                     existing->data.max_array_access);
-                        return;
-                     }
-		     existing->type = var->type;
-		  } else if (existing->type->length != 0
-                             && existing->type->length <=
-                                var->data.max_array_access) {
+            /* Check if types match. Interface blocks have some special
+             * rules so we handle those elsewhere.
+             */
+           if (var->type != existing->type &&
+                !var->is_interface_instance()) {
+	       if (!validate_intrastage_arrays(prog, var, existing)) {
+                  if (var->type->is_record() && existing->type->is_record()
+                      && existing->type->record_compare(var->type)) {
+                     existing->type = var->type;
+                  } else {
                      linker_error(prog, "%s `%s' declared as type "
-                                  "`%s' but outermost dimension has an index"
-                                  " of `%i'\n",
+                                  "`%s' and type `%s'\n",
                                   mode_string(var),
-                                  var->name, existing->type->name,
-                                  var->data.max_array_access);
+                                  var->name, var->type->name,
+                                  existing->type->name);
                      return;
                   }
-               } else if (var->type->is_record()
-		   && existing->type->is_record()
-		   && existing->type->record_compare(var->type)) {
-		  existing->type = var->type;
-	       } else {
-		  linker_error(prog, "%s `%s' declared as type "
-			       "`%s' and type `%s'\n",
-			       mode_string(var),
-			       var->name, var->type->name,
-			       existing->type->name);
-		  return;
 	       }
 	    }
 
diff --git a/mesalib/src/glsl/linker.h b/mesalib/src/glsl/linker.h
index be4da5e0a..ce3dc3232 100644
--- a/mesalib/src/glsl/linker.h
+++ b/mesalib/src/glsl/linker.h
@@ -61,6 +61,11 @@ link_uniform_blocks(void *mem_ctx,
                     unsigned num_shaders,
                     struct gl_uniform_block **blocks_ret);
 
+bool
+validate_intrastage_arrays(struct gl_shader_program *prog,
+                           ir_variable *const var,
+                           ir_variable *const existing);
+
 void
 validate_intrastage_interface_blocks(struct gl_shader_program *prog,
                                      const gl_shader **shader_list,
diff --git a/mesalib/src/glsl/loop_controls.cpp b/mesalib/src/glsl/loop_controls.cpp
index 2459fc1c3..d7f0b2809 100644
--- a/mesalib/src/glsl/loop_controls.cpp
+++ b/mesalib/src/glsl/loop_controls.cpp
@@ -123,7 +123,7 @@ calculate_iterations(ir_rvalue *from, ir_rvalue *to, ir_rvalue *increment,
    const int bias[] = { -1, 0, 1 };
    bool valid_loop = false;
 
-   for (unsigned i = 0; i < Elements(bias); i++) {
+   for (unsigned i = 0; i < ARRAY_SIZE(bias); i++) {
       /* Increment may be of type int, uint or float. */
       switch (increment->type->base_type) {
       case GLSL_TYPE_INT:
diff --git a/mesalib/src/glsl/lower_discard_flow.cpp b/mesalib/src/glsl/lower_discard_flow.cpp
index 1bc56d79e..ee45bf221 100644
--- a/mesalib/src/glsl/lower_discard_flow.cpp
+++ b/mesalib/src/glsl/lower_discard_flow.cpp
@@ -90,7 +90,14 @@ ir_visitor_status
 lower_discard_flow_visitor::visit_enter(ir_discard *ir)
 {
    ir_dereference *lhs = new(mem_ctx) ir_dereference_variable(discarded);
-   ir_rvalue *rhs = new(mem_ctx) ir_constant(true);
+   ir_rvalue *rhs;
+   if (ir->condition) {
+      /* discarded <- condition, use (var_ref discarded) as the condition */
+      rhs = ir->condition;
+      ir->condition = new(mem_ctx) ir_dereference_variable(discarded);
+   } else {
+      rhs = new(mem_ctx) ir_constant(true);
+   }
    ir_assignment *assign = new(mem_ctx) ir_assignment(lhs, rhs);
    ir->insert_before(assign);
 
diff --git a/mesalib/src/glsl/lower_instructions.cpp b/mesalib/src/glsl/lower_instructions.cpp
index 4779de059..845cfff36 100644
--- a/mesalib/src/glsl/lower_instructions.cpp
+++ b/mesalib/src/glsl/lower_instructions.cpp
@@ -124,7 +124,7 @@
  * Converts double trunc, ceil, floor, round to fract
  */
 
-#include "main/core.h" /* for M_LOG2E */
+#include "c99_math.h"
 #include "program/prog_instruction.h" /* for swizzle */
 #include "glsl_types.h"
 #include "ir.h"
diff --git a/mesalib/src/glsl/lower_packed_varyings.cpp b/mesalib/src/glsl/lower_packed_varyings.cpp
index 5e844c792..2c9a1c495 100644
--- a/mesalib/src/glsl/lower_packed_varyings.cpp
+++ b/mesalib/src/glsl/lower_packed_varyings.cpp
@@ -146,7 +146,11 @@
 
 #include "glsl_symbol_table.h"
 #include "ir.h"
+#include "ir_builder.h"
 #include "ir_optimization.h"
+#include "program/prog_instruction.h"
+
+using namespace ir_builder;
 
 namespace {
 
@@ -163,13 +167,14 @@ public:
    lower_packed_varyings_visitor(void *mem_ctx, unsigned locations_used,
                                  ir_variable_mode mode,
                                  unsigned gs_input_vertices,
-                                 exec_list *out_instructions);
+                                 exec_list *out_instructions,
+                                 exec_list *out_variables);
 
    void run(exec_list *instructions);
 
 private:
-   ir_assignment *bitwise_assign_pack(ir_rvalue *lhs, ir_rvalue *rhs);
-   ir_assignment *bitwise_assign_unpack(ir_rvalue *lhs, ir_rvalue *rhs);
+   void bitwise_assign_pack(ir_rvalue *lhs, ir_rvalue *rhs);
+   void bitwise_assign_unpack(ir_rvalue *lhs, ir_rvalue *rhs);
    unsigned lower_rvalue(ir_rvalue *rvalue, unsigned fine_location,
                          ir_variable *unpacked_var, const char *name,
                          bool gs_input_toplevel, unsigned vertex_index);
@@ -221,13 +226,19 @@ private:
     * appropriate place in the shader once the visitor has finished running.
     */
    exec_list *out_instructions;
+
+   /**
+    * Exec list into which the visitor should insert any new variables.
+    */
+   exec_list *out_variables;
 };
 
 } /* anonymous namespace */
 
 lower_packed_varyings_visitor::lower_packed_varyings_visitor(
       void *mem_ctx, unsigned locations_used, ir_variable_mode mode,
-      unsigned gs_input_vertices, exec_list *out_instructions)
+      unsigned gs_input_vertices, exec_list *out_instructions,
+      exec_list *out_variables)
    : mem_ctx(mem_ctx),
      locations_used(locations_used),
      packed_varyings((ir_variable **)
@@ -235,7 +246,8 @@ lower_packed_varyings_visitor::lower_packed_varyings_visitor(
                                         locations_used)),
      mode(mode),
      gs_input_vertices(gs_input_vertices),
-     out_instructions(out_instructions)
+     out_instructions(out_instructions),
+     out_variables(out_variables)
 {
 }
 
@@ -274,6 +286,7 @@ lower_packed_varyings_visitor::run(exec_list *instructions)
    }
 }
 
+#define SWIZZLE_ZWZW MAKE_SWIZZLE4(SWIZZLE_Z, SWIZZLE_W, SWIZZLE_Z, SWIZZLE_W)
 
 /**
  * Make an ir_assignment from \c rhs to \c lhs, performing appropriate
@@ -281,7 +294,7 @@ lower_packed_varyings_visitor::run(exec_list *instructions)
  *
  * This function is called when packing varyings.
  */
-ir_assignment *
+void
 lower_packed_varyings_visitor::bitwise_assign_pack(ir_rvalue *lhs,
                                                    ir_rvalue *rhs)
 {
@@ -300,12 +313,28 @@ lower_packed_varyings_visitor::bitwise_assign_pack(ir_rvalue *lhs,
          rhs = new(this->mem_ctx)
             ir_expression(ir_unop_bitcast_f2i, lhs->type, rhs);
          break;
+      case GLSL_TYPE_DOUBLE:
+         assert(rhs->type->vector_elements <= 2);
+         if (rhs->type->vector_elements == 2) {
+            ir_variable *t = new(mem_ctx) ir_variable(lhs->type, "pack", ir_var_temporary);
+
+            assert(lhs->type->vector_elements == 4);
+            this->out_variables->push_tail(t);
+            this->out_instructions->push_tail(
+                  assign(t, u2i(expr(ir_unop_unpack_double_2x32, swizzle_x(rhs->clone(mem_ctx, NULL)))), 0x3));
+            this->out_instructions->push_tail(
+                  assign(t,  u2i(expr(ir_unop_unpack_double_2x32, swizzle_y(rhs))), 0xc));
+            rhs = deref(t).val;
+         } else {
+            rhs = u2i(expr(ir_unop_unpack_double_2x32, rhs));
+         }
+         break;
       default:
          assert(!"Unexpected type conversion while lowering varyings");
          break;
       }
    }
-   return new(this->mem_ctx) ir_assignment(lhs, rhs);
+   this->out_instructions->push_tail(new (this->mem_ctx) ir_assignment(lhs, rhs));
 }
 
 
@@ -315,7 +344,7 @@ lower_packed_varyings_visitor::bitwise_assign_pack(ir_rvalue *lhs,
  *
  * This function is called when unpacking varyings.
  */
-ir_assignment *
+void
 lower_packed_varyings_visitor::bitwise_assign_unpack(ir_rvalue *lhs,
                                                      ir_rvalue *rhs)
 {
@@ -334,12 +363,27 @@ lower_packed_varyings_visitor::bitwise_assign_unpack(ir_rvalue *lhs,
          rhs = new(this->mem_ctx)
             ir_expression(ir_unop_bitcast_i2f, lhs->type, rhs);
          break;
+      case GLSL_TYPE_DOUBLE:
+         assert(lhs->type->vector_elements <= 2);
+         if (lhs->type->vector_elements == 2) {
+            ir_variable *t = new(mem_ctx) ir_variable(lhs->type, "unpack", ir_var_temporary);
+            assert(rhs->type->vector_elements == 4);
+            this->out_variables->push_tail(t);
+            this->out_instructions->push_tail(
+                  assign(t, expr(ir_unop_pack_double_2x32, i2u(swizzle_xy(rhs->clone(mem_ctx, NULL)))), 0x1));
+            this->out_instructions->push_tail(
+                  assign(t, expr(ir_unop_pack_double_2x32, i2u(swizzle(rhs->clone(mem_ctx, NULL), SWIZZLE_ZWZW, 2))), 0x2));
+            rhs = deref(t).val;
+         } else {
+            rhs = expr(ir_unop_pack_double_2x32, i2u(rhs));
+         }
+         break;
       default:
          assert(!"Unexpected type conversion while lowering varyings");
          break;
       }
    }
-   return new(this->mem_ctx) ir_assignment(lhs, rhs);
+   this->out_instructions->push_tail(new(this->mem_ctx) ir_assignment(lhs, rhs));
 }
 
 
@@ -372,6 +416,7 @@ lower_packed_varyings_visitor::lower_rvalue(ir_rvalue *rvalue,
                                             bool gs_input_toplevel,
                                             unsigned vertex_index)
 {
+   unsigned dmul = rvalue->type->is_double() ? 2 : 1;
    /* When gs_input_toplevel is set, we should be looking at a geometry shader
     * input array.
     */
@@ -405,17 +450,26 @@ lower_packed_varyings_visitor::lower_rvalue(ir_rvalue *rvalue,
       return this->lower_arraylike(rvalue, rvalue->type->matrix_columns,
                                    fine_location, unpacked_var, name,
                                    false, vertex_index);
-   } else if (rvalue->type->vector_elements + fine_location % 4 > 4) {
+   } else if (rvalue->type->vector_elements * dmul +
+              fine_location % 4 > 4) {
       /* This vector is going to be "double parked" across two varying slots,
-       * so handle it as two separate assignments.
+       * so handle it as two separate assignments. For doubles, a dvec3/dvec4
+       * can end up being spread over 3 slots. However the second splitting
+       * will happen later, here we just always want to split into 2.
        */
-      unsigned left_components = 4 - fine_location % 4;
-      unsigned right_components
-         = rvalue->type->vector_elements - left_components;
+      unsigned left_components, right_components;
       unsigned left_swizzle_values[4] = { 0, 0, 0, 0 };
       unsigned right_swizzle_values[4] = { 0, 0, 0, 0 };
       char left_swizzle_name[4] = { 0, 0, 0, 0 };
       char right_swizzle_name[4] = { 0, 0, 0, 0 };
+
+      left_components = 4 - fine_location % 4;
+      if (rvalue->type->is_double()) {
+         /* We might actually end up with 0 left components! */
+         left_components /= 2;
+      }
+      right_components = rvalue->type->vector_elements - left_components;
+
       for (unsigned i = 0; i < left_components; i++) {
          left_swizzle_values[i] = i;
          left_swizzle_name[i] = "xyzw"[i];
@@ -433,9 +487,13 @@ lower_packed_varyings_visitor::lower_rvalue(ir_rvalue *rvalue,
          = ralloc_asprintf(this->mem_ctx, "%s.%s", name, left_swizzle_name);
       char *right_name
          = ralloc_asprintf(this->mem_ctx, "%s.%s", name, right_swizzle_name);
-      fine_location = this->lower_rvalue(left_swizzle, fine_location,
-                                         unpacked_var, left_name, false,
-                                         vertex_index);
+      if (left_components)
+         fine_location = this->lower_rvalue(left_swizzle, fine_location,
+                                            unpacked_var, left_name, false,
+                                            vertex_index);
+      else
+         /* Top up the fine location to the next slot */
+         fine_location++;
       return this->lower_rvalue(right_swizzle, fine_location, unpacked_var,
                                 right_name, false, vertex_index);
    } else {
@@ -443,7 +501,7 @@ lower_packed_varyings_visitor::lower_rvalue(ir_rvalue *rvalue,
        * varying.
        */
       unsigned swizzle_values[4] = { 0, 0, 0, 0 };
-      unsigned components = rvalue->type->vector_elements;
+      unsigned components = rvalue->type->vector_elements * dmul;
       unsigned location = fine_location / 4;
       unsigned location_frac = fine_location % 4;
       for (unsigned i = 0; i < components; ++i)
@@ -454,13 +512,9 @@ lower_packed_varyings_visitor::lower_rvalue(ir_rvalue *rvalue,
       ir_swizzle *swizzle = new(this->mem_ctx)
          ir_swizzle(packed_deref, swizzle_values, components);
       if (this->mode == ir_var_shader_out) {
-         ir_assignment *assignment
-            = this->bitwise_assign_pack(swizzle, rvalue);
-         this->out_instructions->push_tail(assignment);
+         this->bitwise_assign_pack(swizzle, rvalue);
       } else {
-         ir_assignment *assignment
-            = this->bitwise_assign_unpack(rvalue, swizzle);
-         this->out_instructions->push_tail(assignment);
+         this->bitwise_assign_unpack(rvalue, swizzle);
       }
       return fine_location + components;
    }
@@ -598,7 +652,7 @@ lower_packed_varyings_visitor::needs_lowering(ir_variable *var)
    }
    if (type->is_array())
       type = type->fields.array;
-   if (type->vector_elements == 4)
+   if (type->vector_elements == 4 && !type->is_double())
       return false;
    return true;
 }
@@ -657,9 +711,11 @@ lower_packed_varyings(void *mem_ctx, unsigned locations_used,
    exec_list void_parameters;
    ir_function_signature *main_func_sig
       = main_func->matching_signature(NULL, &void_parameters, false);
-   exec_list new_instructions;
+   exec_list new_instructions, new_variables;
    lower_packed_varyings_visitor visitor(mem_ctx, locations_used, mode,
-                                         gs_input_vertices, &new_instructions);
+                                         gs_input_vertices,
+                                         &new_instructions,
+                                         &new_variables);
    visitor.run(instructions);
    if (mode == ir_var_shader_out) {
       if (shader->Stage == MESA_SHADER_GEOMETRY) {
@@ -667,15 +723,22 @@ lower_packed_varyings(void *mem_ctx, unsigned locations_used,
           * to EmitVertex()
           */
          lower_packed_varyings_gs_splicer splicer(mem_ctx, &new_instructions);
+
+         /* Add all the variables in first. */
+         main_func_sig->body.head->insert_before(&new_variables);
+
+         /* Now update all the EmitVertex instances */
          splicer.run(instructions);
       } else {
          /* For other shader types, outputs need to be lowered at the end of
           * main()
           */
+         main_func_sig->body.append_list(&new_variables);
          main_func_sig->body.append_list(&new_instructions);
       }
    } else {
       /* Shader inputs need to be lowered at the beginning of main() */
       main_func_sig->body.head->insert_before(&new_instructions);
+      main_func_sig->body.head->insert_before(&new_variables);
    }
 }
diff --git a/mesalib/src/glsl/nir/glsl_to_nir.cpp b/mesalib/src/glsl/nir/glsl_to_nir.cpp
index 544d0d932..adef19c80 100644
--- a/mesalib/src/glsl/nir/glsl_to_nir.cpp
+++ b/mesalib/src/glsl/nir/glsl_to_nir.cpp
@@ -563,8 +563,14 @@ nir_visitor::visit(ir_discard *ir)
     * discards will be immediately followed by a return.
     */
 
-   nir_intrinsic_instr *discard =
-      nir_intrinsic_instr_create(this->shader, nir_intrinsic_discard);
+   nir_intrinsic_instr *discard;
+   if (ir->condition) {
+      discard = nir_intrinsic_instr_create(this->shader,
+                                           nir_intrinsic_discard_if);
+      discard->src[0] = evaluate_rvalue(ir->condition);
+   } else {
+      discard = nir_intrinsic_instr_create(this->shader, nir_intrinsic_discard);
+   }
    nir_instr_insert_after_cf_list(this->cf_node_list, &discard->instr);
 }
 
diff --git a/mesalib/src/glsl/nir/nir.c b/mesalib/src/glsl/nir/nir.c
index 5b0e4bc50..ab57fd4e2 100644
--- a/mesalib/src/glsl/nir/nir.c
+++ b/mesalib/src/glsl/nir/nir.c
@@ -63,6 +63,7 @@ reg_create(void *mem_ctx, struct exec_list *list)
 {
    nir_register *reg = ralloc(mem_ctx, nir_register);
 
+   reg->parent_instr = NULL;
    reg->uses = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
                                 _mesa_key_pointer_equal);
    reg->defs = _mesa_set_create(mem_ctx, _mesa_hash_pointer,
diff --git a/mesalib/src/glsl/nir/nir.h b/mesalib/src/glsl/nir/nir.h
index d74caa959..d5df59609 100644
--- a/mesalib/src/glsl/nir/nir.h
+++ b/mesalib/src/glsl/nir/nir.h
@@ -66,6 +66,7 @@ name(const in_type *parent)                              \
 struct nir_function_overload;
 struct nir_function;
 struct nir_shader;
+struct nir_instr;
 
 
 /**
@@ -386,6 +387,14 @@ typedef struct {
     */
    bool is_packed;
 
+   /**
+    * If this pointer is non-NULL then this register has exactly one
+    * definition and that definition dominates all of its uses.  This is
+    * set by the out-of-SSA pass so that backends can get SSA-like
+    * information even once they have gone out of SSA.
+    */
+   struct nir_instr *parent_instr;
+
    /** set of nir_instr's where this register is used (read from) */
    struct set *uses;
 
@@ -408,7 +417,7 @@ typedef enum {
    nir_instr_type_parallel_copy,
 } nir_instr_type;
 
-typedef struct {
+typedef struct nir_instr {
    struct exec_node node;
    nir_instr_type type;
    struct nir_block *block;
diff --git a/mesalib/src/glsl/nir/nir_from_ssa.c b/mesalib/src/glsl/nir/nir_from_ssa.c
index 7c5009577..c3090fb06 100644
--- a/mesalib/src/glsl/nir/nir_from_ssa.c
+++ b/mesalib/src/glsl/nir/nir_from_ssa.c
@@ -26,6 +26,7 @@
  */
 
 #include "nir.h"
+#include "nir_vla.h"
 
 /*
  * This file implements an out-of-SSA pass as described in "Revisiting
@@ -181,7 +182,7 @@ merge_merge_sets(merge_set *a, merge_set *b)
 static bool
 merge_sets_interfere(merge_set *a, merge_set *b)
 {
-   merge_node *dom[a->size + b->size];
+   NIR_VLA(merge_node *, dom, a->size + b->size);
    int dom_idx = -1;
 
    struct exec_node *an = exec_list_get_head(&a->nodes);
@@ -508,6 +509,13 @@ get_register_for_ssa_def(nir_ssa_def *def, struct from_ssa_state *state)
       reg->num_components = def->num_components;
       reg->num_array_elems = 0;
 
+      /* This register comes from an SSA definition that was not part of a
+       * phi-web.  Therefore, we know it has a single unique definition
+       * that dominates all of its uses.  Therefore, we can copy the
+       * parent_instr from the SSA def safely.
+       */
+      reg->parent_instr = def->parent_instr;
+
       _mesa_hash_table_insert(state->ssa_table, def, reg);
       return reg;
    }
@@ -666,21 +674,16 @@ resolve_parallel_copy(nir_parallel_copy_instr *pcopy,
    }
 
    /* The register/source corresponding to the given index */
-   nir_src values[num_copies * 2];
-   memset(values, 0, sizeof values);
-
-   /* The current location of a given piece of data */
-   int loc[num_copies * 2];
+   NIR_VLA_ZERO(nir_src, values, num_copies * 2);
 
-   /* The piece of data that the given piece of data is to be copied from */
-   int pred[num_copies * 2];
+   /* The current location of a given piece of data.  We will use -1 for "null" */
+   NIR_VLA_FILL(int, loc, num_copies * 2, -1);
 
-   /* Initialize loc and pred.  We will use -1 for "null" */
-   memset(loc, -1, sizeof loc);
-   memset(pred, -1, sizeof pred);
+   /* The piece of data that the given piece of data is to be copied from.  We will use -1 for "null" */
+   NIR_VLA_FILL(int, pred, num_copies * 2, -1);
 
    /* The destinations we have yet to properly fill */
-   int to_do[num_copies * 2];
+   NIR_VLA(int, to_do, num_copies * 2);
    int to_do_idx = -1;
 
    /* Now we set everything up:
@@ -730,7 +733,7 @@ resolve_parallel_copy(nir_parallel_copy_instr *pcopy,
    }
 
    /* Currently empty destinations we can go ahead and fill */
-   int ready[num_copies * 2];
+   NIR_VLA(int, ready, num_copies * 2);
    int ready_idx = -1;
 
    /* Mark the ones that are ready for copying.  We know an index is a
diff --git a/mesalib/src/glsl/nir/nir_intrinsics.h b/mesalib/src/glsl/nir/nir_intrinsics.h
index d94866c85..3bf102fc1 100644
--- a/mesalib/src/glsl/nir/nir_intrinsics.h
+++ b/mesalib/src/glsl/nir/nir_intrinsics.h
@@ -68,6 +68,8 @@ INTRINSIC(interp_var_at_offset, 1, ARR(2), true, 0, 1, 0,
 #define BARRIER(name) INTRINSIC(name, 0, ARR(), false, 0, 0, 0, 0)
 
 BARRIER(discard)
+/** A conditional discard, with a single boolean source. */
+INTRINSIC(discard_if, 1, ARR(1), false, 0, 0, 0, 0)
 
 INTRINSIC(emit_vertex,   0, ARR(), false, 0, 0, 1, 0)
 INTRINSIC(end_primitive, 0, ARR(), false, 0, 0, 1, 0)
diff --git a/mesalib/src/glsl/nir/nir_live_variables.c b/mesalib/src/glsl/nir/nir_live_variables.c
index 7402dc087..1c96dcf36 100644
--- a/mesalib/src/glsl/nir/nir_live_variables.c
+++ b/mesalib/src/glsl/nir/nir_live_variables.c
@@ -26,6 +26,7 @@
 
 #include "nir.h"
 #include "nir_worklist.h"
+#include "nir_vla.h"
 
 /*
  * Basic liveness analysis.  This works only in SSA form.
@@ -130,8 +131,8 @@ static bool
 propagate_across_edge(nir_block *pred, nir_block *succ,
                       struct live_variables_state *state)
 {
-   BITSET_WORD live[state->bitset_words];
-   memcpy(live, succ->live_in, sizeof live);
+   NIR_VLA(BITSET_WORD, live, state->bitset_words);
+   memcpy(live, succ->live_in, state->bitset_words * sizeof *live);
 
    nir_foreach_instr(succ, instr) {
       if (instr->type != nir_instr_type_phi)
diff --git a/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c b/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
index 8af753029..9e9a418e3 100644
--- a/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
+++ b/mesalib/src/glsl/nir/nir_lower_vars_to_ssa.c
@@ -26,6 +26,8 @@
  */
 
 #include "nir.h"
+#include "nir_vla.h"
+
 
 struct deref_node {
    struct deref_node *parent;
@@ -899,8 +901,8 @@ rename_variables_block(nir_block *block, struct lower_variables_state *state)
 static void
 insert_phi_nodes(struct lower_variables_state *state)
 {
-   unsigned work[state->impl->num_blocks];
-   unsigned has_already[state->impl->num_blocks];
+   NIR_VLA_ZERO(unsigned, work, state->impl->num_blocks);
+   NIR_VLA_ZERO(unsigned, has_already, state->impl->num_blocks);
 
    /*
     * Since the work flags already prevent us from inserting a node that has
@@ -910,10 +912,7 @@ insert_phi_nodes(struct lower_variables_state *state)
     * function. So all we need to handle W is an array and a pointer to the
     * next element to be inserted and the next element to be removed.
     */
-   nir_block *W[state->impl->num_blocks];
-
-   memset(work, 0, sizeof work);
-   memset(has_already, 0, sizeof has_already);
+   NIR_VLA(nir_block *, W, state->impl->num_blocks);
 
    unsigned w_start, w_end;
    unsigned iter_count = 0;
diff --git a/mesalib/src/glsl/nir/nir_opt_copy_propagate.c b/mesalib/src/glsl/nir/nir_opt_copy_propagate.c
index dd0ec01ef..ee78e5aa0 100644
--- a/mesalib/src/glsl/nir/nir_opt_copy_propagate.c
+++ b/mesalib/src/glsl/nir/nir_opt_copy_propagate.c
@@ -53,22 +53,6 @@ static bool is_move(nir_alu_instr *instr)
 
 }
 
-static bool
-is_swizzleless_move(nir_alu_instr *instr)
-{
-   if (!is_move(instr))
-      return false;
-
-   for (unsigned i = 0; i < 4; i++) {
-      if (!((instr->dest.write_mask >> i) & 1))
-         break;
-      if (instr->src[0].swizzle[i] != i)
-         return false;
-   }
-
-   return true;
-}
-
 static bool is_vec(nir_alu_instr *instr)
 {
    for (unsigned i = 0; i < nir_op_infos[instr->op].num_inputs; i++)
@@ -80,6 +64,35 @@ static bool is_vec(nir_alu_instr *instr)
           instr->op == nir_op_vec4;
 }
 
+static bool
+is_swizzleless_move(nir_alu_instr *instr)
+{
+   if (is_move(instr)) {
+      for (unsigned i = 0; i < 4; i++) {
+         if (!((instr->dest.write_mask >> i) & 1))
+            break;
+         if (instr->src[0].swizzle[i] != i)
+            return false;
+      }
+      return true;
+   } else if (is_vec(instr)) {
+      nir_ssa_def *def = NULL;
+      for (unsigned i = 0; i < nir_op_infos[instr->op].num_inputs; i++) {
+         if (instr->src[i].swizzle[0] != i)
+            return false;
+
+         if (def == NULL) {
+            def = instr->src[i].src.ssa;
+         } else if (instr->src[i].src.ssa != def) {
+            return false;
+         }
+      }
+      return true;
+   } else {
+      return false;
+   }
+}
+
 typedef struct {
    nir_ssa_def *def;
    bool found;
diff --git a/mesalib/src/glsl/nir/nir_opt_gcm.c b/mesalib/src/glsl/nir/nir_opt_gcm.c
index bf565b969..b4f5fd3d5 100644
--- a/mesalib/src/glsl/nir/nir_opt_gcm.c
+++ b/mesalib/src/glsl/nir/nir_opt_gcm.c
@@ -121,9 +121,11 @@ gcm_pin_instructions_block(nir_block *block, void *void_state)
          case nir_op_fddy_coarse:
             /* These can only go in uniform control flow; pin them for now */
             instr->pass_flags = GCM_INSTR_PINNED;
+            break;
 
          default:
             instr->pass_flags = 0;
+            break;
          }
          break;
 
@@ -134,9 +136,11 @@ gcm_pin_instructions_block(nir_block *block, void *void_state)
          case nir_texop_lod:
             /* These two take implicit derivatives so they need to be pinned */
             instr->pass_flags = GCM_INSTR_PINNED;
+            break;
 
          default:
             instr->pass_flags = 0;
+            break;
          }
          break;
 
diff --git a/mesalib/src/glsl/nir/nir_vla.h b/mesalib/src/glsl/nir/nir_vla.h
new file mode 100644
index 000000000..753783316
--- /dev/null
+++ b/mesalib/src/glsl/nir/nir_vla.h
@@ -0,0 +1,54 @@
+/**************************************************************************
+ *
+ * Copyright 2015 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.
+ *
+ **************************************************************************/
+
+#pragma once
+
+
+#include "c99_alloca.h"
+
+
+/* Declare a variable length array, with no initialization */
+#define NIR_VLA(_type, _name, _length) \
+   _type *_name = alloca((_length) * sizeof *_name)
+
+
+/* Declare a variable length array, and initialize it with the given byte.
+ *
+ * _length is evaluated twice, so expressions with side-effects must be
+ * avoided.
+ */
+#define NIR_VLA_FILL(_type, _name, _length, _byte) \
+   _type *_name = memset(alloca((_length) * sizeof *_name), _byte, (_length) * sizeof *_name)
+
+
+/* Declare a variable length array, and zero it.
+ *
+ * Just like NIR_VLA_FILL, _length is evaluated twice, so expressions with
+ * side-effects must be avoided.
+ */
+#define NIR_VLA_ZERO(_type, _name, _length) \
+   NIR_VLA_FILL(_type, _name, _length, 0)
diff --git a/mesalib/src/glsl/opt_algebraic.cpp b/mesalib/src/glsl/opt_algebraic.cpp
index 6784242ff..c6040bff8 100644
--- a/mesalib/src/glsl/opt_algebraic.cpp
+++ b/mesalib/src/glsl/opt_algebraic.cpp
@@ -744,48 +744,65 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
        * a saturate operation
        */
       for (int op = 0; op < 2; op++) {
-         ir_expression *minmax = op_expr[op];
+         ir_expression *inner_expr = op_expr[op];
          ir_constant *outer_const = op_const[1 - op];
          ir_expression_operation op_cond = (ir->operation == ir_binop_max) ?
             ir_binop_min : ir_binop_max;
 
-         if (!minmax || !outer_const || (minmax->operation != op_cond))
+         if (!inner_expr || !outer_const || (inner_expr->operation != op_cond))
             continue;
 
+         /* One of these has to be a constant */
+         if (!inner_expr->operands[0]->as_constant() &&
+             !inner_expr->operands[1]->as_constant())
+            break;
+
          /* Found a min(max) combination. Now try to see if its operands
           * meet our conditions that we can do just a single saturate operation
           */
          for (int minmax_op = 0; minmax_op < 2; minmax_op++) {
-            ir_rvalue *inner_val_a = minmax->operands[minmax_op];
-            ir_rvalue *inner_val_b = minmax->operands[1 - minmax_op];
+            ir_rvalue *x = inner_expr->operands[minmax_op];
+            ir_rvalue *y = inner_expr->operands[1 - minmax_op];
 
-            if (!inner_val_a || !inner_val_b)
+            ir_constant *inner_const = y->as_constant();
+            if (!inner_const)
                continue;
 
-            /* Found a {min|max} ({max|min} (x, 0.0), 1.0) operation and its variations */
-            if ((outer_const->is_one() && inner_val_a->is_zero()) ||
-                (inner_val_a->is_one() && outer_const->is_zero()))
-               return saturate(inner_val_b);
-
-            /* Found a {min|max} ({max|min} (x, 0.0), b) where b < 1.0
-             * and its variations
-             */
-            if (is_less_than_one(outer_const) && inner_val_b->is_zero())
-               return expr(ir_binop_min, saturate(inner_val_a), outer_const);
-
-            if (!inner_val_b->as_constant())
-               continue;
-
-            if (is_less_than_one(inner_val_b->as_constant()) && outer_const->is_zero())
-               return expr(ir_binop_min, saturate(inner_val_a), inner_val_b);
-
-            /* Found a {min|max} ({max|min} (x, b), 1.0), where b > 0.0
-             * and its variations
-             */
-            if (outer_const->is_one() && is_greater_than_zero(inner_val_b->as_constant()))
-               return expr(ir_binop_max, saturate(inner_val_a), inner_val_b);
-            if (inner_val_b->as_constant()->is_one() && is_greater_than_zero(outer_const))
-               return expr(ir_binop_max, saturate(inner_val_a), outer_const);
+            /* min(max(x, 0.0), 1.0) is sat(x) */
+            if (ir->operation == ir_binop_min &&
+                inner_const->is_zero() &&
+                outer_const->is_one())
+               return saturate(x);
+
+            /* max(min(x, 1.0), 0.0) is sat(x) */
+            if (ir->operation == ir_binop_max &&
+                inner_const->is_one() &&
+                outer_const->is_zero())
+               return saturate(x);
+
+            /* min(max(x, 0.0), b) where b < 1.0 is sat(min(x, b)) */
+            if (ir->operation == ir_binop_min &&
+                inner_const->is_zero() &&
+                is_less_than_one(outer_const))
+               return saturate(expr(ir_binop_min, x, outer_const));
+
+            /* max(min(x, b), 0.0) where b < 1.0 is sat(min(x, b)) */
+            if (ir->operation == ir_binop_max &&
+                is_less_than_one(inner_const) &&
+                outer_const->is_zero())
+               return saturate(expr(ir_binop_min, x, inner_const));
+
+            /* max(min(x, 1.0), b) where b > 0.0 is sat(max(x, b)) */
+            if (ir->operation == ir_binop_max &&
+                inner_const->is_one() &&
+                is_greater_than_zero(outer_const))
+               return saturate(expr(ir_binop_max, x, outer_const));
+
+            /* min(max(x, b), 1.0) where b > 0.0 is sat(max(x, b)) */
+            if (ir->operation == ir_binop_min &&
+                is_greater_than_zero(inner_const) &&
+                outer_const->is_one())
+               return saturate(expr(ir_binop_max, x, inner_const));
          }
       }
 
@@ -849,6 +866,7 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
             one = new(mem_ctx) ir_constant(1.0, op2_components);
             break;
          default:
+            one = NULL;
             unreachable("unexpected type");
          }
 
diff --git a/mesalib/src/glsl/opt_conditional_discard.cpp b/mesalib/src/glsl/opt_conditional_discard.cpp
new file mode 100644
index 000000000..8a3ad2487
--- /dev/null
+++ b/mesalib/src/glsl/opt_conditional_discard.cpp
@@ -0,0 +1,81 @@
+/*
+ * Copyright © 2014 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * \file opt_conditional_discard.cpp
+ *
+ * Replace
+ *
+ *    if (cond) discard;
+ *
+ * with
+ *
+ *    (discard <condition>)
+ */
+
+#include "glsl_types.h"
+#include "ir.h"
+
+namespace {
+
+class opt_conditional_discard_visitor : public ir_hierarchical_visitor {
+public:
+   opt_conditional_discard_visitor()
+   {
+      progress = false;
+   }
+
+   ir_visitor_status visit_leave(ir_if *);
+
+   bool progress;
+};
+
+} /* anonymous namespace */
+
+bool
+opt_conditional_discard(exec_list *instructions)
+{
+   opt_conditional_discard_visitor v;
+   v.run(instructions);
+   return v.progress;
+}
+
+ir_visitor_status
+opt_conditional_discard_visitor::visit_leave(ir_if *ir)
+{
+   /* Look for "if (...) discard" with no else clause or extra statements. */
+   if (ir->then_instructions.is_empty() ||
+       !ir->then_instructions.head->next->is_tail_sentinel() ||
+       !((ir_instruction *) ir->then_instructions.head)->as_discard() ||
+       !ir->else_instructions.is_empty())
+      return visit_continue;
+
+   /* Move the condition and replace the ir_if with the ir_discard. */
+   ir_discard *discard = (ir_discard *) ir->then_instructions.head;
+   discard->condition = ir->condition;
+   ir->replace_with(discard);
+
+   progress = true;
+
+   return visit_continue;
+}
diff --git a/mesalib/src/glsl/opt_constant_folding.cpp b/mesalib/src/glsl/opt_constant_folding.cpp
index 74b855e5e..4aae3f0dd 100644
--- a/mesalib/src/glsl/opt_constant_folding.cpp
+++ b/mesalib/src/glsl/opt_constant_folding.cpp
@@ -50,6 +50,7 @@ public:
       /* empty */
    }
 
+   virtual ir_visitor_status visit_enter(ir_discard *ir);
    virtual ir_visitor_status visit_enter(ir_assignment *ir);
    virtual ir_visitor_status visit_enter(ir_call *ir);
 
@@ -94,6 +95,29 @@ ir_constant_folding_visitor::handle_rvalue(ir_rvalue **rvalue)
 }
 
 ir_visitor_status
+ir_constant_folding_visitor::visit_enter(ir_discard *ir)
+{
+   if (ir->condition) {
+      ir->condition->accept(this);
+      handle_rvalue(&ir->condition);
+
+      ir_constant *const_val = ir->condition->as_constant();
+      /* If the condition is constant, either remove the condition or
+       * remove the never-executed assignment.
+       */
+      if (const_val) {
+         if (const_val->value.b[0])
+            ir->condition = NULL;
+         else
+            ir->remove();
+         this->progress = true;
+      }
+   }
+
+   return visit_continue_with_parent;
+}
+
+ir_visitor_status
 ir_constant_folding_visitor::visit_enter(ir_assignment *ir)
 {
    ir->rhs->accept(this);
diff --git a/mesalib/src/glsl/s_expression.cpp b/mesalib/src/glsl/s_expression.cpp
index 2928a4db8..7eaa491e2 100644
--- a/mesalib/src/glsl/s_expression.cpp
+++ b/mesalib/src/glsl/s_expression.cpp
@@ -24,6 +24,7 @@
 
 #include <assert.h>
 #include <limits>
+#include <stdio.h>
 #include "s_expression.h"
 
 s_symbol::s_symbol(const char *str, size_t n)
diff --git a/mesalib/src/glsl/s_expression.h b/mesalib/src/glsl/s_expression.h
index 1d4753530..f0dffb1b2 100644
--- a/mesalib/src/glsl/s_expression.h
+++ b/mesalib/src/glsl/s_expression.h
@@ -39,8 +39,8 @@
 #define SX_AS_INT(x)    SX_AS_(int, x)
 
 /* Pattern matching macros */
-#define MATCH(list, pat) s_match(list, Elements(pat), pat, false)
-#define PARTIAL_MATCH(list, pat) s_match(list, Elements(pat), pat, true)
+#define MATCH(list, pat) s_match(list, ARRAY_SIZE(pat), pat, false)
+#define PARTIAL_MATCH(list, pat) s_match(list, ARRAY_SIZE(pat), pat, true)
 
 /* For our purposes, S-Expressions are:
  * - <int>
diff --git a/mesalib/src/glsl/standalone_scaffolding.cpp b/mesalib/src/glsl/standalone_scaffolding.cpp
index ad0d75bf8..6f5a27f85 100644
--- a/mesalib/src/glsl/standalone_scaffolding.cpp
+++ b/mesalib/src/glsl/standalone_scaffolding.cpp
@@ -30,6 +30,7 @@
 #include "standalone_scaffolding.h"
 
 #include <assert.h>
+#include <stdio.h>
 #include <string.h>
 #include "util/ralloc.h"
 
diff --git a/mesalib/src/loader/Makefile.am b/mesalib/src/loader/Makefile.am
index 36ddba82b..3d32279ea 100644
--- a/mesalib/src/loader/Makefile.am
+++ b/mesalib/src/loader/Makefile.am
@@ -30,6 +30,7 @@ libloader_la_CPPFLAGS = \
 	-I$(top_srcdir)/include \
 	-I$(top_srcdir)/src \
 	$(VISIBILITY_CFLAGS) \
+	$(MSVC2013_COMPAT_CFLAGS) \
 	$(LIBUDEV_CFLAGS)
 
 libloader_la_SOURCES = $(LOADER_C_FILES)
diff --git a/mesalib/src/loader/SConscript b/mesalib/src/loader/SConscript
index 359fc1850..16d1053ff 100644
--- a/mesalib/src/loader/SConscript
+++ b/mesalib/src/loader/SConscript
@@ -2,6 +2,8 @@ Import('*')
 
 env = env.Clone()
 
+env.MSVC2013Compat()
+
 env.Prepend(CPPPATH = [
     '#include'
 ])
diff --git a/mesalib/src/mapi/Makefile.am b/mesalib/src/mapi/Makefile.am
index 679468291..50c5b2ebb 100644
--- a/mesalib/src/mapi/Makefile.am
+++ b/mesalib/src/mapi/Makefile.am
@@ -39,11 +39,14 @@ EXTRA_DIST = \
 	glapi/SConscript \
 	shared-glapi/SConscript
 
-AM_CFLAGS = $(PTHREAD_CFLAGS)
+AM_CFLAGS = \
+	$(PTHREAD_CFLAGS) \
+	$(MSVC2013_COMPAT_CFLAGS)
 AM_CPPFLAGS =							\
 	$(DEFINES)						\
 	$(SELINUX_CFLAGS)					\
 	-I$(top_srcdir)/include					\
+	-I$(top_srcdir)/src					\
 	-I$(top_srcdir)/src/mapi				\
 	-I$(top_builddir)/src/mapi
 
@@ -66,7 +69,7 @@ if HAVE_SHARED_GLAPI
 BUILT_SOURCES += shared-glapi/glapi_mapi_tmp.h
 
 lib_LTLIBRARIES += shared-glapi/libglapi.la
-shared_glapi_libglapi_la_SOURCES = $(MAPI_GLAPI_FILES)
+shared_glapi_libglapi_la_SOURCES = $(MAPI_GLAPI_FILES) shared-glapi/glapi_mapi_tmp.h
 shared_glapi_libglapi_la_CPPFLAGS = \
 	$(AM_CPPFLAGS) \
 	-DMAPI_MODE_GLAPI \
@@ -115,7 +118,7 @@ glapi_libglapi_la_CPPFLAGS = \
 	-I$(top_srcdir)/src/mesa
 
 if HAVE_SHARED_GLAPI
-glapi_libglapi_la_SOURCES += $(MAPI_BRIDGE_FILES)
+glapi_libglapi_la_SOURCES += $(MAPI_BRIDGE_FILES) glapi/glapi_mapi_tmp.h
 glapi_libglapi_la_CPPFLAGS += \
 	-DMAPI_MODE_BRIDGE \
 	-DMAPI_ABI_HEADER=\"glapi/glapi_mapi_tmp.h\"
@@ -230,10 +233,4 @@ endif
 es2api/glapi_mapi_tmp.h: glapi/gen/gl_and_es_API.xml $(glapi_gen_mapi_deps)
 	$(call glapi_gen_mapi,$<,es2api)
 
-# XXX: Inline vgapi's Makefile.am here.
-EXTRA_DIST += vgapi
-# if HAVE_OPENVG
-# SUBDIRS += vgapi
-# endif
-
 include $(top_srcdir)/install-lib-links.mk
diff --git a/mesalib/src/mapi/Makefile.sources b/mesalib/src/mapi/Makefile.sources
index 41dbb24c8..07063f390 100644
--- a/mesalib/src/mapi/Makefile.sources
+++ b/mesalib/src/mapi/Makefile.sources
@@ -15,13 +15,10 @@
 #    this mode, compile MAPI_BRIDGE_FILES with MAPI_MODE_BRIDGE defined.
 
 MAPI_UTIL_FILES = \
-	u_compiler.h \
 	u_current.c \
 	u_current.h \
 	u_execmem.c \
-	u_execmem.h \
-	u_macros.h \
-	u_thread.h
+	u_execmem.h
 
 MAPI_BRIDGE_FILES = \
 	entry.c \
diff --git a/mesalib/src/mapi/entry.c b/mesalib/src/mapi/entry.c
index b6e8db28d..27d0db40e 100644
--- a/mesalib/src/mapi/entry.c
+++ b/mesalib/src/mapi/entry.c
@@ -27,7 +27,9 @@
 
 #include "entry.h"
 #include "u_current.h"
-#include "u_macros.h"
+
+#define _U_STRINGIFY(x) #x
+#define U_STRINGIFY(x) _U_STRINGIFY(x)
 
 /* define macros for use by assembly dispatchers */
 #define ENTRY_CURRENT_TABLE U_STRINGIFY(u_current_table)
@@ -51,7 +53,7 @@
 
 #include <stdlib.h>
 
-static INLINE const struct mapi_table *
+static inline const struct mapi_table *
 entry_current_get(void)
 {
 #ifdef MAPI_MODE_BRIDGE
diff --git a/mesalib/src/mapi/entry.h b/mesalib/src/mapi/entry.h
index dba1c06a3..7c8137c0c 100644
--- a/mesalib/src/mapi/entry.h
+++ b/mesalib/src/mapi/entry.h
@@ -28,7 +28,6 @@
 #ifndef _ENTRY_H_
 #define _ENTRY_H_
 
-#include "u_compiler.h"
 
 typedef void (*mapi_func)(void);
 
diff --git a/mesalib/src/mapi/entry_x86-64_tls.h b/mesalib/src/mapi/entry_x86-64_tls.h
index 71e9d6063..5c03b0456 100644
--- a/mesalib/src/mapi/entry_x86-64_tls.h
+++ b/mesalib/src/mapi/entry_x86-64_tls.h
@@ -25,7 +25,6 @@
  *    Chia-I Wu <olv@lunarg.com>
  */
 
-#include "u_macros.h"
 
 __asm__(".text\n"
         ".balign 32\n"
@@ -88,7 +87,7 @@ entry_generate(int slot)
       0x41, 0xff, 0xa3, 0x34, 0x12, 0x00, 0x00,
    };
    unsigned long addr;
-   void *code;
+   char *code;
    mapi_func entry;
 
    addr = x86_64_current_tls();
diff --git a/mesalib/src/mapi/entry_x86_tls.h b/mesalib/src/mapi/entry_x86_tls.h
index fa7bc15b8..46d2eced2 100644
--- a/mesalib/src/mapi/entry_x86_tls.h
+++ b/mesalib/src/mapi/entry_x86_tls.h
@@ -26,7 +26,6 @@
  */
 
 #include <string.h>
-#include "u_macros.h"
 
 __asm__(".text");
 
@@ -113,7 +112,7 @@ entry_generate(int slot)
       0xff, 0xa0, 0x34, 0x12, 0x00, 0x00, /* jmp *0x1234(%eax) */
       0x90, 0x90, 0x90, 0x90              /* nop's */
    };
-   void *code;
+   char *code;
    mapi_func entry;
 
    code = u_execmem_alloc(sizeof(code_templ));
diff --git a/mesalib/src/mapi/entry_x86_tsd.h b/mesalib/src/mapi/entry_x86_tsd.h
index ece00fac0..ea7bacb43 100644
--- a/mesalib/src/mapi/entry_x86_tsd.h
+++ b/mesalib/src/mapi/entry_x86_tsd.h
@@ -25,7 +25,6 @@
  *    Chia-I Wu <olv@lunarg.com>
  */
 
-#include "u_macros.h"
 
 #define X86_ENTRY_SIZE 32
 
diff --git a/mesalib/src/mapi/glapi/SConscript b/mesalib/src/mapi/glapi/SConscript
index 97ebfe638..8ded46f32 100644
--- a/mesalib/src/mapi/glapi/SConscript
+++ b/mesalib/src/mapi/glapi/SConscript
@@ -8,6 +8,8 @@ Import('*')
 
 env = env.Clone()
 
+env.MSVC2013Compat()
+
 env.Append(CPPDEFINES = [
     'MAPI_MODE_UTIL',
 ])
@@ -25,6 +27,7 @@ if env['platform'] == 'windows':
         env.Append(CPPDEFINES = ['_GLAPI_NO_EXPORTS'])
 
 env.Append(CPPPATH = [
+    '#/src',
     '#/src/mapi',
     '#/src/mesa',
     Dir('..'), # src/mapi build path
diff --git a/mesalib/src/mapi/glapi/gen/gl_enums.py b/mesalib/src/mapi/glapi/gen/gl_enums.py
index 35919d68a..d61618f4d 100644
--- a/mesalib/src/mapi/glapi/gen/gl_enums.py
+++ b/mesalib/src/mapi/glapi/gen/gl_enums.py
@@ -83,7 +83,7 @@ const char *_mesa_lookup_enum_by_nr( int nr )
    STATIC_ASSERT(sizeof(enum_string_table) < (1 << 16));
 
    elt = bsearch(& nr, enum_string_table_offsets,
-                 Elements(enum_string_table_offsets),
+                 ARRAY_SIZE(enum_string_table_offsets),
                  sizeof(enum_string_table_offsets[0]),
                  (cfunc) compar_nr);
 
@@ -127,7 +127,7 @@ static const char *prim_names[PRIM_MAX+3] = {
 const char *
 _mesa_lookup_prim_by_nr(GLuint nr)
 {
-   if (nr < Elements(prim_names))
+   if (nr < ARRAY_SIZE(prim_names))
       return prim_names[nr];
    else
       return "invalid mode";
diff --git a/mesalib/src/mapi/glapi/gen/gl_x86-64_asm.py b/mesalib/src/mapi/glapi/gen/gl_x86-64_asm.py
index 19e0e15b0..5a69e0343 100644
--- a/mesalib/src/mapi/glapi/gen/gl_x86-64_asm.py
+++ b/mesalib/src/mapi/glapi/gen/gl_x86-64_asm.py
@@ -138,10 +138,6 @@ class PrintGenericStubs(gl_XML.gl_print_base):
         print '#  define GL_PREFIX(n) GLNAME(CONCAT(gl,n))'
         print '# endif'
         print ''
-        print '#if defined(HAVE_PTHREAD) || defined(_WIN32)'
-        print '#  define THREADS'
-        print '#endif'
-        print ''
         print '\t.text'
         print ''
         print '#ifdef GLX_USE_TLS'
@@ -170,7 +166,7 @@ class PrintGenericStubs(gl_XML.gl_print_base):
         print '\tmovl\t(%rax), %edi'
         print '\tjmp\tpthread_getspecific@PLT'
         print ''
-        print '#elif defined(THREADS)'
+        print '#else'
         print ''
         print '\t.extern\t_glapi_get_dispatch'
         print ''
diff --git a/mesalib/src/mapi/glapi/gen/gl_x86_asm.py b/mesalib/src/mapi/glapi/gen/gl_x86_asm.py
index d87d0bd24..f855dbaaa 100644
--- a/mesalib/src/mapi/glapi/gen/gl_x86_asm.py
+++ b/mesalib/src/mapi/glapi/gen/gl_x86_asm.py
@@ -78,9 +78,6 @@ class PrintGenericStubs(gl_XML.gl_print_base):
         print '#define GLOBL_FN(x) GLOBL x'
         print '#endif'
         print ''
-        print '#if defined(HAVE_PTHREAD) || defined(_WIN32)'
-        print '#  define THREADS'
-        print '#endif'
         print ''
         print '#ifdef GLX_USE_TLS'
         print ''
@@ -109,7 +106,7 @@ class PrintGenericStubs(gl_XML.gl_print_base):
         print '\tJMP(GL_OFFSET(off)) ;\t\t\t\t\\'
         print '1:\tCALL(_x86_get_dispatch) ;\t\t\t\\'
         print '\tJMP(GL_OFFSET(off))'
-        print '#elif defined(THREADS)'
+        print '#else'
         print '#  define GL_STUB(fn,off,fn_alt)\t\t\t\\'
         print 'ALIGNTEXT16;\t\t\t\t\t\t\\'
         print 'GLOBL_FN(GL_PREFIX(fn, fn_alt));\t\t\t\\'
@@ -120,13 +117,6 @@ class PrintGenericStubs(gl_XML.gl_print_base):
         print '\tJMP(GL_OFFSET(off)) ;\t\t\t\t\\'
         print '1:\tCALL(_glapi_get_dispatch) ;\t\t\t\\'
         print '\tJMP(GL_OFFSET(off))'
-        print '#else /* Non-threaded version. */'
-        print '#  define GL_STUB(fn,off,fn_alt)\t\t\t\\'
-        print 'ALIGNTEXT16;\t\t\t\t\t\t\\'
-        print 'GLOBL_FN(GL_PREFIX(fn, fn_alt));\t\t\t\\'
-        print 'GL_PREFIX(fn, fn_alt):\t\t\t\t\t\\'
-        print '\tMOV_L(CONTENT(GLNAME(_glapi_Dispatch)), EAX) ;\t\\'
-        print '\tJMP(GL_OFFSET(off))'
         print '#endif'
         print ''
         print '#ifdef HAVE_ALIAS'
@@ -164,7 +154,7 @@ class PrintGenericStubs(gl_XML.gl_print_base):
         print '\tCALL(GLNAME(pthread_getspecific))'
         print '\tADD_L(CONST(28), ESP)'
         print '\tRET'
-        print '#elif defined(THREADS)'
+        print '#else'
         print 'EXTERN GLNAME(_glapi_get_dispatch)'
         print '#endif'
         print ''
diff --git a/mesalib/src/mapi/glapi/glapi.h b/mesalib/src/mapi/glapi/glapi.h
index e2fa9252f..8d991fb3b 100644
--- a/mesalib/src/mapi/glapi/glapi.h
+++ b/mesalib/src/mapi/glapi/glapi.h
@@ -44,7 +44,7 @@
 #ifndef _GLAPI_H
 #define _GLAPI_H
 
-#include "u_thread.h"
+#include "util/macros.h"
 
 
 #ifdef __cplusplus
@@ -102,21 +102,12 @@ _GLAPI_EXPORT extern const void *_glapi_Context;
 _GLAPI_EXPORT extern struct _glapi_table *_glapi_Dispatch;
 _GLAPI_EXPORT extern void *_glapi_Context;
 
-# ifdef THREADS
-
-#  define GET_DISPATCH() \
+#define GET_DISPATCH() \
      (likely(_glapi_Dispatch) ? _glapi_Dispatch : _glapi_get_dispatch())
 
-#  define GET_CURRENT_CONTEXT(C)  struct gl_context *C = (struct gl_context *) \
+#define GET_CURRENT_CONTEXT(C)  struct gl_context *C = (struct gl_context *) \
      (likely(_glapi_Context) ? _glapi_Context : _glapi_get_context())
 
-# else
-
-#  define GET_DISPATCH() _glapi_Dispatch
-#  define GET_CURRENT_CONTEXT(C)  struct gl_context *C = (struct gl_context *) _glapi_Context
-
-# endif
-
 #endif /* defined (GLX_USE_TLS) */
 
 
diff --git a/mesalib/src/mapi/glapi/glapi_entrypoint.c b/mesalib/src/mapi/glapi/glapi_entrypoint.c
index 362a8f11b..7facd8aca 100644
--- a/mesalib/src/mapi/glapi/glapi_entrypoint.c
+++ b/mesalib/src/mapi/glapi/glapi_entrypoint.c
@@ -29,6 +29,7 @@
  */
 
 
+#include "c11/threads.h"
 #include "glapi/glapi_priv.h"
 #include "u_execmem.h"
 
@@ -120,11 +121,9 @@ fill_in_entrypoint_offset(_glapi_proc entrypoint, unsigned int offset)
 
 #if defined(GLX_USE_TLS)
    *((unsigned int *)(code +  8)) = 4 * offset;
-#elif defined(THREADS)
+#else
    *((unsigned int *)(code + 11)) = 4 * offset;
    *((unsigned int *)(code + 22)) = 4 * offset;
-#else
-   *((unsigned int *)(code +  7)) = 4 * offset;
 #endif
 }
 
@@ -338,7 +337,7 @@ void
 init_glapi_relocs_once( void )
 {
 #if defined(HAVE_PTHREAD) || defined(GLX_USE_TLS)
-   static pthread_once_t once_control = PTHREAD_ONCE_INIT;
-   pthread_once( & once_control, init_glapi_relocs );
+   static once_flag flag = ONCE_FLAG_INIT;
+   call_once(&flag, init_glapi_relocs);
 #endif
 }
diff --git a/mesalib/src/mapi/glapi/glapi_priv.h b/mesalib/src/mapi/glapi/glapi_priv.h
index 92925faac..d3682603d 100644
--- a/mesalib/src/mapi/glapi/glapi_priv.h
+++ b/mesalib/src/mapi/glapi/glapi_priv.h
@@ -86,10 +86,8 @@ get_entrypoint_address(unsigned int functionOffset);
 #if defined(USE_X86_ASM)
 # if defined(GLX_USE_TLS)
 #  define DISPATCH_FUNCTION_SIZE  16
-# elif defined(THREADS)
-#  define DISPATCH_FUNCTION_SIZE  32
 # else
-#  define DISPATCH_FUNCTION_SIZE  16
+#  define DISPATCH_FUNCTION_SIZE  32
 # endif
 #endif
 
diff --git a/mesalib/src/mapi/mapi.c b/mesalib/src/mapi/mapi.c
index aa6b91b42..c235adc4a 100644
--- a/mesalib/src/mapi/mapi.c
+++ b/mesalib/src/mapi/mapi.c
@@ -29,7 +29,6 @@
 #include <string.h>
 
 #include "u_current.h"
-#include "u_thread.h"
 #include "mapi.h"
 #include "stub.h"
 #include "table.h"
diff --git a/mesalib/src/mapi/mapi.h b/mesalib/src/mapi/mapi.h
index 2aa8d9fe0..9adf8b5a3 100644
--- a/mesalib/src/mapi/mapi.h
+++ b/mesalib/src/mapi/mapi.h
@@ -28,8 +28,6 @@
 #ifndef _MAPI_H_
 #define _MAPI_H_
 
-#include "u_compiler.h"
-
 #ifdef _WIN32
 #ifdef MAPI_DLL_EXPORTS
 #define MAPI_EXPORT __declspec(dllexport)
diff --git a/mesalib/src/mapi/mapi_abi.py b/mesalib/src/mapi/mapi_abi.py
index d783ed32f..012a5c3f4 100644
--- a/mesalib/src/mapi/mapi_abi.py
+++ b/mesalib/src/mapi/mapi_abi.py
@@ -816,25 +816,8 @@ typedef int GLclampx;
 
         return header
 
-class VGAPIPrinter(ABIPrinter):
-    """OpenVG API Printer"""
-
-    def __init__(self, entries):
-        super(VGAPIPrinter, self).__init__(entries)
-
-        self.api_defines = ['VG_VGEXT_PROTOTYPES']
-        self.api_headers = ['"VG/openvg.h"', '"VG/vgext.h"']
-        self.api_call = 'VG_API_CALL'
-        self.api_entry = 'VG_API_ENTRY'
-        self.api_attrs = 'VG_API_EXIT'
-
-        self.prefix_lib = 'vg'
-        self.prefix_app = 'vega'
-        self.prefix_noop = 'noop'
-        self.prefix_warn = 'vg'
-
 def parse_args():
-    printers = ['vgapi', 'glapi', 'es1api', 'es2api', 'shared-glapi']
+    printers = ['glapi', 'es1api', 'es2api', 'shared-glapi']
     modes = ['lib', 'app']
 
     parser = OptionParser(usage='usage: %prog [options] <filename>')
@@ -853,7 +836,6 @@ def parse_args():
 
 def main():
     printers = {
-        'vgapi': VGAPIPrinter,
         'glapi': GLAPIPrinter,
         'es1api': ES1APIPrinter,
         'es2api': ES2APIPrinter,
diff --git a/mesalib/src/mapi/stub.c b/mesalib/src/mapi/stub.c
index dfadbe1a5..05436bab6 100644
--- a/mesalib/src/mapi/stub.c
+++ b/mesalib/src/mapi/stub.c
@@ -28,14 +28,14 @@
 #include <stdlib.h>
 #include <string.h>
 #include <assert.h>
+#include "c11/threads.h"
 
+#include "util/macros.h"
 #include "u_current.h"
-#include "u_thread.h"
 #include "entry.h"
 #include "stub.h"
 #include "table.h"
 
-#define ARRAY_SIZE(x) (sizeof(x)/sizeof((x)[0]))
 
 struct mapi_stub {
    const void *name;
@@ -54,16 +54,8 @@ static int next_dynamic_slot = MAPI_TABLE_NUM_STATIC;
 void
 stub_init_once(void)
 {
-#ifdef HAVE_PTHREAD
-   static pthread_once_t once = PTHREAD_ONCE_INIT;
-   pthread_once(&once, entry_patch_public);
-#else
-   static int first = 1;
-   if (first) {
-      first = 0;
-      entry_patch_public();
-   }
-#endif
+   static once_flag flag = ONCE_FLAG_INIT;
+   call_once(&flag, entry_patch_public);
 }
 
 static int
diff --git a/mesalib/src/mapi/table.h b/mesalib/src/mapi/table.h
index df59aed4f..e2d6ef03d 100644
--- a/mesalib/src/mapi/table.h
+++ b/mesalib/src/mapi/table.h
@@ -28,7 +28,7 @@
 #ifndef _TABLE_H_
 #define _TABLE_H_
 
-#include "u_compiler.h"
+#include "c99_compat.h"
 #include "entry.h"
 
 #define MAPI_TMP_TABLE
@@ -37,12 +37,14 @@
 #define MAPI_TABLE_NUM_SLOTS (MAPI_TABLE_NUM_STATIC + MAPI_TABLE_NUM_DYNAMIC)
 #define MAPI_TABLE_SIZE (MAPI_TABLE_NUM_SLOTS * sizeof(mapi_func))
 
+struct mapi_table;
+
 extern const mapi_func table_noop_array[];
 
 /**
  * Get the no-op dispatch table.
  */
-static INLINE const struct mapi_table *
+static inline const struct mapi_table *
 table_get_noop(void)
 {
    return (const struct mapi_table *) table_noop_array;
@@ -51,7 +53,7 @@ table_get_noop(void)
 /**
  * Set the function of a slot.
  */
-static INLINE void
+static inline void
 table_set_func(struct mapi_table *tbl, int slot, mapi_func func)
 {
    mapi_func *funcs = (mapi_func *) tbl;
@@ -61,7 +63,7 @@ table_set_func(struct mapi_table *tbl, int slot, mapi_func func)
 /**
  * Return the function of a slot.
  */
-static INLINE mapi_func
+static inline mapi_func
 table_get_func(const struct mapi_table *tbl, int slot)
 {
    const mapi_func *funcs = (const mapi_func *) tbl;
diff --git a/mesalib/src/mapi/u_compiler.h b/mesalib/src/mapi/u_compiler.h
deleted file mode 100644
index f376e97a0..000000000
--- a/mesalib/src/mapi/u_compiler.h
+++ /dev/null
@@ -1,33 +0,0 @@
-#ifndef _U_COMPILER_H_
-#define _U_COMPILER_H_
-
-#include "c99_compat.h" /* inline, __func__, etc. */
-
-
-/* XXX: Use standard `inline` keyword instead */
-#ifndef INLINE
-#  define INLINE inline
-#endif
-
-/* Function visibility */
-#ifndef PUBLIC
-#  if defined(__GNUC__) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x590))
-#    define PUBLIC __attribute__((visibility("default")))
-#  elif defined(_MSC_VER)
-#    define PUBLIC __declspec(dllexport)
-#  else
-#    define PUBLIC
-#  endif
-#endif
-
-#ifndef likely
-#  if defined(__GNUC__)
-#    define likely(x)   __builtin_expect(!!(x), 1)
-#    define unlikely(x) __builtin_expect(!!(x), 0)
-#  else
-#    define likely(x)   (x)
-#    define unlikely(x) (x)
-#  endif
-#endif
-
-#endif /* _U_COMPILER_H_ */
diff --git a/mesalib/src/mapi/u_current.c b/mesalib/src/mapi/u_current.c
index afa887e06..7e7e275f2 100644
--- a/mesalib/src/mapi/u_current.c
+++ b/mesalib/src/mapi/u_current.c
@@ -48,8 +48,8 @@
  *                 drivers!  No changes to the public glapi interface.
  */
 
+#include "c11/threads.h"
 #include "u_current.h"
-#include "u_thread.h"
 
 #ifndef MAPI_MODE_UTIL
 
@@ -112,11 +112,9 @@ struct mapi_table *u_current_table =
    (struct mapi_table *) table_noop_array;
 void *u_current_context;
 
-#ifdef THREADS
-struct u_tsd u_current_table_tsd;
-static struct u_tsd u_current_context_tsd;
+tss_t u_current_table_tsd;
+static tss_t u_current_context_tsd;
 static int ThreadSafe;
-#endif /* THREADS */
 
 #endif /* defined(GLX_USE_TLS) */
 /*@}*/
@@ -125,20 +123,20 @@ static int ThreadSafe;
 void
 u_current_destroy(void)
 {
-#if defined(THREADS) && defined(_WIN32)
-   u_tsd_destroy(&u_current_table_tsd);
-   u_tsd_destroy(&u_current_context_tsd);
+#if !defined(GLX_USE_TLS)
+   tss_delete(u_current_table_tsd);
+   tss_delete(u_current_context_tsd);
 #endif
 }
 
 
-#if defined(THREADS) && !defined(GLX_USE_TLS)
+#if !defined(GLX_USE_TLS)
 
 static void
 u_current_init_tsd(void)
 {
-   u_tsd_init(&u_current_table_tsd);
-   u_tsd_init(&u_current_context_tsd);
+   tss_create(&u_current_table_tsd, NULL);
+   tss_create(&u_current_context_tsd, NULL);
 }
 
 /**
@@ -146,6 +144,43 @@ u_current_init_tsd(void)
  */
 static mtx_t ThreadCheckMutex = _MTX_INITIALIZER_NP;
 
+
+#ifdef _WIN32
+typedef DWORD thread_id;
+#else
+typedef thrd_t thread_id;
+#endif
+
+
+static inline thread_id
+get_thread_id(void)
+{
+   /*
+    * XXX: Callers of of this function assume it is a lightweight function.
+    * But unfortunately C11's thrd_current() gives no such guarantees.  In
+    * fact, it's pretty hard to have a compliant implementation of
+    * thrd_current() on Windows with such characteristics.  So for now, we
+    * side-step this mess and use Windows thread primitives directly here.
+    */
+#ifdef _WIN32
+   return GetCurrentThreadId();
+#else
+   return thrd_current();
+#endif
+}
+
+
+static inline int
+thread_id_equal(thread_id t1, thread_id t2)
+{
+#ifdef _WIN32
+   return t1 == t2;
+#else
+   return thrd_equal(t1, t2);
+#endif
+}
+
+
 /**
  * We should call this periodically from a function such as glXMakeCurrent
  * in order to test if multiple threads are being used.
@@ -153,7 +188,7 @@ static mtx_t ThreadCheckMutex = _MTX_INITIALIZER_NP;
 void
 u_current_init(void)
 {
-   static unsigned long knownID;
+   static thread_id knownID;
    static int firstCall = 1;
 
    if (ThreadSafe)
@@ -163,10 +198,10 @@ u_current_init(void)
    if (firstCall) {
       u_current_init_tsd();
 
-      knownID = u_thread_self();
+      knownID = get_thread_id();
       firstCall = 0;
    }
-   else if (knownID != u_thread_self()) {
+   else if (!thread_id_equal(knownID, get_thread_id())) {
       ThreadSafe = 1;
       u_current_set_table(NULL);
       u_current_set_context(NULL);
@@ -197,11 +232,9 @@ u_current_set_context(const void *ptr)
 
 #if defined(GLX_USE_TLS)
    u_current_context = (void *) ptr;
-#elif defined(THREADS)
-   u_tsd_set(&u_current_context_tsd, (void *) ptr);
-   u_current_context = (ThreadSafe) ? NULL : (void *) ptr;
 #else
-   u_current_context = (void *) ptr;
+   tss_set(u_current_context_tsd, (void *) ptr);
+   u_current_context = (ThreadSafe) ? NULL : (void *) ptr;
 #endif
 }
 
@@ -215,12 +248,8 @@ u_current_get_context_internal(void)
 {
 #if defined(GLX_USE_TLS)
    return u_current_context;
-#elif defined(THREADS)
-   return (ThreadSafe)
-      ? u_tsd_get(&u_current_context_tsd)
-      : u_current_context;
 #else
-   return u_current_context;
+   return ThreadSafe ? tss_get(u_current_context_tsd) : u_current_context;
 #endif
 }
 
@@ -241,11 +270,9 @@ u_current_set_table(const struct mapi_table *tbl)
 
 #if defined(GLX_USE_TLS)
    u_current_table = (struct mapi_table *) tbl;
-#elif defined(THREADS)
-   u_tsd_set(&u_current_table_tsd, (void *) tbl);
-   u_current_table = (ThreadSafe) ? NULL : (void *) tbl;
 #else
-   u_current_table = (struct mapi_table *) tbl;
+   tss_set(u_current_table_tsd, (void *) tbl);
+   u_current_table = (ThreadSafe) ? NULL : (void *) tbl;
 #endif
 }
 
@@ -257,10 +284,10 @@ u_current_get_table_internal(void)
 {
 #if defined(GLX_USE_TLS)
    return u_current_table;
-#elif defined(THREADS)
-   return (struct mapi_table *) ((ThreadSafe) ?
-         u_tsd_get(&u_current_table_tsd) : (void *) u_current_table);
 #else
-   return u_current_table;
+   if (ThreadSafe)
+      return (struct mapi_table *) tss_get(u_current_table_tsd);
+   else
+      return (struct mapi_table *) u_current_table;
 #endif
 }
diff --git a/mesalib/src/mapi/u_current.h b/mesalib/src/mapi/u_current.h
index 72708d433..ea4f81723 100644
--- a/mesalib/src/mapi/u_current.h
+++ b/mesalib/src/mapi/u_current.h
@@ -1,6 +1,10 @@
 #ifndef _U_CURRENT_H_
 #define _U_CURRENT_H_
 
+#include "c99_compat.h"
+#include "util/macros.h"
+
+
 #if defined(MAPI_MODE_UTIL) || defined(MAPI_MODE_GLAPI) || \
     defined(MAPI_MODE_BRIDGE)
 
@@ -24,8 +28,6 @@
 
 #else /* MAPI_MODE_UTIL || MAPI_MODE_GLAPI || MAPI_MODE_BRIDGE */
 
-#include "u_compiler.h"
-
 struct mapi_table;
 
 #ifdef GLX_USE_TLS
@@ -63,7 +65,7 @@ u_current_set_context(const void *ptr);
 void *
 u_current_get_context_internal(void);
 
-static INLINE const struct mapi_table *
+static inline const struct mapi_table *
 u_current_get_table(void)
 {
 #ifdef GLX_USE_TLS
@@ -74,7 +76,7 @@ u_current_get_table(void)
 #endif
 }
 
-static INLINE const void *
+static inline const void *
 u_current_get_context(void)
 {
 #ifdef GLX_USE_TLS
diff --git a/mesalib/src/mapi/u_execmem.c b/mesalib/src/mapi/u_execmem.c
index ac1cae093..89d5c1d08 100644
--- a/mesalib/src/mapi/u_execmem.c
+++ b/mesalib/src/mapi/u_execmem.c
@@ -32,8 +32,8 @@
  */
 
 
-#include "u_compiler.h"
-#include "u_thread.h"
+#include "c99_compat.h"
+#include "c11/threads.h"
 #include "u_execmem.h"
 
 
diff --git a/mesalib/src/mapi/u_macros.h b/mesalib/src/mapi/u_macros.h
deleted file mode 100644
index 72345b5f1..000000000
--- a/mesalib/src/mapi/u_macros.h
+++ /dev/null
@@ -1,12 +0,0 @@
-#ifndef _U_MACROS_
-#define _U_MACROS_
-
-#define _U_STRINGIFY(x) #x
-#define _U_CONCAT(x, y) x ## y
-#define _U_CONCAT_STR(x, y) #x#y
-
-#define U_STRINGIFY(x) _U_STRINGIFY(x)
-#define U_CONCAT(x, y) _U_CONCAT(x, y)
-#define U_CONCAT_STR(x, y) _U_CONCAT_STR(x, y)
-
-#endif /* _U_MACROS_ */
diff --git a/mesalib/src/mapi/u_thread.h b/mesalib/src/mapi/u_thread.h
deleted file mode 100644
index 57c3b076a..000000000
--- a/mesalib/src/mapi/u_thread.h
+++ /dev/null
@@ -1,156 +0,0 @@
-/*
- * Mesa 3-D graphics library
- *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-
-/*
- * Thread support for gl dispatch.
- *
- * Initial version by John Stone (j.stone@acm.org) (johns@cs.umr.edu)
- *                and Christoph Poliwoda (poliwoda@volumegraphics.com)
- * Revised by Keith Whitwell
- * Adapted for new gl dispatcher by Brian Paul
- * Modified for use in mapi by Chia-I Wu
- */
-
-/*
- * If this file is accidentally included by a non-threaded build,
- * it should not cause the build to fail, or otherwise cause problems.
- * In general, it should only be included when needed however.
- */
-
-#ifndef _U_THREAD_H_
-#define _U_THREAD_H_
-
-#include <stdio.h>
-#include <stdlib.h>
-#include "u_compiler.h"
-
-#include "c11/threads.h"
-
-#if defined(HAVE_PTHREAD) || defined(_WIN32)
-#ifndef THREADS
-#define THREADS
-#endif
-#endif
-
-/*
- * Error messages
- */
-#define INIT_TSD_ERROR "Mesa: failed to allocate key for thread specific data"
-#define GET_TSD_ERROR "Mesa: failed to get thread specific data"
-#define SET_TSD_ERROR "Mesa: thread failed to set thread specific data"
-
-
-/*
- * Magic number to determine if a TSD object has been initialized.
- * Kind of a hack but there doesn't appear to be a better cross-platform
- * solution.
- */
-#define INIT_MAGIC 0xff8adc98
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-
-struct u_tsd {
-   tss_t key;
-   unsigned initMagic;
-};
-
-
-static INLINE unsigned long
-u_thread_self(void)
-{
-   /*
-    * XXX: Callers of u_thread_self assume it is a lightweight function,
-    * returning a numeric value.  But unfortunately C11's thrd_current() gives
-    * no such guarantees.  In fact, it's pretty hard to have a compliant
-    * implementation of thrd_current() on Windows with such characteristics.
-    * So for now, we side-step this mess and use Windows thread primitives
-    * directly here.
-    *
-    * FIXME: On the other hand, u_thread_self() is a bad
-    * abstraction.  Even with pthreads, there is no guarantee that
-    * pthread_self() will return numeric IDs -- we should be using
-    * pthread_equal() instead of assuming we can compare thread ids...
-    */
-#ifdef _WIN32
-   return GetCurrentThreadId();
-#else
-   return (unsigned long) (uintptr_t) thrd_current();
-#endif
-}
-
-
-static INLINE void
-u_tsd_init(struct u_tsd *tsd)
-{
-   if (tss_create(&tsd->key, NULL/*free*/) != 0) {
-      perror(INIT_TSD_ERROR);
-      exit(-1);
-   }
-   tsd->initMagic = INIT_MAGIC;
-}
-
-
-static INLINE void *
-u_tsd_get(struct u_tsd *tsd)
-{
-   if (tsd->initMagic != INIT_MAGIC) {
-      u_tsd_init(tsd);
-   }
-   return tss_get(tsd->key);
-}
-
-
-static INLINE void
-u_tsd_set(struct u_tsd *tsd, void *ptr)
-{
-   if (tsd->initMagic != INIT_MAGIC) {
-      u_tsd_init(tsd);
-   }
-   if (tss_set(tsd->key, ptr) != 0) {
-      perror(SET_TSD_ERROR);
-      exit(-1);
-   }
-}
-
-
-static INLINE void
-u_tsd_destroy(struct u_tsd *tsd)
-{
-   if (tsd->initMagic != INIT_MAGIC) {
-      return;
-   }
-   tss_delete(tsd->key);
-   tsd->initMagic = 0x0;
-}
-
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif /* _U_THREAD_H_ */
diff --git a/mesalib/src/mesa/Android.gen.mk b/mesalib/src/mesa/Android.gen.mk
index c7b7f7edf..caae2c107 100644
--- a/mesalib/src/mesa/Android.gen.mk
+++ b/mesalib/src/mesa/Android.gen.mk
@@ -122,5 +122,5 @@ format_info_deps := \
 	$(LOCAL_PATH)/main/format_parser.py \
 	$(FORMAT_INFO)
 
-$(intermediates)/main/format_info.c: $(format_info_deps)
+$(intermediates)/main/format_info.h: $(format_info_deps)
 	@$(MESA_PYTHON2) $(FORMAT_INFO) $< > $@
diff --git a/mesalib/src/mesa/Makefile.am b/mesalib/src/mesa/Makefile.am
index b6cb8f111..3dab8f03f 100644
--- a/mesalib/src/mesa/Makefile.am
+++ b/mesalib/src/mesa/Makefile.am
@@ -62,7 +62,6 @@ include Makefile.sources
 EXTRA_DIST = \
 	drivers/haiku \
 	drivers/SConscript \
-	drivers/windows \
 	main/format_info.py \
 	main/format_pack.py \
 	main/format_parser.py \
@@ -80,7 +79,7 @@ EXTRA_DIST = \
 
 BUILT_SOURCES = \
 	main/get_hash.h \
-        main/format_info.c \
+	main/format_info.h \
 	main/git_sha1.h \
 	main/format_pack.c \
 	main/format_unpack.c \
@@ -94,13 +93,13 @@ CLEANFILES = \
 GET_HASH_GEN = main/get_hash_generator.py
 
 main/get_hash.h: ../mapi/glapi/gen/gl_and_es_API.xml main/get_hash_params.py 	\
-		 $(GET_HASH_GEN) Makefile
+		 $(GET_HASH_GEN)
 	$(AM_V_GEN)set -e;						\
 	$(PYTHON2) $(PYTHON_FLAGS) $(srcdir)/$(GET_HASH_GEN)		\
 		-f $< > $@.tmp;						\
 	mv $@.tmp $@;
 
-main/format_info.c: main/formats.csv                                    \
+main/format_info.h: main/formats.csv                                    \
                     main/format_parser.py main/format_info.py
 	$(AM_V_GEN)set -e;						\
 	$(PYTHON2) $(PYTHON_FLAGS) $(srcdir)/main/format_info.py        \
@@ -123,7 +122,7 @@ main/format_unpack.c: main/format_unpack.py main/formats.csv	\
 			$(srcdir)/main/formats.csv			\
 		| $(INDENT) $(INDENT_FLAGS) > $@;
 
-main/formats.c: main/format_info.c
+main/formats.c: main/format_info.h
 
 noinst_LTLIBRARIES = $(ARCH_LIBS)
 if NEED_LIBMESA
@@ -136,8 +135,14 @@ noinst_LTLIBRARIES += libmesagallium.la
 endif
 
 AM_CPPFLAGS = $(DEFINES) $(INCLUDE_DIRS)
-AM_CFLAGS = $(LLVM_CFLAGS) $(VISIBILITY_CFLAGS)
-AM_CXXFLAGS = $(LLVM_CFLAGS) $(VISIBILITY_CXXFLAGS)
+AM_CFLAGS = \
+	$(LLVM_CFLAGS) \
+	$(VISIBILITY_CFLAGS) \
+	$(MSVC2013_COMPAT_CFLAGS)
+AM_CXXFLAGS = \
+	$(LLVM_CFLAGS) \
+	$(VISIBILITY_CXXFLAGS) \
+	$(MSVC2013_COMPAT_CXXFLAGS)
 
 ARCH_LIBS =
 
diff --git a/mesalib/src/mesa/Makefile.sources b/mesalib/src/mesa/Makefile.sources
index 5b4e71253..217be9ab3 100644
--- a/mesalib/src/mesa/Makefile.sources
+++ b/mesalib/src/mesa/Makefile.sources
@@ -86,6 +86,7 @@ MAIN_FILES = \
 	main/ffvertex_prog.h \
 	main/fog.c \
 	main/fog.h \
+	main/format_info.h \
 	main/format_pack.h \
 	main/format_pack.c \
 	main/format_unpack.h \
@@ -100,6 +101,7 @@ MAIN_FILES = \
 	main/framebuffer.h \
 	main/get.c \
 	main/get.h \
+	main/get_hash.h \
 	main/genmipmap.c \
 	main/genmipmap.h \
 	main/getstring.c \
diff --git a/mesalib/src/mesa/SConscript b/mesalib/src/mesa/SConscript
index 62e81ced1..81939f981 100644
--- a/mesalib/src/mesa/SConscript
+++ b/mesalib/src/mesa/SConscript
@@ -10,6 +10,8 @@ from sys import executable as python_cmd
 
 env = env.Clone()
 
+env.MSVC2013Compat()
+
 env.Append(CPPPATH = [
     '#/src',
     '#/src/mapi',
@@ -62,7 +64,7 @@ get_hash_header = env.CodeGenerate(
 )
 
 format_info = env.CodeGenerate(
-      target = 'main/format_info.c',
+      target = 'main/format_info.h',
       script = 'main/format_info.py',
       source = 'main/formats.csv',
       command = python_cmd + ' $SCRIPT ' + ' $SOURCE > $TARGET'
diff --git a/mesalib/src/mesa/drivers/SConscript b/mesalib/src/mesa/drivers/SConscript
index 9888e3306..17813da96 100644
--- a/mesalib/src/mesa/drivers/SConscript
+++ b/mesalib/src/mesa/drivers/SConscript
@@ -11,8 +11,5 @@ if env['dri']:
         'dri/common/SConscript',
     ])
 
-if env['platform'] == 'windows':
-    SConscript('windows/gdi/SConscript')
-
 if env['platform'] == 'haiku':
     SConscript('haiku/swrast/SConscript')
diff --git a/mesalib/src/mesa/drivers/common/driverfuncs.h b/mesalib/src/mesa/drivers/common/driverfuncs.h
index 6b9a90074..385ccb8dd 100644
--- a/mesalib/src/mesa/drivers/common/driverfuncs.h
+++ b/mesalib/src/mesa/drivers/common/driverfuncs.h
@@ -26,6 +26,10 @@
 #ifndef DRIVERFUNCS_H
 #define DRIVERFUNCS_H
 
+#ifdef __cplusplus
+extern "C" {
+#endif
+
 extern void
 _mesa_init_driver_functions(struct dd_function_table *driver);
 
@@ -33,5 +37,8 @@ _mesa_init_driver_functions(struct dd_function_table *driver);
 extern void
 _mesa_init_driver_state(struct gl_context *ctx);
 
+#ifdef __cplusplus
+} // extern "C"
+#endif
 
 #endif
diff --git a/mesalib/src/mesa/drivers/common/meta.c b/mesalib/src/mesa/drivers/common/meta.c
index 3636ee83b..fdc4cf1e9 100644
--- a/mesalib/src/mesa/drivers/common/meta.c
+++ b/mesalib/src/mesa/drivers/common/meta.c
@@ -393,7 +393,7 @@ _mesa_meta_setup_vertex_objects(GLuint *VAO, GLuint *VBO,
 void
 _mesa_meta_init(struct gl_context *ctx)
 {
-   ASSERT(!ctx->Meta);
+   assert(!ctx->Meta);
 
    ctx->Meta = CALLOC_STRUCT(gl_meta_state);
 }
@@ -823,6 +823,8 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
 void
 _mesa_meta_end(struct gl_context *ctx)
 {
+   assert(ctx->Meta->SaveStackDepth > 0);
+
    struct save_state *save = &ctx->Meta->Save[ctx->Meta->SaveStackDepth - 1];
    const GLbitfield state = save->SavedState;
    int i;
@@ -1045,7 +1047,7 @@ _mesa_meta_end(struct gl_context *ctx)
    if (state & MESA_META_TEXTURE) {
       GLuint u, tgt;
 
-      ASSERT(ctx->Texture.CurrentUnit == 0);
+      assert(ctx->Texture.CurrentUnit == 0);
 
       /* restore texenv for unit[0] */
       _mesa_TexEnvi(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, save->EnvMode);
@@ -1332,8 +1334,8 @@ _mesa_meta_alloc_texture(struct temp_texture *tex,
 {
    GLboolean newTex = GL_FALSE;
 
-   ASSERT(width <= tex->MaxSize);
-   ASSERT(height <= tex->MaxSize);
+   assert(width <= tex->MaxSize);
+   assert(height <= tex->MaxSize);
 
    if (width > tex->Width ||
        height > tex->Height ||
@@ -1747,7 +1749,7 @@ meta_clear(struct gl_context *ctx, GLbitfield buffers, bool glsl)
          _mesa_ClampColor(GL_CLAMP_FRAGMENT_COLOR, GL_FALSE);
    }
    else {
-      ASSERT(metaSave & MESA_META_COLOR_MASK);
+      assert(metaSave & MESA_META_COLOR_MASK);
       _mesa_ColorMask(GL_FALSE, GL_FALSE, GL_FALSE, GL_FALSE);
    }
 
diff --git a/mesalib/src/mesa/drivers/common/meta_blit.c b/mesalib/src/mesa/drivers/common/meta_blit.c
index 3406be1ed..bb2164276 100644
--- a/mesalib/src/mesa/drivers/common/meta_blit.c
+++ b/mesalib/src/mesa/drivers/common/meta_blit.c
@@ -82,7 +82,7 @@ setup_glsl_msaa_blit_scaled_shader(struct gl_context *ctx,
    y_scale = samples * 0.5;
 
    /* We expect only power of 2 samples in source multisample buffer. */
-   assert((samples & (samples - 1)) == 0);
+   assert(samples > 0 && (samples & (samples - 1)) == 0);
    while (samples >> (shader_offset + 1)) {
       shader_offset++;
    }
@@ -133,6 +133,8 @@ setup_glsl_msaa_blit_scaled_shader(struct gl_context *ctx,
       sample_map = ctx->Const.SampleMap8x;
       break;
    default:
+      sample_number = NULL;
+      sample_map = NULL;
       _mesa_problem(ctx, "Unsupported sample count %d\n", samples);
       unreachable("Unsupported sample count");
    }
@@ -261,7 +263,7 @@ setup_glsl_msaa_blit_shader(struct gl_context *ctx,
    }
 
    /* We expect only power of 2 samples in source multisample buffer. */
-   assert((samples & (samples - 1)) == 0);
+   assert(samples > 0 && (samples & (samples - 1)) == 0);
    while (samples >> (shader_offset + 1)) {
       shader_offset++;
    }
diff --git a/mesalib/src/mesa/drivers/common/meta_tex_subimage.c b/mesalib/src/mesa/drivers/common/meta_tex_subimage.c
index 68c8273fe..e29addb73 100644
--- a/mesalib/src/mesa/drivers/common/meta_tex_subimage.c
+++ b/mesalib/src/mesa/drivers/common/meta_tex_subimage.c
@@ -44,7 +44,7 @@
 
 static struct gl_texture_image *
 create_texture_for_pbo(struct gl_context *ctx, bool create_pbo,
-                       GLenum pbo_target, int width, int height, int depth,
+                       GLenum pbo_target, int width, int height,
                        GLenum format, GLenum type, const void *pixels,
                        const struct gl_pixelstore_attrib *packing,
                        GLuint *tmp_pbo, GLuint *tmp_tex)
@@ -57,8 +57,7 @@ create_texture_for_pbo(struct gl_context *ctx, bool create_pbo,
    struct gl_texture_image *tex_image;
    bool read_only;
 
-   if ((packing->ImageHeight != 0 && packing->ImageHeight != height) ||
-       packing->SwapBytes ||
+   if (packing->SwapBytes ||
        packing->LsbFirst ||
        packing->Invert)
       return NULL;
@@ -79,6 +78,8 @@ create_texture_for_pbo(struct gl_context *ctx, bool create_pbo,
       *tmp_pbo = 0;
       buffer_obj = packing->BufferObj;
    } else {
+      bool is_pixel_pack = pbo_target == GL_PIXEL_PACK_BUFFER;
+
       assert(create_pbo);
 
       _mesa_GenBuffers(1, tmp_pbo);
@@ -89,9 +90,17 @@ create_texture_for_pbo(struct gl_context *ctx, bool create_pbo,
        */
       _mesa_BindBuffer(pbo_target, *tmp_pbo);
 
-      _mesa_BufferData(pbo_target, row_stride * height, pixels, GL_STREAM_DRAW);
+      /* In case of GL_PIXEL_PACK_BUFFER, pass null pointer for the pixel
+       * data to avoid unnecessary data copying in _mesa_BufferData().
+       */
+      if (is_pixel_pack)
+         _mesa_BufferData(pbo_target, row_stride * height, NULL,
+                          GL_STREAM_READ);
+      else
+         _mesa_BufferData(pbo_target, row_stride * height, pixels,
+                          GL_STREAM_DRAW);
 
-      buffer_obj = ctx->Unpack.BufferObj;
+      buffer_obj = packing->BufferObj;
       pixels = NULL;
 
       _mesa_BindBuffer(pbo_target, 0);
@@ -99,14 +108,16 @@ create_texture_for_pbo(struct gl_context *ctx, bool create_pbo,
 
    _mesa_GenTextures(1, tmp_tex);
    tex_obj = _mesa_lookup_texture(ctx, *tmp_tex);
-   tex_obj->Target = depth > 1 ? GL_TEXTURE_2D_ARRAY : GL_TEXTURE_2D;
-   tex_obj->Immutable = GL_TRUE;
    _mesa_initialize_texture_object(ctx, tex_obj, *tmp_tex, GL_TEXTURE_2D);
+   /* This must be set after _mesa_initialize_texture_object, not before. */
+   tex_obj->Immutable = GL_TRUE;
+   /* This is required for interactions with ARB_texture_view. */
+   tex_obj->NumLayers = 1;
 
    internal_format = _mesa_get_format_base_format(pbo_format);
 
    tex_image = _mesa_get_tex_image(ctx, tex_obj, tex_obj->Target, 0);
-   _mesa_init_teximage_fields(ctx, tex_image, width, height, depth,
+   _mesa_init_teximage_fields(ctx, tex_image, width, height, 1,
                               0, internal_format, pbo_format);
 
    read_only = pbo_target == GL_PIXEL_UNPACK_BUFFER;
@@ -133,6 +144,7 @@ _mesa_meta_pbo_TexSubImage(struct gl_context *ctx, GLuint dims,
                            const struct gl_pixelstore_attrib *packing)
 {
    GLuint pbo = 0, pbo_tex = 0, fbos[2] = { 0, 0 };
+   int full_height, image_height;
    struct gl_texture_image *pbo_tex_image;
    GLenum status;
    bool success = false;
@@ -166,9 +178,16 @@ _mesa_meta_pbo_TexSubImage(struct gl_context *ctx, GLuint dims,
       return true;
    }
 
+   /* For arrays, use a tall (height * depth) 2D texture but taking into
+    * account the inter-image padding specified with the image height packing
+    * property.
+    */
+   image_height = packing->ImageHeight == 0 ? height : packing->ImageHeight;
+   full_height = image_height * (depth - 1) + height;
+
    pbo_tex_image = create_texture_for_pbo(ctx, create_pbo,
                                           GL_PIXEL_UNPACK_BUFFER,
-                                          width, height, depth,
+                                          width, full_height,
                                           format, type, pixels, packing,
                                           &pbo, &pbo_tex);
    if (!pbo_tex_image)
@@ -177,8 +196,8 @@ _mesa_meta_pbo_TexSubImage(struct gl_context *ctx, GLuint dims,
    if (allocate_storage)
       ctx->Driver.AllocTextureImageBuffer(ctx, tex_image);
 
-   /* Only stash the current FBO */
-   _mesa_meta_begin(ctx, 0);
+   _mesa_meta_begin(ctx, ~(MESA_META_PIXEL_TRANSFER |
+                           MESA_META_PIXEL_STORE));
 
    _mesa_GenFramebuffers(2, fbos);
    _mesa_BindFramebuffer(GL_READ_FRAMEBUFFER, fbos[0]);
@@ -186,8 +205,12 @@ _mesa_meta_pbo_TexSubImage(struct gl_context *ctx, GLuint dims,
 
    if (tex_image->TexObject->Target == GL_TEXTURE_1D_ARRAY) {
       assert(depth == 1);
+      assert(zoffset == 0);
       depth = height;
       height = 1;
+      image_height = 1;
+      zoffset = yoffset;
+      yoffset = 0;
    }
 
    _mesa_meta_bind_fbo_image(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0,
@@ -214,15 +237,14 @@ _mesa_meta_pbo_TexSubImage(struct gl_context *ctx, GLuint dims,
       goto fail;
 
    for (z = 1; z < depth; z++) {
-      _mesa_meta_bind_fbo_image(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0,
-                                pbo_tex_image, z);
       _mesa_meta_bind_fbo_image(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0,
                                 tex_image, zoffset + z);
 
       _mesa_update_state(ctx);
 
       _mesa_meta_BlitFramebuffer(ctx, ctx->ReadBuffer, ctx->DrawBuffer,
-                                 0, 0, width, height,
+                                 0, z * image_height,
+                                 width, z * image_height + height,
                                  xoffset, yoffset,
                                  xoffset + width, yoffset + height,
                                  GL_COLOR_BUFFER_BIT, GL_NEAREST);
@@ -249,6 +271,7 @@ _mesa_meta_pbo_GetTexSubImage(struct gl_context *ctx, GLuint dims,
                               const struct gl_pixelstore_attrib *packing)
 {
    GLuint pbo = 0, pbo_tex = 0, fbos[2] = { 0, 0 };
+   int full_height, image_height;
    struct gl_texture_image *pbo_tex_image;
    GLenum status;
    bool success = false;
@@ -282,22 +305,33 @@ _mesa_meta_pbo_GetTexSubImage(struct gl_context *ctx, GLuint dims,
       return true;
    }
 
+   /* For arrays, use a tall (height * depth) 2D texture but taking into
+    * account the inter-image padding specified with the image height packing
+    * property.
+    */
+   image_height = packing->ImageHeight == 0 ? height : packing->ImageHeight;
+   full_height = image_height * (depth - 1) + height;
+
    pbo_tex_image = create_texture_for_pbo(ctx, false, GL_PIXEL_PACK_BUFFER,
-                                          width, height, depth,
+                                          width, full_height * depth,
                                           format, type, pixels, packing,
                                           &pbo, &pbo_tex);
    if (!pbo_tex_image)
       return false;
 
-   /* Only stash the current FBO */
-   _mesa_meta_begin(ctx, 0);
+   _mesa_meta_begin(ctx, ~(MESA_META_PIXEL_TRANSFER |
+                           MESA_META_PIXEL_STORE));
 
    _mesa_GenFramebuffers(2, fbos);
 
    if (tex_image && tex_image->TexObject->Target == GL_TEXTURE_1D_ARRAY) {
       assert(depth == 1);
+      assert(zoffset == 0);
       depth = height;
       height = 1;
+      image_height = 1;
+      zoffset = yoffset;
+      yoffset = 0;
    }
 
    /* If we were given a texture, bind it to the read framebuffer.  If not,
@@ -336,15 +370,14 @@ _mesa_meta_pbo_GetTexSubImage(struct gl_context *ctx, GLuint dims,
    for (z = 1; z < depth; z++) {
       _mesa_meta_bind_fbo_image(GL_READ_FRAMEBUFFER, GL_COLOR_ATTACHMENT0,
                                 tex_image, zoffset + z);
-      _mesa_meta_bind_fbo_image(GL_DRAW_FRAMEBUFFER, GL_COLOR_ATTACHMENT0,
-                                pbo_tex_image, z);
 
       _mesa_update_state(ctx);
 
       _mesa_meta_BlitFramebuffer(ctx, ctx->ReadBuffer, ctx->DrawBuffer,
                                  xoffset, yoffset,
                                  xoffset + width, yoffset + height,
-                                 0, 0, width, height,
+                                 0, z * image_height,
+                                 width, z * image_height + height,
                                  GL_COLOR_BUFFER_BIT, GL_NEAREST);
    }
 
diff --git a/mesalib/src/mesa/drivers/dri/common/utils.c b/mesalib/src/mesa/drivers/dri/common/utils.c
index f2e63c0b9..bb22107e9 100644
--- a/mesalib/src/mesa/drivers/dri/common/utils.c
+++ b/mesalib/src/mesa/drivers/dri/common/utils.c
@@ -30,6 +30,7 @@
  */
 
 #include <string.h>
+#include <stdio.h>
 #include <stdlib.h>
 #include <stdbool.h>
 #include <stdint.h>
@@ -485,6 +486,7 @@ driIndexConfigAttrib(const __DRIconfig *config, int index,
  * Currently only the following queries are supported by this function:
  *
  *     - \c __DRI2_RENDERER_VERSION
+ *     - \c __DRI2_RENDERER_PREFERRED_PROFILE
  *     - \c __DRI2_RENDERER_OPENGL_CORE_PROFILE_VERSION
  *     - \c __DRI2_RENDERER_OPENGL_COMPATIBLITY_PROFILE_VERSION
  *     - \c __DRI2_RENDERER_ES_PROFILE_VERSION
diff --git a/mesalib/src/mesa/drivers/dri/common/xmlconfig.c b/mesalib/src/mesa/drivers/dri/common/xmlconfig.c
index d69eb0196..2b284ccf2 100644
--- a/mesalib/src/mesa/drivers/dri/common/xmlconfig.c
+++ b/mesalib/src/mesa/drivers/dri/common/xmlconfig.c
@@ -27,10 +27,13 @@
  * \author Felix Kuehling
  */
 
+#include <stdarg.h>
+#include <stdio.h>
 #include <string.h>
 #include <assert.h>
 #include <expat.h>
 #include <fcntl.h>
+#include <math.h>
 #include <unistd.h>
 #include <errno.h>
 #include "main/imports.h"
diff --git a/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am b/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am
index da7d03480..555771685 100644
--- a/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am
+++ b/mesalib/src/mesa/drivers/dri/common/xmlpool/Makefile.am
@@ -52,7 +52,7 @@ POT=xmlpool.pot
 
 .PHONY: all clean pot po mo
 
-EXTRA_DIST = gen_xmlpool.py t_options.h $(POS) SConscript
+EXTRA_DIST = gen_xmlpool.py options.h t_options.h $(POS) SConscript
 BUILT_SOURCES = options.h
 CLEANFILES = $(MOS) options.h
 
diff --git a/mesalib/src/mesa/drivers/dri/common/xmlpool/ca.po b/mesalib/src/mesa/drivers/dri/common/xmlpool/ca.po
index 23e9f4211..03bf29613 100644
--- a/mesalib/src/mesa/drivers/dri/common/xmlpool/ca.po
+++ b/mesalib/src/mesa/drivers/dri/common/xmlpool/ca.po
@@ -25,15 +25,15 @@ msgid ""
 msgstr ""
 "Project-Id-Version: Mesa 10.1.0-devel\n"
 "Report-Msgid-Bugs-To: \n"
-"POT-Creation-Date: 2014-09-25 22:29-0600\n"
-"PO-Revision-Date: 2014-09-26 14:43-0700\n"
+"POT-Creation-Date: 2015-02-07 02:08-0700\n"
+"PO-Revision-Date: 2015-02-23 14:28-0700\n"
 "Last-Translator: Alex Henrie <alexhenrie24@gmail.com>\n"
 "Language-Team: Catalan <ca@li.org>\n"
 "Language: ca\n"
 "MIME-Version: 1.0\n"
 "Content-Type: text/plain; charset=UTF-8\n"
 "Content-Transfer-Encoding: 8bit\n"
-"X-Generator: Poedit 1.6.9\n"
+"X-Generator: Poedit 1.7.4\n"
 
 #: t_options.h:56
 msgid "Debugging"
@@ -83,8 +83,8 @@ msgstr "Deshabilita el GL_ARB_shader_bit_encoding"
 msgid ""
 "Force a default GLSL version for shaders that lack an explicit #version line"
 msgstr ""
-"Força una versió GLSL per defecte en els shaders als quals falta una línia "
-"#version explícita"
+"Força una versió GLSL per defecte en els shaders als quals lis manca una "
+"línia #version explícita"
 
 #: t_options.h:110
 msgid "Allow GLSL #extension directives in the middle of shaders"
@@ -92,7 +92,7 @@ msgstr "Permet les directives #extension GLSL en el mitjà dels shaders"
 
 #: t_options.h:120
 msgid "Image Quality"
-msgstr "Qualitat d'Imatge"
+msgstr "Qualitat d'imatge"
 
 #: t_options.h:133
 msgid "Texture color depth"
@@ -149,11 +149,11 @@ msgstr "Mètode d'arrodoniment de color"
 
 #: t_options.h:171
 msgid "Round color components downward"
-msgstr "Arrondeix els components de color a baix"
+msgstr "Arrodoneix els components de color a baix"
 
 #: t_options.h:172
 msgid "Round to nearest color"
-msgstr "Arrondeix al color més proper"
+msgstr "Arrodoneix al color més proper"
 
 #: t_options.h:181
 msgid "Color dithering method"
@@ -181,15 +181,15 @@ msgstr "Un filtre de postprocessament per a aplicar cel shading a la sortida"
 
 #: t_options.h:200
 msgid "A post-processing filter to remove the red channel"
-msgstr "Un filtre de postprocessament per a treure el canal vermell"
+msgstr "Un filtre de postprocessament per a eliminar el canal vermell"
 
 #: t_options.h:205
 msgid "A post-processing filter to remove the green channel"
-msgstr "Un filtre de postprocessament per a treure el canal verd"
+msgstr "Un filtre de postprocessament per a eliminar el canal verd"
 
 #: t_options.h:210
 msgid "A post-processing filter to remove the blue channel"
-msgstr "Un filtre de postprocessament per a treure el canal blau"
+msgstr "Un filtre de postprocessament per a eliminar el canal blau"
 
 #: t_options.h:215
 msgid ""
@@ -222,17 +222,17 @@ msgstr "Utilitza la canonada TCL de programari"
 
 #: t_options.h:240
 msgid "Use hardware TCL as first TCL pipeline stage"
-msgstr "Utilitza el TCL maquinàri com la primera fase de la canonada TCL"
+msgstr "Utilitza el TCL de maquinari com a la primera fase de la canonada TCL"
 
 #: t_options.h:241
 msgid "Bypass the TCL pipeline"
-msgstr "Passar per alt la canonada TCL"
+msgstr "Passa per alt la canonada TCL"
 
 #: t_options.h:242
 msgid ""
 "Bypass the TCL pipeline with state-based machine code generated on-the-fly"
 msgstr ""
-"Passar per alt la canonada TCL amb codi màquina basat en estats, generat "
+"Passa per alt la canonada TCL amb codi de màquina basat en estats, generat "
 "sobre la marxa"
 
 #: t_options.h:251
@@ -245,7 +245,7 @@ msgstr "Espera activa pel maquinari de gràfics"
 
 #: t_options.h:253
 msgid "Sleep for brief intervals while waiting for the graphics hardware"
-msgstr "Dormi per intervals breus mentre s'espera al maquinari de gràfics"
+msgstr "Dorm per intervals breus mentre s'espera al maquinari de gràfics"
 
 #: t_options.h:254
 msgid "Let the graphics hardware emit a software interrupt and sleep"
@@ -259,7 +259,7 @@ msgstr "Sincronització amb refresc vertical (intervals d'intercanvi)"
 #: t_options.h:265
 msgid "Never synchronize with vertical refresh, ignore application's choice"
 msgstr ""
-"Mai sincronitzis amb el refresc vertial, ignora l'elecció de l'aplicació"
+"Mai sincronitzis amb el refresc vertical, ignora l'elecció de l'aplicació"
 
 #: t_options.h:266
 msgid "Initial swap interval 0, obey application's choice"
@@ -301,7 +301,7 @@ msgstr "Tota la memòria disponible"
 
 #: t_options.h:296
 msgid "Only card memory (if available)"
-msgstr "Només memòria de tarjeta (si està disponible)"
+msgstr "Només memòria de targeta (si està disponible)"
 
 #: t_options.h:297
 msgid "Only GART (AGP/PCIE) memory (if available)"
@@ -329,4 +329,20 @@ msgstr "Inicialització"
 
 #: t_options.h:341
 msgid "Define the graphic device to use if possible"
-msgstr "Defineix el dispositiu de gràfics que usar si és possible"
+msgstr "Defineix el dispositiu de gràfics que utilitzar si és possible"
+
+#: t_options.h:350
+msgid "Gallium Nine"
+msgstr "Gallium Nine"
+
+#: t_options.h:354
+msgid ""
+"Define the throttling value. -1 for no throttling, -2 for default (usually "
+"2), 0 for glfinish behaviour"
+msgstr ""
+"Defineix el valor de regulació. -1 per a no regular, -2 per al predeterminat "
+"(generalment 2), 0 per al comportament de glfinish"
+
+#: t_options.h:359
+msgid "Use an additional thread to submit buffers."
+msgstr "Utilitza un fil addicional per a entregar els buffers."
diff --git a/mesalib/src/mesa/drivers/dri/common/xmlpool/es.po b/mesalib/src/mesa/drivers/dri/common/xmlpool/es.po
index a68c329a5..f9d950ac1 100644
--- a/mesalib/src/mesa/drivers/dri/common/xmlpool/es.po
+++ b/mesalib/src/mesa/drivers/dri/common/xmlpool/es.po
@@ -9,8 +9,8 @@ msgid ""
 msgstr ""
 "Project-Id-Version: es\n"
 "Report-Msgid-Bugs-To: \n"
-"POT-Creation-Date: 2014-09-25 22:29-0600\n"
-"PO-Revision-Date: 2014-09-26 14:22-0700\n"
+"POT-Creation-Date: 2015-02-07 02:08-0700\n"
+"PO-Revision-Date: 2015-02-23 14:54-0700\n"
 "Last-Translator: Alex Henrie <alexhenrie24@gmail.com>\n"
 "Language-Team: Spanish <es@li.org>\n"
 "Language: es\n"
@@ -18,7 +18,7 @@ msgstr ""
 "Content-Type: text/plain; charset=UTF-8\n"
 "Content-Transfer-Encoding: 8bit\n"
 "Plural-Forms: nplurals=2; plural=(n != 1);\n"
-"X-Generator: Poedit 1.6.9\n"
+"X-Generator: Poedit 1.7.4\n"
 
 #: t_options.h:56
 msgid "Debugging"
@@ -301,11 +301,11 @@ msgstr "Habilitar la extensión GL_ARB_vertex_program"
 
 #: t_options.h:323
 msgid "Miscellaneous"
-msgstr "Misceláneo"
+msgstr "Misceláneos"
 
 #: t_options.h:327
 msgid "Create all visuals with a depth buffer"
-msgstr "Crear todos los visuales con buffer de profundidad"
+msgstr "Crear todos los visuales con búfer de profundidad"
 
 #: t_options.h:337
 msgid "Initialization"
@@ -314,3 +314,19 @@ msgstr "Inicialización"
 #: t_options.h:341
 msgid "Define the graphic device to use if possible"
 msgstr "Define el dispositivo de gráficos que usar si es posible"
+
+#: t_options.h:350
+msgid "Gallium Nine"
+msgstr "Gallium Nine"
+
+#: t_options.h:354
+msgid ""
+"Define the throttling value. -1 for no throttling, -2 for default (usually "
+"2), 0 for glfinish behaviour"
+msgstr ""
+"Define el valor de regulación. -1 para no regular, -2 para el por defecto "
+"(generalmente 2), 0 para el comportamiento de glfinish"
+
+#: t_options.h:359
+msgid "Use an additional thread to submit buffers."
+msgstr "Usar un hilo adicional para entregar los búfer."
diff --git a/mesalib/src/mesa/drivers/dri/swrast/swrast.c b/mesalib/src/mesa/drivers/dri/swrast/swrast.c
index 8005f7d69..fb29078b6 100644
--- a/mesalib/src/mesa/drivers/dri/swrast/swrast.c
+++ b/mesalib/src/mesa/drivers/dri/swrast/swrast.c
@@ -32,6 +32,7 @@
  * The back-buffer is allocated by the driver and is private.
  */
 
+#include <stdio.h>
 #include "main/api_exec.h"
 #include "main/context.h"
 #include "main/extensions.h"
@@ -341,7 +342,7 @@ swrast_delete_renderbuffer(struct gl_context *ctx, struct gl_renderbuffer *rb)
 }
 
 /* see bytes_per_line in libGL */
-static INLINE int
+static inline int
 bytes_per_line(unsigned pitch_bits, unsigned mul)
 {
    unsigned mask = mul - 1;
@@ -481,7 +482,7 @@ swrast_map_renderbuffer(struct gl_context *ctx,
       return;
    }
 
-   ASSERT(xrb->Base.Buffer);
+   assert(xrb->Base.Buffer);
 
    if (rb->AllocStorage == swrast_alloc_back_storage) {
       map += (rb->Height - 1) * stride;
diff --git a/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h b/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h
index 535f6fe59..1f3a48f38 100644
--- a/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h
+++ b/mesalib/src/mesa/drivers/dri/swrast/swrast_priv.h
@@ -65,13 +65,13 @@ struct dri_context
     __DRIcontext *cPriv;
 };
 
-static INLINE struct dri_context *
+static inline struct dri_context *
 dri_context(__DRIcontext * driContextPriv)
 {
     return (struct dri_context *)driContextPriv->driverPrivate;
 }
 
-static INLINE struct dri_context *
+static inline struct dri_context *
 swrast_context(struct gl_context *ctx)
 {
     return (struct dri_context *) ctx;
@@ -89,13 +89,13 @@ struct dri_drawable
     char *row;
 };
 
-static INLINE struct dri_drawable *
+static inline struct dri_drawable *
 dri_drawable(__DRIdrawable * driDrawPriv)
 {
     return (struct dri_drawable *)driDrawPriv->driverPrivate;
 }
 
-static INLINE struct dri_drawable *
+static inline struct dri_drawable *
 swrast_drawable(struct gl_framebuffer *fb)
 {
     return (struct dri_drawable *) fb;
@@ -115,7 +115,7 @@ struct dri_swrast_renderbuffer {
     GLuint bpp;
 };
 
-static INLINE struct dri_swrast_renderbuffer *
+static inline struct dri_swrast_renderbuffer *
 dri_swrast_renderbuffer(struct gl_renderbuffer *rb)
 {
     return (struct dri_swrast_renderbuffer *) rb;
diff --git a/mesalib/src/mesa/drivers/windows/gdi/SConscript b/mesalib/src/mesa/drivers/windows/gdi/SConscript
deleted file mode 100644
index 1f4d7e1fa..000000000
--- a/mesalib/src/mesa/drivers/windows/gdi/SConscript
+++ /dev/null
@@ -1,40 +0,0 @@
-Import('*')
-
-env = env.Clone()
-
-env.Prepend(CPPPATH = [
-    '#src',
-    '#src/mapi',
-    '#src/mesa',
-])
-
-env.AppendUnique(CPPDEFINES = [
-    '_GDI32_', # prevent wgl* being declared __declspec(dllimport)
-    'BUILD_GL32', # declare gl* as __declspec(dllexport) in Mesa headers 
-])
-if not env['gles']:
-    # prevent _glapi_* from being declared __declspec(dllimport)
-    env.Append(CPPDEFINES = ['_GLAPI_NO_EXPORTS'])
-
-env.Prepend(LIBS = [
-    mesautil,
-    glapi,
-    mesa,
-    glsl,
-    'gdi32',
-    'user32',
-    'kernel32',
-])
-
-sources = [
-    'mesa.def',
-    'wgl.c',
-    'wmesa.c',
-]
-
-mesagdi = env.SharedLibrary(
-    target = 'opengl32',
-    source = sources,
-)
-
-env.Alias('mesagdi', mesagdi)
diff --git a/mesalib/src/mesa/drivers/windows/gdi/colors.h b/mesalib/src/mesa/drivers/windows/gdi/colors.h
deleted file mode 100644
index 03e512c1f..000000000
--- a/mesalib/src/mesa/drivers/windows/gdi/colors.h
+++ /dev/null
@@ -1,29 +0,0 @@
-/* Values for wmesa->pixelformat: */
-
-#define PF_8A8B8G8R	3	/* 32-bit TrueColor:  8-A, 8-B, 8-G, 8-R */
-#define PF_8R8G8B	4	/* 32-bit TrueColor:  8-R, 8-G, 8-B */
-#define PF_5R6G5B	5	/* 16-bit TrueColor:  5-R, 6-G, 5-B bits */
-#define PF_DITHER8	6	/* Dithered RGB using a lookup table */
-#define PF_LOOKUP	7	/* Undithered RGB using a lookup table */
-#define PF_GRAYSCALE	10	/* Grayscale or StaticGray */
-#define PF_BADFORMAT	11
-#define PF_INDEX8	12
-
-
-#define BGR8(r,g,b) (unsigned)(((BYTE)((b & 0xc0) | ((g & 0xe0)>>2) | \
-                                      ((r & 0xe0)>>5))))
-
-/* Windows uses 5,5,5 for 16-bit */
-#define BGR16(r,g,b) (  (((unsigned short)b       ) >> 3) | \
-                        (((unsigned short)g & 0xf8) << 2) | \
-                        (((unsigned short)r & 0xf8) << 7) )
-
-#define BGR24(r,g,b) (unsigned long)((DWORD)(((BYTE)(b)| \
-                                    ((WORD)((BYTE)(g))<<8))| \
-                                    (((DWORD)(BYTE)(r))<<16)))
-
-#define BGR32(r,g,b) (unsigned long)((DWORD)(((BYTE)(b)| \
-                                    ((WORD)((BYTE)(g))<<8))| \
-                                    (((DWORD)(BYTE)(r))<<16)))
-
-
diff --git a/mesalib/src/mesa/drivers/windows/gdi/mesa.def b/mesalib/src/mesa/drivers/windows/gdi/mesa.def
deleted file mode 100644
index a52207b63..000000000
--- a/mesalib/src/mesa/drivers/windows/gdi/mesa.def
+++ /dev/null
@@ -1,385 +0,0 @@
-;
-; Module definition file for Mesa (OPENGL32.DLL)
-;
-; Note: The OpenGL functions use the STDCALL
-; function calling convention.  Microsoft's
-; OPENGL32 uses this convention and so must the
-; Mesa OPENGL32 so that the Mesa DLL can be used
-; as a drop-in replacement.
-;
-; The linker exports STDCALL entry points with
-; 'decorated' names; e.g., _glBegin@0, where the
-; trailing number is the number of bytes of 
-; parameter data pushed onto the stack.  The
-; callee is responsible for popping this data
-; off the stack, usually via a RETF n instruction.
-;
-; However, the Microsoft OPENGL32.DLL does not export
-; the decorated names, even though the calling convention
-; is STDCALL.  So, this module definition file is
-; needed to force the Mesa OPENGL32.DLL to export the
-; symbols in the same manner as the Microsoft DLL.
-; Were it not for this problem, this file would not
-; be needed (for the gl* functions) since the entry
-; points are compiled with dllexport declspec.
-;
-EXPORTS
-	glAccum
-	glAlphaFunc
-	glAreTexturesResident
-	glArrayElement
-	glBegin
-	glBindTexture
-	glBitmap
-	glBlendFunc
-	glCallList
-	glCallLists
-	glClear
-	glClearAccum
-	glClearColor
-	glClearDepth
-	glClearIndex
-	glClearStencil
-	glClipPlane
-	glColor3b
-	glColor3bv
-	glColor3d
-	glColor3dv
-	glColor3f
-	glColor3fv
-	glColor3i
-	glColor3iv
-	glColor3s
-	glColor3sv
-	glColor3ub
-	glColor3ubv
-	glColor3ui
-	glColor3uiv
-	glColor3us
-	glColor3usv
-	glColor4b
-	glColor4bv
-	glColor4d
-	glColor4dv
-	glColor4f
-	glColor4fv
-	glColor4i
-	glColor4iv
-	glColor4s
-	glColor4sv
-	glColor4ub
-	glColor4ubv
-	glColor4ui
-	glColor4uiv
-	glColor4us
-	glColor4usv
-	glColorMask
-	glColorMaterial
-	glColorPointer
-	glCopyPixels
-	glCopyTexImage1D
-	glCopyTexImage2D
-	glCopyTexSubImage1D
-	glCopyTexSubImage2D
-	glCullFace
-	glDeleteLists
-	glDeleteTextures
-	glDepthFunc
-	glDepthMask
-	glDepthRange
-	glDisable
-	glDisableClientState
-	glDrawArrays
-	glDrawBuffer
-	glDrawElements
-	glDrawPixels
-	glEdgeFlag
-	glEdgeFlagPointer
-	glEdgeFlagv
-	glEnable
-	glEnableClientState
-	glEnd
-	glEndList
-	glEvalCoord1d
-	glEvalCoord1dv
-	glEvalCoord1f
-	glEvalCoord1fv
-	glEvalCoord2d
-	glEvalCoord2dv
-	glEvalCoord2f
-	glEvalCoord2fv
-	glEvalMesh1
-	glEvalMesh2
-	glEvalPoint1
-	glEvalPoint2
-	glFeedbackBuffer
-	glFinish
-	glFlush
-	glFogf
-	glFogfv
-	glFogi
-	glFogiv
-	glFrontFace
-	glFrustum
-	glGenLists
-	glGenTextures
-	glGetBooleanv
-	glGetClipPlane
-	glGetDoublev
-	glGetError
-	glGetFloatv
-	glGetIntegerv
-	glGetLightfv
-	glGetLightiv
-	glGetMapdv
-	glGetMapfv
-	glGetMapiv
-	glGetMaterialfv
-	glGetMaterialiv
-	glGetPixelMapfv
-	glGetPixelMapuiv
-	glGetPixelMapusv
-	glGetPointerv
-	glGetPolygonStipple
-	glGetString
-	glGetTexEnvfv
-	glGetTexEnviv
-	glGetTexGendv
-	glGetTexGenfv
-	glGetTexGeniv
-	glGetTexImage
-	glGetTexLevelParameterfv
-	glGetTexLevelParameteriv
-	glGetTexParameterfv
-	glGetTexParameteriv
-	glHint
-	glIndexMask
-	glIndexPointer
-	glIndexd
-	glIndexdv
-	glIndexf
-	glIndexfv
-	glIndexi
-	glIndexiv
-	glIndexs
-	glIndexsv
-	glIndexub
-	glIndexubv
-	glInitNames
-	glInterleavedArrays
-	glIsEnabled
-	glIsList
-	glIsTexture
-	glLightModelf
-	glLightModelfv
-	glLightModeli
-	glLightModeliv
-	glLightf
-	glLightfv
-	glLighti
-	glLightiv
-	glLineStipple
-	glLineWidth
-	glListBase
-	glLoadIdentity
-	glLoadMatrixd
-	glLoadMatrixf
-	glLoadName
-	glLogicOp
-	glMap1d
-	glMap1f
-	glMap2d
-	glMap2f
-	glMapGrid1d
-	glMapGrid1f
-	glMapGrid2d
-	glMapGrid2f
-	glMaterialf
-	glMaterialfv
-	glMateriali
-	glMaterialiv
-	glMatrixMode
-	glMultMatrixd
-	glMultMatrixf
-	glNewList
-	glNormal3b
-	glNormal3bv
-	glNormal3d
-	glNormal3dv
-	glNormal3f
-	glNormal3fv
-	glNormal3i
-	glNormal3iv
-	glNormal3s
-	glNormal3sv
-	glNormalPointer
-	glOrtho
-	glPassThrough
-	glPixelMapfv
-	glPixelMapuiv
-	glPixelMapusv
-	glPixelStoref
-	glPixelStorei
-	glPixelTransferf
-	glPixelTransferi
-	glPixelZoom
-	glPointSize
-	glPolygonMode
-	glPolygonOffset
-	glPolygonStipple
-	glPopAttrib
-	glPopClientAttrib
-	glPopMatrix
-	glPopName
-	glPrioritizeTextures
-	glPushAttrib
-	glPushClientAttrib
-	glPushMatrix
-	glPushName
-	glRasterPos2d
-	glRasterPos2dv
-	glRasterPos2f
-	glRasterPos2fv
-	glRasterPos2i
-	glRasterPos2iv
-	glRasterPos2s
-	glRasterPos2sv
-	glRasterPos3d
-	glRasterPos3dv
-	glRasterPos3f
-	glRasterPos3fv
-	glRasterPos3i
-	glRasterPos3iv
-	glRasterPos3s
-	glRasterPos3sv
-	glRasterPos4d
-	glRasterPos4dv
-	glRasterPos4f
-	glRasterPos4fv
-	glRasterPos4i
-	glRasterPos4iv
-	glRasterPos4s
-	glRasterPos4sv
-	glReadBuffer
-	glReadPixels
-	glRectd
-	glRectdv
-	glRectf
-	glRectfv
-	glRecti
-	glRectiv
-	glRects
-	glRectsv
-	glRenderMode
-	glRotated
-	glRotatef
-	glScaled
-	glScalef
-	glScissor
-	glSelectBuffer
-	glShadeModel
-	glStencilFunc
-	glStencilMask
-	glStencilOp
-	glTexCoord1d
-	glTexCoord1dv
-	glTexCoord1f
-	glTexCoord1fv
-	glTexCoord1i
-	glTexCoord1iv
-	glTexCoord1s
-	glTexCoord1sv
-	glTexCoord2d
-	glTexCoord2dv
-	glTexCoord2f
-	glTexCoord2fv
-	glTexCoord2i
-	glTexCoord2iv
-	glTexCoord2s
-	glTexCoord2sv
-	glTexCoord3d
-	glTexCoord3dv
-	glTexCoord3f
-	glTexCoord3fv
-	glTexCoord3i
-	glTexCoord3iv
-	glTexCoord3s
-	glTexCoord3sv
-	glTexCoord4d
-	glTexCoord4dv
-	glTexCoord4f
-	glTexCoord4fv
-	glTexCoord4i
-	glTexCoord4iv
-	glTexCoord4s
-	glTexCoord4sv
-	glTexCoordPointer
-	glTexEnvf
-	glTexEnvfv
-	glTexEnvi
-	glTexEnviv
-	glTexGend
-	glTexGendv
-	glTexGenf
-	glTexGenfv
-	glTexGeni
-	glTexGeniv
-	glTexImage1D
-	glTexImage2D
-	glTexParameterf
-	glTexParameterfv
-	glTexParameteri
-	glTexParameteriv
-	glTexSubImage1D
-	glTexSubImage2D
-	glTranslated
-	glTranslatef
-	glVertex2d
-	glVertex2dv
-	glVertex2f
-	glVertex2fv
-	glVertex2i
-	glVertex2iv
-	glVertex2s
-	glVertex2sv
-	glVertex3d
-	glVertex3dv
-	glVertex3f
-	glVertex3fv
-	glVertex3i
-	glVertex3iv
-	glVertex3s
-	glVertex3sv
-	glVertex4d
-	glVertex4dv
-	glVertex4f
-	glVertex4fv
-	glVertex4i
-	glVertex4iv
-	glVertex4s
-	glVertex4sv
-	glVertexPointer
-	glViewport
-	wglChoosePixelFormat
-	wglCopyContext
-	wglCreateContext
-	wglCreateLayerContext
-	wglDeleteContext
-	wglDescribeLayerPlane
-	wglDescribePixelFormat
-	wglGetCurrentContext
-	wglGetCurrentDC
-	wglGetLayerPaletteEntries
-	wglGetPixelFormat
-	wglGetProcAddress
-	wglMakeCurrent
-	wglRealizeLayerPalette
-	wglSetLayerPaletteEntries
-	wglSetPixelFormat
-	wglShareLists
-	wglSwapBuffers
-	wglSwapLayerBuffers
-	wglUseFontBitmapsA
-	wglUseFontBitmapsW
-	wglUseFontOutlinesA
-	wglUseFontOutlinesW
diff --git a/mesalib/src/mesa/drivers/windows/gdi/wgl.c b/mesalib/src/mesa/drivers/windows/gdi/wgl.c
deleted file mode 100644
index 3458cc9f3..000000000
--- a/mesalib/src/mesa/drivers/windows/gdi/wgl.c
+++ /dev/null
@@ -1,689 +0,0 @@
-
-/*
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Library General Public
- * License as published by the Free Software Foundation; either
- * version 2 of the License, or (at your option) any later version.
- *
- * This library is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * Library General Public License for more details.
- *
- * You should have received a copy of the GNU Library General Public
- * License along with this library; if not, write to the Free
- * Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
- *
- */
-
-/*
- * File name 	: wgl.c
- * WGL stuff. Added by Oleg Letsinsky, ajl@ultersys.ru
- * Some things originated from the 3Dfx WGL functions
- */
-
-/* 
- * This file contains the implementation of the wgl* functions for
- * Mesa on Windows.  Since these functions are provided by Windows in
- * GDI/OpenGL, we must supply our versions that work with Mesa here.
- */
-
-
-#include <windows.h>
-
-#include "main/config.h"
-#include "glapi/glapi.h"
-#include "swrast/swrast.h"
-#include "GL/wmesa.h"   /* protos for wmesa* functions */
-
-/*
- * Pixel Format Descriptors
- */
-
-/* Extend the PFD to include DB flag */
-struct __pixelformat__
-{
-    PIXELFORMATDESCRIPTOR pfd;
-    GLboolean doubleBuffered;
-};
-
-
-
-/* These are the PFD's supported by this driver. */
-struct __pixelformat__	pfd[] =
-{
-#if 0 
-    /* Double Buffer, alpha */
-    {	
-	{	
-	    sizeof(PIXELFORMATDESCRIPTOR),	1,
-	    PFD_DRAW_TO_WINDOW|PFD_SUPPORT_OPENGL|
-	    PFD_GENERIC_FORMAT|PFD_DOUBLEBUFFER|PFD_SWAP_COPY,
-	    PFD_TYPE_RGBA,
-	    24,	
-	    8, 0,	
-	    8, 8,	
-	    8, 16,	
-	    8, 24,
-	    0, 0, 0, 0, 0,	
-	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
-	    0, 0, 0,	
-	    0, 0, 0 
-	},
-        GL_TRUE
-    },
-    /* Single Buffer, alpha */
-    {	
-	{	
-	    sizeof(PIXELFORMATDESCRIPTOR),	1,
-	    PFD_DRAW_TO_WINDOW|PFD_SUPPORT_OPENGL|
-	    PFD_GENERIC_FORMAT,
-	    PFD_TYPE_RGBA,
-	    24,	
-	    8, 0,	
-	    8, 8,	
-	    8, 16,	
-	    8, 24,
-	    0, 0, 0, 0,	0,	
-	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
-	    0, 0, 0,	
-	    0, 0, 0
-	},
-        GL_FALSE
-    },
-#endif 
-    /* Double Buffer, no alpha */
-    {	
-	{	
-	    sizeof(PIXELFORMATDESCRIPTOR),	1,
-	    PFD_DRAW_TO_WINDOW|PFD_SUPPORT_OPENGL|
-	    PFD_GENERIC_FORMAT|PFD_DOUBLEBUFFER|PFD_SWAP_COPY,
-	    PFD_TYPE_RGBA,
-	    24,	
-	    8, 0,
-	    8, 8,
-	    8, 16,
-	    0, 0,
-	    0, 0, 0, 0,	0,
-	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
-	    0, 0, 0, 
-	    0, 0, 0 
-	},
-        GL_TRUE
-    },
-    /* Single Buffer, no alpha */
-    {	
-	{
-	    sizeof(PIXELFORMATDESCRIPTOR),	1,
-	    PFD_DRAW_TO_WINDOW|PFD_SUPPORT_OPENGL|
-	    PFD_GENERIC_FORMAT,
-	    PFD_TYPE_RGBA,
-	    24,	
-	    8, 0,
-	    8, 8,
-	    8, 16,
-	    0, 0,
-	    0, 0, 0, 0,	0,
-	    DEFAULT_SOFTWARE_DEPTH_BITS,	8,	
-	    0, 0, 0,
-	    0, 0, 0 
-	},
-        GL_FALSE
-    },
-};
-
-int npfd = sizeof(pfd) / sizeof(pfd[0]);
-
-
-/*
- * Contexts
- */
-
-typedef struct {
-    WMesaContext ctx;
-} MesaWglCtx;
-
-#define MESAWGL_CTX_MAX_COUNT 20
-
-static MesaWglCtx wgl_ctx[MESAWGL_CTX_MAX_COUNT];
-
-static unsigned ctx_count = 0;
-static int ctx_current = -1;
-static unsigned curPFD = 0;
-
-static HDC CurrentHDC = 0;
-
-
-WINGDIAPI HGLRC GLAPIENTRY wglCreateContext(HDC hdc)
-{
-    int i = 0;
-    if (!ctx_count) {
-	for(i=0;i<MESAWGL_CTX_MAX_COUNT;i++) {
-	    wgl_ctx[i].ctx = NULL;
-	}
-    }
-    for( i = 0; i < MESAWGL_CTX_MAX_COUNT; i++ ) {
-        if ( wgl_ctx[i].ctx == NULL ) {
-            wgl_ctx[i].ctx = 
-		WMesaCreateContext(hdc, NULL, (GLboolean)GL_TRUE,
-				   (GLboolean) (pfd[curPFD-1].doubleBuffered ?
-                                   GL_TRUE : GL_FALSE), 
-				   (GLboolean)(pfd[curPFD-1].pfd.cAlphaBits ? 
-				   GL_TRUE : GL_FALSE) );
-            if (wgl_ctx[i].ctx == NULL)
-                break;
-            ctx_count++;
-            return ((HGLRC)wgl_ctx[i].ctx);
-        }
-    }
-    SetLastError(0);
-    return(NULL);
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglDeleteContext(HGLRC hglrc)
-{
-    int i;
-    for ( i = 0; i < MESAWGL_CTX_MAX_COUNT; i++ ) {
-    	if ( wgl_ctx[i].ctx == (WMesaContext) hglrc ){
-            WMesaMakeCurrent((WMesaContext) hglrc, NULL);
-            WMesaDestroyContext(wgl_ctx[i].ctx);
-            wgl_ctx[i].ctx = NULL;
-            ctx_count--;
-            return(TRUE);
-    	}
-    }
-    SetLastError(0);
-    return(FALSE);
-}
-
-WINGDIAPI HGLRC GLAPIENTRY wglGetCurrentContext(VOID)
-{
-    if (ctx_current < 0)
-	return 0;
-    else
-	return (HGLRC) wgl_ctx[ctx_current].ctx;
-}
-
-WINGDIAPI HDC GLAPIENTRY wglGetCurrentDC(VOID)
-{
-    return CurrentHDC;
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglMakeCurrent(HDC hdc, HGLRC hglrc)
-{
-    int i;
-    
-    CurrentHDC = hdc;
-
-    if (!hdc || !hglrc) {
-	WMesaMakeCurrent(NULL, NULL);
-	ctx_current = -1;
-	return TRUE;
-    }
-    
-    for ( i = 0; i < MESAWGL_CTX_MAX_COUNT; i++ ) {
-	if ( wgl_ctx[i].ctx == (WMesaContext) hglrc ) {
-	    WMesaMakeCurrent( (WMesaContext) hglrc, hdc );
-	    ctx_current = i;
-	    return TRUE;
-	}
-    }
-    return FALSE;
-}
-
-
-WINGDIAPI int GLAPIENTRY wglChoosePixelFormat(HDC hdc,
-					      CONST 
-					      PIXELFORMATDESCRIPTOR *ppfd)
-{
-    int		i,best = -1,bestdelta = 0x7FFFFFFF,delta;
-    (void) hdc;
-    
-    if(ppfd->nSize != sizeof(PIXELFORMATDESCRIPTOR) || ppfd->nVersion != 1)
-	{
-	    SetLastError(0);
-	    return(0);
-	}
-    for(i = 0; i < npfd;i++)
-	{
-	    delta = 0;
-	    if(
-		(ppfd->dwFlags & PFD_DRAW_TO_WINDOW) &&
-		!(pfd[i].pfd.dwFlags & PFD_DRAW_TO_WINDOW))
-		continue;
-	    if(
-		(ppfd->dwFlags & PFD_DRAW_TO_BITMAP) &&
-		!(pfd[i].pfd.dwFlags & PFD_DRAW_TO_BITMAP))
-		continue;
-	    if(
-		(ppfd->dwFlags & PFD_SUPPORT_GDI) &&
-		!(pfd[i].pfd.dwFlags & PFD_SUPPORT_GDI))
-		continue;
-	    if(
-		(ppfd->dwFlags & PFD_SUPPORT_OPENGL) &&
-		!(pfd[i].pfd.dwFlags & PFD_SUPPORT_OPENGL))
-		continue;
-	    if(
-		!(ppfd->dwFlags & PFD_DOUBLEBUFFER_DONTCARE) &&
-		((ppfd->dwFlags & PFD_DOUBLEBUFFER) != 
-		 (pfd[i].pfd.dwFlags & PFD_DOUBLEBUFFER)))
-		continue;
-	    if(
-		!(ppfd->dwFlags & PFD_STEREO_DONTCARE) &&
-		((ppfd->dwFlags & PFD_STEREO) != 
-		 (pfd[i].pfd.dwFlags & PFD_STEREO)))
-		continue;
-	    if(ppfd->iPixelType != pfd[i].pfd.iPixelType)
-		delta++;
-	    if(ppfd->cAlphaBits != pfd[i].pfd.cAlphaBits)
-		delta++;
-	    if(delta < bestdelta)
-		{
-		    best = i + 1;
-		    bestdelta = delta;
-		    if(bestdelta == 0)
-			break;
-		}
-	}
-    if(best == -1)
-	{
-	    SetLastError(0);
-	    return(0);
-	}
-    return(best);
-}
-
-WINGDIAPI int GLAPIENTRY wglDescribePixelFormat(HDC hdc,
-					        int iPixelFormat,
-					        UINT nBytes,
-					        LPPIXELFORMATDESCRIPTOR ppfd)
-{
-    (void) hdc;
-    
-    if(ppfd == NULL)
-	return(npfd);
-    if(iPixelFormat < 1 || iPixelFormat > npfd || 
-       nBytes != sizeof(PIXELFORMATDESCRIPTOR))
-	{
-	    SetLastError(0);
-	    return(0);
-	}
-    *ppfd = pfd[iPixelFormat - 1].pfd;
-    return(npfd);
-}
-
-WINGDIAPI PROC GLAPIENTRY wglGetProcAddress(LPCSTR lpszProc)
-{
-    PROC p = (PROC) _glapi_get_proc_address((const char *) lpszProc);
-    if (p)
-	return p;
-    
-    SetLastError(0);
-    return(NULL);
-}
-
-WINGDIAPI int GLAPIENTRY wglGetPixelFormat(HDC hdc)
-{
-    (void) hdc;
-    if(curPFD == 0) {
-	SetLastError(0);
-	return(0);
-    }
-    return(curPFD);
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglSetPixelFormat(HDC hdc,int iPixelFormat,
-					const PIXELFORMATDESCRIPTOR *ppfd)
-{
-    (void) hdc;
-
-    /* SetPixelFormat (hence wglSetPixelFormat) must not touch ppfd, per
-     * http://msdn.microsoft.com/en-us/library/dd369049(v=vs.85).aspx
-     */
-    (void) ppfd;
-
-    if(iPixelFormat < 1 || iPixelFormat > npfd) {
-	SetLastError(0);
-	return(FALSE);
-    }
-    curPFD = iPixelFormat;
-    return(TRUE);
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglSwapBuffers(HDC hdc)
-{
-    WMesaSwapBuffers(hdc);
-    return TRUE;
-}
-
-static FIXED FixedFromDouble(double d)
-{
-   long l = (long) (d * 65536L);
-   return *(FIXED *) (void *) &l;
-}
-
-
-/*
-** This is cribbed from FX/fxwgl.c, and seems to implement support
-** for bitmap fonts where the wglUseFontBitmapsA() code implements
-** support for outline fonts.  In combination they hopefully give
-** fairly generic support for fonts.
-*/
-static BOOL wglUseFontBitmaps_FX(HDC fontDevice, DWORD firstChar,
-                                 DWORD numChars, DWORD listBase)
-{
-#define VERIFY(a) (void)(a)
-    
-    TEXTMETRIC metric;
-    BITMAPINFO *dibInfo;
-    HDC bitDevice;
-    COLORREF tempColor;
-    int i;
-    
-    VERIFY(GetTextMetrics(fontDevice, &metric));
-    
-    dibInfo = calloc(sizeof(BITMAPINFO) + sizeof(RGBQUAD), 1);
-    dibInfo->bmiHeader.biSize = sizeof(BITMAPINFOHEADER);
-    dibInfo->bmiHeader.biPlanes = 1;
-    dibInfo->bmiHeader.biBitCount = 1;
-    dibInfo->bmiHeader.biCompression = BI_RGB;
-    
-    bitDevice = CreateCompatibleDC(fontDevice);
-    
-    /* Swap fore and back colors so the bitmap has the right polarity */
-    tempColor = GetBkColor(bitDevice);
-    SetBkColor(bitDevice, GetTextColor(bitDevice));
-    SetTextColor(bitDevice, tempColor);
-    
-    /* Place chars based on base line */
-    VERIFY(SetTextAlign(bitDevice, TA_BASELINE) != GDI_ERROR ? 1 : 0);
-    
-    for(i = 0; i < (int)numChars; i++) {
-	SIZE size;
-	char curChar;
-	int charWidth,charHeight,bmapWidth,bmapHeight,numBytes,res;
-	HBITMAP bitObject;
-	HGDIOBJ origBmap;
-	unsigned char *bmap;
-	
-	curChar = (char)(i + firstChar);
-	
-	/* Find how high/wide this character is */
-	VERIFY(GetTextExtentPoint32(bitDevice, &curChar, 1, &size));
-	
-	/* Create the output bitmap */
-	charWidth = size.cx;
-	charHeight = size.cy;
-	/* Round up to the next multiple of 32 bits */
-	bmapWidth = ((charWidth + 31) / 32) * 32;   
-	bmapHeight = charHeight;
-	bitObject = CreateCompatibleBitmap(bitDevice,
-					   bmapWidth,
-					   bmapHeight);
-	/* VERIFY(bitObject); */
-	
-	/* Assign the output bitmap to the device */
-	origBmap = SelectObject(bitDevice, bitObject);
-	(void) VERIFY(origBmap);
-	
-	VERIFY( PatBlt( bitDevice, 0, 0, bmapWidth, bmapHeight,BLACKNESS ) );
-	
-	/* Use our source font on the device */
-	VERIFY(SelectObject(bitDevice, GetCurrentObject(fontDevice,OBJ_FONT)));
-	
-	/* Draw the character */
-	VERIFY(TextOut(bitDevice, 0, metric.tmAscent, &curChar, 1));
-	
-	/* Unselect our bmap object */
-	VERIFY(SelectObject(bitDevice, origBmap));
-	
-	/* Convert the display dependant representation to a 1 bit deep DIB */
-	numBytes = (bmapWidth * bmapHeight) / 8;
-	bmap = malloc(numBytes);
-	dibInfo->bmiHeader.biWidth = bmapWidth;
-	dibInfo->bmiHeader.biHeight = bmapHeight;
-	res = GetDIBits(bitDevice, bitObject, 0, bmapHeight, bmap,
-			dibInfo,
-			DIB_RGB_COLORS);
-	/* VERIFY(res); */
-	
-	/* Create the GL object */
-	glNewList(i + listBase, GL_COMPILE);
-	glBitmap(bmapWidth, bmapHeight, 0.0, (GLfloat)metric.tmDescent,
-		 (GLfloat)charWidth, 0.0,
-		 bmap);
-	glEndList();
-	/* CheckGL(); */
-	
-	/* Destroy the bmap object */
-	DeleteObject(bitObject);
-	
-	/* Deallocate the bitmap data */
-	free(bmap);
-    }
-    
-    /* Destroy the DC */
-    VERIFY(DeleteDC(bitDevice));
-    
-    free(dibInfo);
-    
-    return TRUE;
-#undef VERIFY
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglUseFontBitmapsA(HDC hdc, DWORD first,
-					     DWORD count, DWORD listBase)
-{
-    int i;
-    GLuint font_list;
-    DWORD size;
-    GLYPHMETRICS gm;
-    HANDLE hBits;
-    LPSTR lpBits;
-    MAT2 mat;
-    int  success = TRUE;
-    
-    if (count == 0)
-	return FALSE;
-    
-    font_list = listBase;
-    
-    mat.eM11 = FixedFromDouble(1);
-    mat.eM12 = FixedFromDouble(0);
-    mat.eM21 = FixedFromDouble(0);
-    mat.eM22 = FixedFromDouble(-1);
-    
-    memset(&gm,0,sizeof(gm));
-    
-    /*
-    ** If we can't get the glyph outline, it may be because this is a fixed
-    ** font.  Try processing it that way.
-    */
-    if( GetGlyphOutline(hdc, first, GGO_BITMAP, &gm, 0, NULL, &mat)
-	== GDI_ERROR ) {
-	return wglUseFontBitmaps_FX( hdc, first, count, listBase );
-    }
-    
-    /*
-    ** Otherwise process all desired characters.
-    */
-    for (i = 0; i < (int)count; i++) {
-	DWORD err;
-	
-	glNewList( font_list+i, GL_COMPILE );
-	
-	/* allocate space for the bitmap/outline */
-	size = GetGlyphOutline(hdc, first + i, GGO_BITMAP, 
-			       &gm, 0, NULL, &mat);
-	if (size == GDI_ERROR) {
-	    glEndList( );
-	    err = GetLastError();
-	    success = FALSE;
-	    continue;
-	}
-	
-	hBits  = GlobalAlloc(GHND, size+1);
-	lpBits = GlobalLock(hBits);
-	
-	err = 
-	    GetGlyphOutline(hdc,         /* handle to device context */
-			    first + i,   /* character to query */
-			    GGO_BITMAP,  /* format of data to return */
-			    &gm,         /* ptr to structure for metrics*/
-			    size,        /* size of buffer for data */
-			    lpBits,      /* pointer to buffer for data */
-			    &mat         /* pointer to transformation */
-			    /* matrix structure */
-		);
-	
-	if (err == GDI_ERROR) {
-	    GlobalUnlock(hBits);
-	    GlobalFree(hBits);
-	    
-	    glEndList( );
-	    err = GetLastError();
-	    success = FALSE;
-	    continue;
-	}
-	
-	glBitmap(gm.gmBlackBoxX,gm.gmBlackBoxY,
-		 (GLfloat)-gm.gmptGlyphOrigin.x,
-		 (GLfloat)gm.gmptGlyphOrigin.y,
-		 (GLfloat)gm.gmCellIncX,
-		 (GLfloat)gm.gmCellIncY,
-		 (const GLubyte * )lpBits);
-	
-	GlobalUnlock(hBits);
-	GlobalFree(hBits);
-	
-	glEndList( );
-    }
-    
-    return success;
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglShareLists(HGLRC hglrc1,
-					HGLRC hglrc2)
-{
-    WMesaShareLists((WMesaContext)hglrc1, (WMesaContext)hglrc2);
-    return(TRUE);
-}
-
-
-
-/* NOT IMPLEMENTED YET */
-WINGDIAPI BOOL GLAPIENTRY wglCopyContext(HGLRC hglrcSrc,
-					 HGLRC hglrcDst,
-					 UINT mask)
-{
-    (void) hglrcSrc; (void) hglrcDst; (void) mask;
-    return(FALSE);
-}
-
-WINGDIAPI HGLRC GLAPIENTRY wglCreateLayerContext(HDC hdc,
-						 int iLayerPlane)
-{
-    SetLastError(0);
-    if (iLayerPlane == 0)
-        return wglCreateContext( hdc );
-    return(NULL);
-}
-
-
-WINGDIAPI BOOL GLAPIENTRY wglUseFontBitmapsW(HDC hdc,
-					     DWORD first,
-					     DWORD count,
-					     DWORD listBase)
-{
-    (void) hdc; (void) first; (void) count; (void) listBase;
-    return FALSE;
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglUseFontOutlinesA(HDC hdc,
-					      DWORD first,
-					      DWORD count,
-					      DWORD listBase,
-					      FLOAT deviation,
-					      FLOAT extrusion,
-					      int format,
-					      LPGLYPHMETRICSFLOAT lpgmf)
-{
-    (void) hdc; (void) first; (void) count;
-    (void) listBase; (void) deviation; (void) extrusion; (void) format;
-    (void) lpgmf;
-    SetLastError(0);
-    return(FALSE);
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglUseFontOutlinesW(HDC hdc,
-					      DWORD first,
-					      DWORD count,
-					      DWORD listBase,
-					      FLOAT deviation,
-					      FLOAT extrusion,
-					      int format,
-					      LPGLYPHMETRICSFLOAT lpgmf)
-{
-    (void) hdc; (void) first; (void) count;
-    (void) listBase; (void) deviation; (void) extrusion; (void) format;
-    (void) lpgmf;
-    SetLastError(0);
-    return(FALSE);
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglDescribeLayerPlane(HDC hdc,
-						int iPixelFormat,
-						int iLayerPlane,
-						UINT nBytes,
-						LPLAYERPLANEDESCRIPTOR plpd)
-{
-    (void) hdc; (void) iPixelFormat; (void) iLayerPlane; 
-    (void) nBytes; (void) plpd;
-    SetLastError(0);
-    return(FALSE);
-}
-
-WINGDIAPI int GLAPIENTRY wglSetLayerPaletteEntries(HDC hdc,
-						   int iLayerPlane,
-						   int iStart,
-						   int cEntries,
-						   CONST COLORREF *pcr)
-{
-    (void) hdc; (void) iLayerPlane; (void) iStart; 
-    (void) cEntries; (void) pcr;
-    SetLastError(0);
-    return(0);
-}
-
-WINGDIAPI int GLAPIENTRY wglGetLayerPaletteEntries(HDC hdc,
-						   int iLayerPlane,
-						   int iStart,
-						   int cEntries,
-						   COLORREF *pcr)
-{
-    (void) hdc; (void) iLayerPlane; (void) iStart; (void) cEntries; (void) pcr;
-    SetLastError(0);
-    return(0);
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglRealizeLayerPalette(HDC hdc,
-						 int iLayerPlane,
-						 BOOL bRealize)
-{
-    (void) hdc; (void) iLayerPlane; (void) bRealize;
-    SetLastError(0);
-    return(FALSE);
-}
-
-WINGDIAPI BOOL GLAPIENTRY wglSwapLayerBuffers(HDC hdc,
-					      UINT fuPlanes)
-{
-    (void) hdc; (void) fuPlanes;
-    SetLastError(0);
-    return(FALSE);
-}
-
-WINGDIAPI const char * GLAPIENTRY wglGetExtensionsStringARB(HDC hdc)
-{
-    return "WGL_ARB_extensions_string";
-}
diff --git a/mesalib/src/mesa/drivers/windows/gdi/wmesa.c b/mesalib/src/mesa/drivers/windows/gdi/wmesa.c
deleted file mode 100644
index c1a9c28b1..000000000
--- a/mesalib/src/mesa/drivers/windows/gdi/wmesa.c
+++ /dev/null
@@ -1,791 +0,0 @@
-/*
- * Windows (Win32/Win64) device driver for Mesa
- *
- */
-
-#include "wmesadef.h"
-#include "colors.h"
-#include "GL/wmesa.h"
-#include <winuser.h>
-#include "main/context.h"
-#include "main/api_exec.h"
-#include "main/extensions.h"
-#include "main/framebuffer.h"
-#include "main/renderbuffer.h"
-#include "main/macros.h"
-#include "main/version.h"
-#include "main/vtxfmt.h"
-#include "drivers/common/driverfuncs.h"
-#include "drivers/common/meta.h"
-#include "vbo/vbo.h"
-#include "swrast/swrast.h"
-#include "swrast/s_renderbuffer.h"
-#include "swrast_setup/swrast_setup.h"
-#include "tnl/tnl.h"
-#include "tnl/t_context.h"
-#include "tnl/t_pipeline.h"
-
-
-/* linked list of our Framebuffers (windows) */
-static WMesaFramebuffer FirstFramebuffer = NULL;
-
-
-/**
- * Create a new WMesaFramebuffer object which will correspond to the
- * given HDC (Window handle).
- */
-static WMesaFramebuffer
-wmesa_new_framebuffer(HDC hdc, struct gl_config *visual)
-{
-    WMesaFramebuffer pwfb
-        = malloc(sizeof(struct wmesa_framebuffer));
-    if (pwfb) {
-        _mesa_initialize_window_framebuffer(&pwfb->Base, visual);
-        pwfb->hDC = hdc;
-        /* insert at head of list */
-        pwfb->next = FirstFramebuffer;
-        FirstFramebuffer = pwfb;
-    }
-    return pwfb;
-}
-
-/**
- * Given an hdc, free the corresponding WMesaFramebuffer
- */
-static void
-wmesa_free_framebuffer(HDC hdc)
-{
-    WMesaFramebuffer pwfb, prev;
-    for (pwfb = FirstFramebuffer; pwfb; pwfb = pwfb->next) {
-        if (pwfb->hDC == hdc)
-            break;
-	prev = pwfb;
-    }
-    if (pwfb) {
-        struct gl_framebuffer *fb;
-	if (pwfb == FirstFramebuffer)
-	    FirstFramebuffer = pwfb->next;
-	else
-	    prev->next = pwfb->next;
-        fb = &pwfb->Base;
-        _mesa_reference_framebuffer(&fb, NULL); 
-    }
-}
-
-/**
- * Given an hdc, return the corresponding WMesaFramebuffer
- */
-static WMesaFramebuffer
-wmesa_lookup_framebuffer(HDC hdc)
-{
-    WMesaFramebuffer pwfb;
-    for (pwfb = FirstFramebuffer; pwfb; pwfb = pwfb->next) {
-        if (pwfb->hDC == hdc)
-            return pwfb;
-    }
-    return NULL;
-}
-
-
-/**
- * Given a struct gl_framebuffer, return the corresponding WMesaFramebuffer.
- */
-static WMesaFramebuffer wmesa_framebuffer(struct gl_framebuffer *fb)
-{
-    return (WMesaFramebuffer) fb;
-}
-
-
-/**
- * Given a struct gl_context, return the corresponding WMesaContext.
- */
-static WMesaContext wmesa_context(const struct gl_context *ctx)
-{
-    return (WMesaContext) ctx;
-}
-
-
-/*
- * Every driver should implement a GetString function in order to
- * return a meaningful GL_RENDERER string.
- */
-static const GLubyte *wmesa_get_string(struct gl_context *ctx, GLenum name)
-{
-    return (name == GL_RENDERER) ? 
-	(GLubyte *) "Mesa Windows GDI Driver" : NULL;
-}
-
-
-/*
- * Determine the pixel format based on the pixel size.
- */
-static void wmSetPixelFormat(WMesaFramebuffer pwfb, HDC hDC)
-{
-    pwfb->cColorBits = GetDeviceCaps(hDC, BITSPIXEL);
-
-    /* Only 16 and 32 bit targets are supported now */
-    assert(pwfb->cColorBits == 0 ||
-	   pwfb->cColorBits == 16 || 
-	   pwfb->cColorBits == 24 || 
-	   pwfb->cColorBits == 32);
-
-    switch(pwfb->cColorBits){
-    case 8:
-	pwfb->pixelformat = PF_INDEX8;
-	break;
-    case 16:
-	pwfb->pixelformat = PF_5R6G5B;
-	break;
-    case 24:
-    case 32:
-	pwfb->pixelformat = PF_8R8G8B;
-	break;
-    default:
-	pwfb->pixelformat = PF_BADFORMAT;
-    }
-}
-
-
-/**
- * Create DIB for back buffer.
- * We write into this memory with the span routines and then blit it
- * to the window on a buffer swap.
- */
-static BOOL wmCreateBackingStore(WMesaFramebuffer pwfb, long lxSize, long lySize)
-{
-    LPBITMAPINFO pbmi = &(pwfb->bmi);
-    HDC          hic;
-
-    pbmi->bmiHeader.biSize = sizeof(BITMAPINFOHEADER);
-    pbmi->bmiHeader.biWidth = lxSize;
-    pbmi->bmiHeader.biHeight= -lySize;
-    pbmi->bmiHeader.biPlanes = 1;
-    pbmi->bmiHeader.biBitCount = GetDeviceCaps(pwfb->hDC, BITSPIXEL);
-    pbmi->bmiHeader.biCompression = BI_RGB;
-    pbmi->bmiHeader.biSizeImage = 0;
-    pbmi->bmiHeader.biXPelsPerMeter = 0;
-    pbmi->bmiHeader.biYPelsPerMeter = 0;
-    pbmi->bmiHeader.biClrUsed = 0;
-    pbmi->bmiHeader.biClrImportant = 0;
-    
-    pwfb->cColorBits = pbmi->bmiHeader.biBitCount;
-    pwfb->ScanWidth = (lxSize * (pwfb->cColorBits / 8) + 3) & ~3;
-    
-    hic = CreateIC("display", NULL, NULL, NULL);
-    pwfb->dib_hDC = CreateCompatibleDC(hic);
-    
-    pwfb->hbmDIB = CreateDIBSection(hic,
-				   &pwfb->bmi,
-				   DIB_RGB_COLORS,
-				   (void **)&(pwfb->pbPixels),
-				   0,
-				   0);
-    pwfb->hOldBitmap = SelectObject(pwfb->dib_hDC, pwfb->hbmDIB);
-    
-    DeleteDC(hic);
-
-    wmSetPixelFormat(pwfb, pwfb->hDC);
-    return TRUE;
-}
-
-
-static void wmDeleteBackingStore(WMesaFramebuffer pwfb)
-{
-    if (pwfb->hbmDIB) {
-	SelectObject(pwfb->dib_hDC, pwfb->hOldBitmap);
-	DeleteDC(pwfb->dib_hDC);
-	DeleteObject(pwfb->hbmDIB);
-    }
-}
-
-
-/**
- * Find the width and height of the window named by hdc.
- */
-static void
-get_window_size(HDC hdc, GLuint *width, GLuint *height)
-{
-    if (WindowFromDC(hdc)) {
-        RECT rect;
-        GetClientRect(WindowFromDC(hdc), &rect);
-        *width = rect.right - rect.left;
-        *height = rect.bottom - rect.top;
-    }
-    else { /* Memory context */
-        /* From contributed code - use the size of the desktop
-         * for the size of a memory context (?) */
-        *width = GetDeviceCaps(hdc, HORZRES);
-        *height = GetDeviceCaps(hdc, VERTRES);
-    }
-}
-
-
-static void
-wmesa_get_buffer_size(struct gl_framebuffer *buffer, GLuint *width, GLuint *height)
-{
-    WMesaFramebuffer pwfb = wmesa_framebuffer(buffer);
-    get_window_size(pwfb->hDC, width, height);
-}
-
-
-static void wmesa_flush(struct gl_context *ctx)
-{
-    WMesaFramebuffer pwfb = wmesa_framebuffer(ctx->WinSysDrawBuffer);
-
-    if (ctx->Visual.doubleBufferMode == 1) {
-	BitBlt(pwfb->hDC, 0, 0, pwfb->Base.Width, pwfb->Base.Height,
-	       pwfb->dib_hDC, 0, 0, SRCCOPY);
-    }
-    else {
-	/* Do nothing for single buffer */
-    }
-}
-
-
-/**********************************************************************/
-/*****                   CLEAR Functions                          *****/
-/**********************************************************************/
-
-/* 
- * Clear the color/depth/stencil buffers.
- */ 
-static void clear(struct gl_context *ctx, GLbitfield mask)
-{
-#define FLIP(Y)  (ctx->DrawBuffer->Height - (Y) - 1)
-    const GLint x = ctx->DrawBuffer->_Xmin;
-    const GLint y = ctx->DrawBuffer->_Ymin;
-    const GLint height = ctx->DrawBuffer->_Ymax - ctx->DrawBuffer->_Ymin;
-    const GLint width  = ctx->DrawBuffer->_Xmax - ctx->DrawBuffer->_Xmin;
-
-    WMesaContext pwc = wmesa_context(ctx);
-    WMesaFramebuffer pwfb = wmesa_framebuffer(ctx->DrawBuffer);
-    int done = 0;
-
-    /* Let swrast do all the work if the masks are not set to
-     * clear all channels. */
-    if (!ctx->Color.ColorMask[0][0] ||
-	!ctx->Color.ColorMask[0][1] ||
-	!ctx->Color.ColorMask[0][2] ||
-	!ctx->Color.ColorMask[0][3]) {
-	_swrast_Clear(ctx, mask);
-	return;
-    }
-
-    if (mask & BUFFER_BITS_COLOR) {
-       /* setup the clearing color */
-       const union gl_color_union color = ctx->Color.ClearColor;
-       GLubyte col[3];
-       UNCLAMPED_FLOAT_TO_UBYTE(col[0], color.f[0]);
-       UNCLAMPED_FLOAT_TO_UBYTE(col[1], color.f[1]);
-       UNCLAMPED_FLOAT_TO_UBYTE(col[2], color.f[2]);
-       pwc->clearColorRef = RGB(col[0], col[1], col[2]);
-       DeleteObject(pwc->clearPen);
-       DeleteObject(pwc->clearBrush);
-       pwc->clearPen = CreatePen(PS_SOLID, 1, pwc->clearColorRef); 
-       pwc->clearBrush = CreateSolidBrush(pwc->clearColorRef); 
-    }
-
-    /* Back buffer */
-    if (mask & BUFFER_BIT_BACK_LEFT) { 
-	
-	int     i, rowSize; 
-	UINT    bytesPerPixel = pwfb->cColorBits / 8; 
-	LPBYTE  lpb, clearRow;
-	LPWORD  lpw;
-	BYTE    bColor; 
-	WORD    wColor; 
-	BYTE    r, g, b; 
-	DWORD   dwColor; 
-	LPDWORD lpdw; 
-	
-	/* Try for a fast clear - clearing entire buffer with a single
-	 * byte value. */
-	if (width == ctx->DrawBuffer->Width &&
-            height == ctx->DrawBuffer->Height) { /* entire buffer */
-	    /* Now check for an easy clear value */
-	    switch (bytesPerPixel) {
-	    case 1:
-		bColor = BGR8(GetRValue(pwc->clearColorRef), 
-			      GetGValue(pwc->clearColorRef), 
-			      GetBValue(pwc->clearColorRef));
-		memset(pwfb->pbPixels, bColor, 
-		       pwfb->ScanWidth * height);
-		done = 1;
-		break;
-	    case 2:
-		wColor = BGR16(GetRValue(pwc->clearColorRef), 
-			       GetGValue(pwc->clearColorRef), 
-			       GetBValue(pwc->clearColorRef)); 
-		if (((wColor >> 8) & 0xff) == (wColor & 0xff)) {
-		    memset(pwfb->pbPixels, wColor & 0xff, 
-			   pwfb->ScanWidth * height);
-		    done = 1;
-		}
-		break;
-	    case 3:
-		/* fall through */
-	    case 4:
-		if (GetRValue(pwc->clearColorRef) == 
-		    GetGValue(pwc->clearColorRef) &&
-		    GetRValue(pwc->clearColorRef) == 
-		    GetBValue(pwc->clearColorRef)) {
-		    memset(pwfb->pbPixels, 
-			   GetRValue(pwc->clearColorRef), 
-			   pwfb->ScanWidth * height);
-		    done = 1;
-		}
-		break;
-	    default:
-		break;
-	    }
-	} /* all */
-
-	if (!done) {
-	    /* Need to clear a row at a time.  Begin by setting the first
-	     * row in the area to be cleared to the clear color. */
-	    
-	    clearRow = pwfb->pbPixels + 
-		pwfb->ScanWidth * FLIP(y) +
-		bytesPerPixel * x; 
-	    switch (bytesPerPixel) {
-	    case 1:
-		lpb = clearRow;
-		bColor = BGR8(GetRValue(pwc->clearColorRef), 
-			      GetGValue(pwc->clearColorRef), 
-			      GetBValue(pwc->clearColorRef));
-		memset(lpb, bColor, width);
-		break;
-	    case 2:
-		lpw = (LPWORD)clearRow;
-		wColor = BGR16(GetRValue(pwc->clearColorRef), 
-			       GetGValue(pwc->clearColorRef), 
-			       GetBValue(pwc->clearColorRef)); 
-		for (i=0; i<width; i++)
-		    *lpw++ = wColor;
-		break;
-	    case 3: 
-		lpb = clearRow;
-		r = GetRValue(pwc->clearColorRef); 
-		g = GetGValue(pwc->clearColorRef); 
-		b = GetBValue(pwc->clearColorRef); 
-		for (i=0; i<width; i++) {
-		    *lpb++ = b; 
-		    *lpb++ = g; 
-		    *lpb++ = r; 
-		} 
-		break;
-	    case 4: 
-		lpdw = (LPDWORD)clearRow; 
-		dwColor = BGR32(GetRValue(pwc->clearColorRef), 
-				GetGValue(pwc->clearColorRef), 
-				GetBValue(pwc->clearColorRef)); 
-		for (i=0; i<width; i++)
-		    *lpdw++ = dwColor;
-		break;
-	    default:
-		break;
-	    } /* switch */
-	    
-	    /* copy cleared row to other rows in buffer */
-	    lpb = clearRow - pwfb->ScanWidth;
-	    rowSize = width * bytesPerPixel;
-	    for (i=1; i<height; i++) { 
-		memcpy(lpb, clearRow, rowSize); 
-		lpb -= pwfb->ScanWidth; 
-	    } 
-	} /* not done */
-	mask &= ~BUFFER_BIT_BACK_LEFT;
-    } /* back buffer */ 
-
-    /* front buffer */
-    if (mask & BUFFER_BIT_FRONT_LEFT) { 
-	HDC DC = pwc->hDC; 
-	HPEN Old_Pen = SelectObject(DC, pwc->clearPen); 
-	HBRUSH Old_Brush = SelectObject(DC, pwc->clearBrush);
-	Rectangle(DC,
-		  x,
-		  FLIP(y) + 1,
-		  x + width + 1,
-		  FLIP(y) - height + 1);
-	SelectObject(DC, Old_Pen); 
-	SelectObject(DC, Old_Brush); 
-	mask &= ~BUFFER_BIT_FRONT_LEFT;
-    } /* front buffer */ 
-    
-    /* Call swrast if there is anything left to clear (like DEPTH) */ 
-    if (mask) 
-	_swrast_Clear(ctx, mask);
-  
-#undef FLIP
-} 
-
-
-
-/**********************************************************************/
-/*****                   BUFFER Functions                         *****/
-/**********************************************************************/
-
-
-
-
-static void
-wmesa_delete_renderbuffer(struct gl_context *ctx, struct gl_renderbuffer *rb)
-{
-    _mesa_delete_renderbuffer(ctx, rb);
-}
-
-
-/**
- * This is called by Mesa whenever it determines that the window size
- * has changed.  Do whatever's needed to cope with that.
- */
-static GLboolean
-wmesa_renderbuffer_storage(struct gl_context *ctx, 
-			   struct gl_renderbuffer *rb,
-			   GLenum internalFormat, 
-			   GLuint width, 
-			   GLuint height)
-{
-    rb->Width = width;
-    rb->Height = height;
-    return GL_TRUE;
-}
-
-
-/**
- * Called by ctx->Driver.ResizeBuffers()
- * Resize the front/back colorbuffers to match the latest window size.
- */
-static void
-wmesa_resize_buffers(struct gl_context *ctx, struct gl_framebuffer *buffer,
-                     GLuint width, GLuint height)
-{
-    WMesaFramebuffer pwfb = wmesa_framebuffer(buffer);
-
-    if (pwfb->Base.Width != width || pwfb->Base.Height != height) {
-	/* Realloc back buffer */
-	if (ctx->Visual.doubleBufferMode == 1) {
-	    wmDeleteBackingStore(pwfb);
-	    wmCreateBackingStore(pwfb, width, height);
-	}
-    }
-    _mesa_resize_framebuffer(ctx, buffer, width, height);
-}
-
-
-/**
- * Called by glViewport.
- * This is a good time for us to poll the current window size and adjust
- * our renderbuffers to match the current window size.
- * Remember, we have no opportunity to respond to conventional
- * resize events since the driver has no event loop.
- * Thus, we poll.
- * MakeCurrent also ends up making a call here, so that ensures
- * we get the viewport set correctly, even if the app does not call
- * glViewport and relies on the defaults.
- */
-static void wmesa_viewport(struct gl_context *ctx)
-{
-    GLuint new_width, new_height;
-
-    wmesa_get_buffer_size(ctx->WinSysDrawBuffer, &new_width, &new_height);
-
-    /**
-     * Resize buffers if the window size changed.
-     */
-    wmesa_resize_buffers(ctx, ctx->WinSysDrawBuffer, new_width, new_height);
-    ctx->NewState |= _NEW_BUFFERS;  /* to update scissor / window bounds */
-}
-
-
-
-
-/**
- * Called when the driver should update it's state, based on the new_state
- * flags.
- */
-static void wmesa_update_state(struct gl_context *ctx, GLuint new_state)
-{
-    _swrast_InvalidateState(ctx, new_state);
-    _swsetup_InvalidateState(ctx, new_state);
-    _vbo_InvalidateState(ctx, new_state);
-    _tnl_InvalidateState(ctx, new_state);
-
-    /* TODO - This code is not complete yet because I 
-     * don't know what to do for all state updates.
-     */
-
-    if (new_state & _NEW_BUFFERS) {
-    }
-}
-
-
-
-
-
-/**********************************************************************/
-/*****                   WMESA Functions                          *****/
-/**********************************************************************/
-
-WMesaContext WMesaCreateContext(HDC hDC, 
-				HPALETTE* Pal,
-				GLboolean rgb_flag,
-				GLboolean db_flag,
-				GLboolean alpha_flag)
-{
-    WMesaContext c;
-    struct dd_function_table functions;
-    GLint red_bits, green_bits, blue_bits, alpha_bits;
-    struct gl_context *ctx;
-    struct gl_config *visual;
-
-    (void) Pal;
-    
-    /* Indexed mode not supported */
-    if (!rgb_flag)
-	return NULL;
-
-    /* Allocate wmesa context */
-    c = CALLOC_STRUCT(wmesa_context);
-    if (!c)
-	return NULL;
-
-#if 0
-    /* I do not understand this contributed code */
-    /* Support memory and device contexts */
-    if(WindowFromDC(hDC) != NULL) {
-	c->hDC = GetDC(WindowFromDC(hDC)); /* huh ???? */
-    }
-    else {
-	c->hDC = hDC;
-    }
-#else
-    c->hDC = hDC;
-#endif
-
-    /* Get data for visual */
-    /* Dealing with this is actually a bit of overkill because Mesa will end
-     * up treating all color component size requests less than 8 by using 
-     * a single byte per channel.  In addition, the interface to the span
-     * routines passes colors as an entire byte per channel anyway, so there
-     * is nothing to be saved by telling the visual to be 16 bits if the device
-     * is 16 bits.  That is, Mesa is going to compute colors down to 8 bits per
-     * channel anyway.
-     * But we go through the motions here anyway.
-     */
-    switch (GetDeviceCaps(c->hDC, BITSPIXEL)) {
-    case 16:
-	red_bits = green_bits = blue_bits = 5;
-	alpha_bits = 0;
-	break;
-    default:
-	red_bits = green_bits = blue_bits = 8;
-	alpha_bits = 8;
-	break;
-    }
-    /* Create visual based on flags */
-    visual = _mesa_create_visual(db_flag,    /* db_flag */
-                                 GL_FALSE,   /* stereo */
-                                 red_bits, green_bits, blue_bits, /* color RGB */
-                                 alpha_flag ? alpha_bits : 0, /* color A */
-                                 DEFAULT_SOFTWARE_DEPTH_BITS, /* depth_bits */
-                                 8,          /* stencil_bits */
-                                 16,16,16,   /* accum RGB */
-                                 alpha_flag ? 16 : 0, /* accum A */
-                                 1);         /* num samples */
-    
-    if (!visual) {
-	free(c);
-	return NULL;
-    }
-
-    /* Set up driver functions */
-    _mesa_init_driver_functions(&functions);
-    functions.GetString = wmesa_get_string;
-    functions.UpdateState = wmesa_update_state;
-    functions.Flush = wmesa_flush;
-    functions.Clear = clear;
-    functions.ResizeBuffers = wmesa_resize_buffers;
-    functions.Viewport = wmesa_viewport;
-
-    /* initialize the Mesa context data */
-    ctx = &c->gl_ctx;
-    _mesa_initialize_context(ctx, API_OPENGL_COMPAT, visual,
-                             NULL, &functions);
-
-    /* visual no longer needed - it was copied by _mesa_initialize_context() */
-    _mesa_destroy_visual(visual);
-
-    _mesa_enable_sw_extensions(ctx);
-  
-    _mesa_meta_init(ctx);
-
-    /* Initialize the software rasterizer and helper modules. */
-    if (!_swrast_CreateContext(ctx) ||
-        !_vbo_CreateContext(ctx) ||
-        !_tnl_CreateContext(ctx) ||
-	!_swsetup_CreateContext(ctx)) {
-	_mesa_free_context_data(ctx);
-	free(c);
-	return NULL;
-    }
-    _swsetup_Wakeup(ctx);
-    TNL_CONTEXT(ctx)->Driver.RunPipeline = _tnl_run_pipeline;
-
-    _mesa_compute_version(ctx);
-
-    /* Exec table initialization requires the version to be computed */
-    _mesa_initialize_dispatch_tables(ctx);
-    _mesa_initialize_vbo_vtxfmt(ctx);
-
-    return c;
-}
-
-
-void WMesaDestroyContext( WMesaContext pwc )
-{
-    struct gl_context *ctx = &pwc->gl_ctx;
-    WMesaFramebuffer pwfb;
-    GET_CURRENT_CONTEXT(cur_ctx);
-
-    if (cur_ctx == ctx) {
-        /* unbind current if deleting current context */
-        WMesaMakeCurrent(NULL, NULL);
-    }
-
-    /* clean up frame buffer resources */
-    pwfb = wmesa_lookup_framebuffer(pwc->hDC);
-    if (pwfb) {
-	if (ctx->Visual.doubleBufferMode == 1)
-	    wmDeleteBackingStore(pwfb);
-	wmesa_free_framebuffer(pwc->hDC);
-    }
-
-    /* Release for device, not memory contexts */
-    if (WindowFromDC(pwc->hDC) != NULL)
-    {
-      ReleaseDC(WindowFromDC(pwc->hDC), pwc->hDC);
-    }
-    DeleteObject(pwc->clearPen); 
-    DeleteObject(pwc->clearBrush); 
-    
-    _mesa_meta_free(ctx);
-
-    _swsetup_DestroyContext(ctx);
-    _tnl_DestroyContext(ctx);
-    _vbo_DestroyContext(ctx);
-    _swrast_DestroyContext(ctx);
-    
-    _mesa_free_context_data(ctx);
-    free(pwc);
-}
-
-
-/**
- * Create a new color renderbuffer.
- */
-static struct gl_renderbuffer *
-wmesa_new_renderbuffer(void)
-{
-    struct gl_renderbuffer *rb = CALLOC_STRUCT(gl_renderbuffer);
-    if (!rb)
-        return NULL;
-
-    _mesa_init_renderbuffer(rb, (GLuint)0);
-    
-    rb->_BaseFormat = GL_RGBA;
-    rb->InternalFormat = GL_RGBA;
-    rb->Delete = wmesa_delete_renderbuffer;
-    rb->AllocStorage = wmesa_renderbuffer_storage;
-    return rb;
-}
-
-
-void WMesaMakeCurrent(WMesaContext c, HDC hdc)
-{
-    WMesaFramebuffer pwfb;
-
-    {
-        /* return if already current */
-        GET_CURRENT_CONTEXT(ctx);
-        WMesaContext pwc = wmesa_context(ctx);
-        if (pwc && c == pwc && pwc->hDC == hdc)
-            return;
-    }
-
-    pwfb = wmesa_lookup_framebuffer(hdc);
-
-    /* Lazy creation of framebuffers */
-    if (c && !pwfb && hdc) {
-        struct gl_renderbuffer *rb;
-        struct gl_config *visual = &c->gl_ctx.Visual;
-        GLuint width, height;
-
-        get_window_size(hdc, &width, &height);
-
-	c->clearPen = CreatePen(PS_SOLID, 1, 0); 
-	c->clearBrush = CreateSolidBrush(0); 
-
-        pwfb = wmesa_new_framebuffer(hdc, visual);
-
-	/* Create back buffer if double buffered */
-	if (visual->doubleBufferMode == 1) {
-	    wmCreateBackingStore(pwfb, width, height);
-	}
-	
-        /* make render buffers */
-        if (visual->doubleBufferMode == 1) {
-            rb = wmesa_new_renderbuffer();
-            _mesa_add_renderbuffer(&pwfb->Base, BUFFER_BACK_LEFT, rb);
-	}
-        rb = wmesa_new_renderbuffer();
-        _mesa_add_renderbuffer(&pwfb->Base, BUFFER_FRONT_LEFT, rb);
-
-	/* Let Mesa own the Depth, Stencil, and Accum buffers */
-        _swrast_add_soft_renderbuffers(&pwfb->Base,
-                                       GL_FALSE, /* color */
-                                       visual->depthBits > 0,
-                                       visual->stencilBits > 0,
-                                       visual->accumRedBits > 0,
-                                       visual->alphaBits >0, 
-                                       GL_FALSE);
-    }
-
-    if (c && pwfb)
-	_mesa_make_current(&c->gl_ctx, &pwfb->Base, &pwfb->Base);
-    else
-        _mesa_make_current(NULL, NULL, NULL);
-}
-
-
-void WMesaSwapBuffers( HDC hdc )
-{
-    GET_CURRENT_CONTEXT(ctx);
-    WMesaContext pwc = wmesa_context(ctx);
-    WMesaFramebuffer pwfb = wmesa_lookup_framebuffer(hdc);
-
-    if (!pwfb) {
-        _mesa_problem(NULL, "wmesa: swapbuffers on unknown hdc");
-        return;
-    }
-
-    /* If we're swapping the buffer associated with the current context
-     * we have to flush any pending rendering commands first.
-     */
-    if (pwc->hDC == hdc) {
-	_mesa_notifySwapBuffers(ctx);
-
-	BitBlt(pwfb->hDC, 0, 0, pwfb->Base.Width, pwfb->Base.Height,
-	       pwfb->dib_hDC, 0, 0, SRCCOPY);
-    }
-    else {
-        /* XXX for now only allow swapping current window */
-        _mesa_problem(NULL, "wmesa: can't swap non-current window");
-    }
-}
-
-void WMesaShareLists(WMesaContext ctx_to_share, WMesaContext ctx)
-{
-	_mesa_share_state(&ctx->gl_ctx, &ctx_to_share->gl_ctx);	
-}
-
diff --git a/mesalib/src/mesa/drivers/windows/gdi/wmesadef.h b/mesalib/src/mesa/drivers/windows/gdi/wmesadef.h
deleted file mode 100644
index 9fda88390..000000000
--- a/mesalib/src/mesa/drivers/windows/gdi/wmesadef.h
+++ /dev/null
@@ -1,43 +0,0 @@
-#ifndef WMESADEF_H
-#define WMESADEF_H
-
-#include <windows.h>
-
-#include "main/context.h"
-
-
-/**
- * The Windows Mesa rendering context, derived from struct gl_context.
- */
-struct wmesa_context {
-    struct gl_context           gl_ctx;	        /* The core GL/Mesa context */
-    HDC                 hDC;
-    COLORREF		clearColorRef;
-    HPEN                clearPen;
-    HBRUSH              clearBrush;
-};
-
-
-/**
- * Windows framebuffer, derived from gl_framebuffer
- */
-struct wmesa_framebuffer
-{
-    struct gl_framebuffer Base;
-    HDC                 hDC;
-    int			pixelformat;
-    GLuint		ScanWidth;
-    int			cColorBits;
-    /* back buffer DIB fields */
-    HDC                 dib_hDC;
-    BITMAPINFO          bmi;
-    HBITMAP             hbmDIB;
-    HBITMAP             hOldBitmap;
-    PBYTE               pbPixels;
-    struct wmesa_framebuffer *next;
-};
-
-typedef struct wmesa_framebuffer *WMesaFramebuffer;
-
-
-#endif /* WMESADEF_H */
diff --git a/mesalib/src/mesa/main/.gitignore b/mesalib/src/mesa/main/.gitignore
index e65472d63..8256ad792 100644
--- a/mesalib/src/mesa/main/.gitignore
+++ b/mesalib/src/mesa/main/.gitignore
@@ -1,13 +1,11 @@
 api_exec.c
 dispatch.h
 enums.c
-get_es1.c
-get_es2.c
 git_sha1.h
 git_sha1.h.tmp
 remap_helper.h
 get_hash.h
 get_hash.h.tmp
-format_info.c
+format_info.h
 format_pack.c
 format_unpack.c
diff --git a/mesalib/src/mesa/main/api_arrayelt.c b/mesalib/src/mesa/main/api_arrayelt.c
index 536326f6e..ea015fd65 100644
--- a/mesalib/src/mesa/main/api_arrayelt.c
+++ b/mesalib/src/mesa/main/api_arrayelt.c
@@ -1548,7 +1548,7 @@ _ae_update_state(struct gl_context *ctx)
           */
          at->array = attribArray;
          at->binding = &vao->VertexBinding[attribArray->VertexBinding];
-         ASSERT(!at->array->Normalized);
+         assert(!at->array->Normalized);
          at->func = AttribFuncsNV[at->array->Normalized]
                                  [at->array->Size-1]
                                  [TYPE_IDX(at->array->Type)];
@@ -1610,8 +1610,8 @@ _ae_update_state(struct gl_context *ctx)
 
    check_vbo(actx, vao->IndexBufferObj);
 
-   ASSERT(at - actx->attribs <= VERT_ATTRIB_MAX);
-   ASSERT(aa - actx->arrays < 32);
+   assert(at - actx->attribs <= VERT_ATTRIB_MAX);
+   assert(aa - actx->arrays < 32);
    at->func = NULL;  /* terminate the list */
    aa->offset = -1;  /* terminate the list */
 
diff --git a/mesalib/src/mesa/main/api_exec.h b/mesalib/src/mesa/main/api_exec.h
index ff35b78bb..1e4a9d61a 100644
--- a/mesalib/src/mesa/main/api_exec.h
+++ b/mesalib/src/mesa/main/api_exec.h
@@ -26,6 +26,9 @@
 #ifndef API_EXEC_H
 #define API_EXEC_H
 
+#ifdef __cplusplus
+extern "C" {
+#endif
 
 struct _glapi_table;
 struct gl_context;
@@ -39,4 +42,8 @@ _mesa_initialize_exec_table(struct gl_context *ctx);
 extern void
 _mesa_initialize_dispatch_tables(struct gl_context *ctx);
 
+#ifdef __cplusplus
+} // extern "C"
+#endif
+
 #endif
diff --git a/mesalib/src/mesa/main/api_loopback.h b/mesalib/src/mesa/main/api_loopback.h
index 2195e01e3..44514405b 100644
--- a/mesalib/src/mesa/main/api_loopback.h
+++ b/mesalib/src/mesa/main/api_loopback.h
@@ -27,13 +27,7 @@
 #ifndef API_LOOPBACK_H
 #define API_LOOPBACK_H
 
-#include "main/compiler.h"
-#include "main/glheader.h" // ?
-#include "main/macros.h" // ?
-#include "main/mtypes.h" // ?
-#include "glapi/glapi.h" // ?
-#include "main/dispatch.h" // ?
-#include "main/context.h" // ?
+#include "main/glheader.h"
 
 struct _glapi_table;
 struct gl_context;
diff --git a/mesalib/src/mesa/main/arbprogram.c b/mesalib/src/mesa/main/arbprogram.c
index fe8cd6bbc..f474951d7 100644
--- a/mesalib/src/mesa/main/arbprogram.c
+++ b/mesalib/src/mesa/main/arbprogram.c
@@ -118,8 +118,8 @@ _mesa_BindProgramARB(GLenum target, GLuint id)
    }
 
    /* Never null pointers */
-   ASSERT(ctx->VertexProgram.Current);
-   ASSERT(ctx->FragmentProgram.Current);
+   assert(ctx->VertexProgram.Current);
+   assert(ctx->FragmentProgram.Current);
 
    if (ctx->Driver.BindProgram)
       ctx->Driver.BindProgram(ctx, target, newProg);
@@ -524,7 +524,7 @@ _mesa_ProgramLocalParameter4fARB(GLenum target, GLuint index,
 
    if (get_local_param_pointer(ctx, "glProgramLocalParameterARB",
 			       target, index, &param)) {
-      ASSERT(index < MAX_PROGRAM_LOCAL_PARAMS);
+      assert(index < MAX_PROGRAM_LOCAL_PARAMS);
       ASSIGN_4V(param, x, y, z, w);
    }
 }
@@ -639,8 +639,8 @@ _mesa_GetProgramivARB(GLenum target, GLenum pname, GLint *params)
       return;
    }
 
-   ASSERT(prog);
-   ASSERT(limits);
+   assert(prog);
+   assert(limits);
 
    /* Queries supported for both vertex and fragment programs */
    switch (pname) {
@@ -817,7 +817,7 @@ _mesa_GetProgramStringARB(GLenum target, GLenum pname, GLvoid *string)
       return;
    }
 
-   ASSERT(prog);
+   assert(prog);
 
    if (pname != GL_PROGRAM_STRING_ARB) {
       _mesa_error(ctx, GL_INVALID_ENUM, "glGetProgramStringARB(pname)");
diff --git a/mesalib/src/mesa/main/arrayobj.c b/mesalib/src/mesa/main/arrayobj.c
index 6440ea670..3c8ffb5a4 100644
--- a/mesalib/src/mesa/main/arrayobj.c
+++ b/mesalib/src/mesa/main/arrayobj.c
@@ -84,10 +84,10 @@ unbind_array_object_vbos(struct gl_context *ctx, struct gl_vertex_array_object *
 {
    GLuint i;
 
-   for (i = 0; i < Elements(obj->VertexBinding); i++)
+   for (i = 0; i < ARRAY_SIZE(obj->VertexBinding); i++)
       _mesa_reference_buffer_object(ctx, &obj->VertexBinding[i].BufferObj, NULL);
 
-   for (i = 0; i < Elements(obj->_VertexAttrib); i++)
+   for (i = 0; i < ARRAY_SIZE(obj->_VertexAttrib); i++)
       _mesa_reference_buffer_object(ctx, &obj->_VertexAttrib[i].BufferObj, NULL);
 }
 
@@ -143,7 +143,7 @@ _mesa_reference_vao_(struct gl_context *ctx,
       struct gl_vertex_array_object *oldObj = *ptr;
 
       mtx_lock(&oldObj->Mutex);
-      ASSERT(oldObj->RefCount > 0);
+      assert(oldObj->RefCount > 0);
       oldObj->RefCount--;
 #if 0
       printf("ArrayObj %p %d DECR to %d\n",
@@ -153,13 +153,13 @@ _mesa_reference_vao_(struct gl_context *ctx,
       mtx_unlock(&oldObj->Mutex);
 
       if (deleteFlag) {
-	 ASSERT(ctx->Driver.DeleteArrayObject);
+	 assert(ctx->Driver.DeleteArrayObject);
          ctx->Driver.DeleteArrayObject(ctx, oldObj);
       }
 
       *ptr = NULL;
    }
-   ASSERT(!*ptr);
+   assert(!*ptr);
 
    if (vao) {
       /* reference new array object */
@@ -230,7 +230,7 @@ _mesa_initialize_vao(struct gl_context *ctx,
    obj->RefCount = 1;
 
    /* Init the individual arrays */
-   for (i = 0; i < Elements(obj->VertexAttrib); i++) {
+   for (i = 0; i < ARRAY_SIZE(obj->VertexAttrib); i++) {
       switch (i) {
       case VERT_ATTRIB_WEIGHT:
          init_array(ctx, obj, VERT_ATTRIB_WEIGHT, 1, GL_FLOAT);
@@ -335,7 +335,7 @@ bind_vertex_array(struct gl_context *ctx, GLuint id, GLboolean genRequired)
    struct gl_vertex_array_object * const oldObj = ctx->Array.VAO;
    struct gl_vertex_array_object *newObj = NULL;
 
-   ASSERT(oldObj != NULL);
+   assert(oldObj != NULL);
 
    if ( oldObj->Name == id )
       return;   /* rebinding the same array object- no change */
@@ -456,7 +456,7 @@ _mesa_DeleteVertexArrays(GLsizei n, const GLuint *ids)
       struct gl_vertex_array_object *obj = _mesa_lookup_vao(ctx, ids[i]);
 
       if ( obj != NULL ) {
-	 ASSERT( obj->Name == ids[i] );
+	 assert( obj->Name == ids[i] );
 
 	 /* If the array object is currently bound, the spec says "the binding
 	  * for that object reverts to zero and the default vertex array
diff --git a/mesalib/src/mesa/main/atifragshader.c b/mesalib/src/mesa/main/atifragshader.c
index 1eab7731a..9d967b9e6 100644
--- a/mesalib/src/mesa/main/atifragshader.c
+++ b/mesalib/src/mesa/main/atifragshader.c
@@ -255,7 +255,7 @@ _mesa_BindFragmentShaderATI(GLuint id)
    /* do actual bind */
    ctx->ATIFragmentShader.Current = newProg;
 
-   ASSERT(ctx->ATIFragmentShader.Current);
+   assert(ctx->ATIFragmentShader.Current);
    if (newProg)
       newProg->RefCount++;
 
diff --git a/mesalib/src/mesa/main/atifragshader.h b/mesalib/src/mesa/main/atifragshader.h
index afaea009d..590113410 100644
--- a/mesalib/src/mesa/main/atifragshader.h
+++ b/mesalib/src/mesa/main/atifragshader.h
@@ -8,10 +8,9 @@
 #ifndef ATIFRAGSHADER_H
 #define ATIFRAGSHADER_H
 
-#include "compiler.h"
 #include "glheader.h"
 
-struct _glapi_table;
+
 struct gl_context;
 
 #define MAX_NUM_INSTRUCTIONS_PER_PASS_ATI 8
diff --git a/mesalib/src/mesa/main/attrib.c b/mesalib/src/mesa/main/attrib.c
index 07934b9bc..20216a87d 100644
--- a/mesalib/src/mesa/main/attrib.c
+++ b/mesalib/src/mesa/main/attrib.c
@@ -1453,7 +1453,7 @@ copy_array_object(struct gl_context *ctx,
    /* In theory must be the same anyway, but on recreate make sure it matches */
    dest->ARBsemantics = src->ARBsemantics;
 
-   for (i = 0; i < Elements(src->VertexAttrib); i++) {
+   for (i = 0; i < ARRAY_SIZE(src->VertexAttrib); i++) {
       _mesa_copy_client_array(ctx, &dest->_VertexAttrib[i], &src->_VertexAttrib[i]);
       _mesa_copy_vertex_attrib_array(ctx, &dest->VertexAttrib[i], &src->VertexAttrib[i]);
       _mesa_copy_vertex_buffer_binding(ctx, &dest->VertexBinding[i], &src->VertexBinding[i]);
diff --git a/mesalib/src/mesa/main/attrib.h b/mesalib/src/mesa/main/attrib.h
index 44413a6e8..f4fd4929f 100644
--- a/mesalib/src/mesa/main/attrib.h
+++ b/mesalib/src/mesa/main/attrib.h
@@ -26,7 +26,6 @@
 #define ATTRIB_H
 
 
-#include "compiler.h"
 #include "glheader.h"
 
 struct _glapi_table;
diff --git a/mesalib/src/mesa/main/blend.c b/mesalib/src/mesa/main/blend.c
index c37c0fea5..774fc888e 100644
--- a/mesalib/src/mesa/main/blend.c
+++ b/mesalib/src/mesa/main/blend.c
@@ -799,7 +799,7 @@ get_clamp_color(const struct gl_framebuffer *fb, GLenum clamp)
    if (clamp == GL_TRUE || clamp == GL_FALSE)
       return clamp;
 
-   ASSERT(clamp == GL_FIXED_ONLY);
+   assert(clamp == GL_FIXED_ONLY);
    if (!fb)
       return GL_TRUE;
 
@@ -896,7 +896,7 @@ void _mesa_init_color( struct gl_context * ctx )
    ctx->Color.AlphaFunc = GL_ALWAYS;
    ctx->Color.AlphaRef = 0;
    ctx->Color.BlendEnabled = 0x0;
-   for (i = 0; i < Elements(ctx->Color.Blend); i++) {
+   for (i = 0; i < ARRAY_SIZE(ctx->Color.Blend); i++) {
       ctx->Color.Blend[i].SrcRGB = GL_ONE;
       ctx->Color.Blend[i].DstRGB = GL_ZERO;
       ctx->Color.Blend[i].SrcA = GL_ONE;
diff --git a/mesalib/src/mesa/main/blit.c b/mesalib/src/mesa/main/blit.c
index b97b56479..0694466eb 100644
--- a/mesalib/src/mesa/main/blit.c
+++ b/mesalib/src/mesa/main/blit.c
@@ -28,6 +28,7 @@
  */
 
 #include <stdbool.h>
+#include <stdio.h>
 
 #include "context.h"
 #include "enums.h"
@@ -48,7 +49,7 @@ find_attachment(const struct gl_framebuffer *fb,
                 const struct gl_renderbuffer *rb)
 {
    GLuint i;
-   for (i = 0; i < Elements(fb->Attachment); i++) {
+   for (i = 0; i < ARRAY_SIZE(fb->Attachment); i++) {
       if (fb->Attachment[i].Renderbuffer == rb)
          return &fb->Attachment[i];
    }
@@ -505,7 +506,7 @@ _mesa_BlitFramebuffer(GLint srcX0, GLint srcY0, GLint srcX1, GLint srcY1,
       return;
    }
 
-   ASSERT(ctx->Driver.BlitFramebuffer);
+   assert(ctx->Driver.BlitFramebuffer);
    ctx->Driver.BlitFramebuffer(ctx, ctx->ReadBuffer, ctx->DrawBuffer,
                                srcX0, srcY0, srcX1, srcY1,
                                dstX0, dstY0, dstX1, dstY1,
diff --git a/mesalib/src/mesa/main/blit.h b/mesalib/src/mesa/main/blit.h
index 533d6e5d1..01a958af5 100644
--- a/mesalib/src/mesa/main/blit.h
+++ b/mesalib/src/mesa/main/blit.h
@@ -26,7 +26,6 @@
 #ifndef BLIT_H
 #define BLIT_H
 
-#include "compiler.h"
 #include "glheader.h"
 
 
diff --git a/mesalib/src/mesa/main/bufferobj.c b/mesalib/src/mesa/main/bufferobj.c
index b372c68f2..e1c587779 100644
--- a/mesalib/src/mesa/main/bufferobj.c
+++ b/mesalib/src/mesa/main/bufferobj.c
@@ -446,7 +446,7 @@ _mesa_reference_buffer_object_(struct gl_context *ctx,
       struct gl_buffer_object *oldObj = *ptr;
 
       mtx_lock(&oldObj->Mutex);
-      ASSERT(oldObj->RefCount > 0);
+      assert(oldObj->RefCount > 0);
       oldObj->RefCount--;
 #if 0
       printf("BufferObj %p %d DECR to %d\n",
@@ -460,18 +460,18 @@ _mesa_reference_buffer_object_(struct gl_context *ctx,
          /* some sanity checking: don't delete a buffer still in use */
 #if 0
          /* unfortunately, these tests are invalid during context tear-down */
-	 ASSERT(ctx->Array.ArrayBufferObj != bufObj);
-	 ASSERT(ctx->Array.VAO->IndexBufferObj != bufObj);
-	 ASSERT(ctx->Array.VAO->Vertex.BufferObj != bufObj);
+	 assert(ctx->Array.ArrayBufferObj != bufObj);
+	 assert(ctx->Array.VAO->IndexBufferObj != bufObj);
+	 assert(ctx->Array.VAO->Vertex.BufferObj != bufObj);
 #endif
 
-	 ASSERT(ctx->Driver.DeleteBuffer);
+	 assert(ctx->Driver.DeleteBuffer);
          ctx->Driver.DeleteBuffer(ctx, oldObj);
       }
 
       *ptr = NULL;
    }
-   ASSERT(!*ptr);
+   assert(!*ptr);
 
    if (bufObj) {
       /* reference new buffer */
@@ -614,7 +614,7 @@ _mesa_buffer_subdata( struct gl_context *ctx, GLintptrARB offset,
    (void) ctx;
 
    /* this should have been caught in _mesa_BufferSubData() */
-   ASSERT(size + offset <= bufObj->Size);
+   assert(size + offset <= bufObj->Size);
 
    if (bufObj->Data) {
       memcpy( (GLubyte *) bufObj->Data + offset, data, size );
@@ -679,7 +679,7 @@ _mesa_buffer_clear_subdata(struct gl_context *ctx,
    GLsizeiptr i;
    GLubyte *dest;
 
-   ASSERT(ctx->Driver.MapBufferRange);
+   assert(ctx->Driver.MapBufferRange);
    dest = ctx->Driver.MapBufferRange(ctx, offset, size,
                                      GL_MAP_WRITE_BIT |
                                      GL_MAP_INVALIDATE_RANGE_BIT,
@@ -908,7 +908,7 @@ _mesa_handle_bind_buffer_gen(struct gl_context *ctx,
       /* If this is a new buffer object id, or one which was generated but
        * never used before, allocate a buffer object now.
        */
-      ASSERT(ctx->Driver.NewBufferObject);
+      assert(ctx->Driver.NewBufferObject);
       buf = ctx->Driver.NewBufferObject(ctx, buffer);
       if (!buf) {
 	 _mesa_error(ctx, GL_OUT_OF_MEMORY, "%s", caller);
@@ -1124,7 +1124,7 @@ _mesa_buffer_unmap_all_mappings(struct gl_context *ctx,
    for (i = 0; i < MAP_COUNT; i++) {
       if (_mesa_bufferobj_mapped(bufObj, i)) {
          ctx->Driver.UnmapBuffer(ctx, bufObj, i);
-         ASSERT(bufObj->Mappings[i].Pointer == NULL);
+         assert(bufObj->Mappings[i].Pointer == NULL);
          bufObj->Mappings[i].AccessFlags = 0;
       }
    }
@@ -1174,12 +1174,12 @@ _mesa_DeleteBuffers(GLsizei n, const GLuint *ids)
          struct gl_vertex_array_object *vao = ctx->Array.VAO;
          GLuint j;
 
-         ASSERT(bufObj->Name == ids[i] || bufObj == &DummyBufferObject);
+         assert(bufObj->Name == ids[i] || bufObj == &DummyBufferObject);
 
          _mesa_buffer_unmap_all_mappings(ctx, bufObj);
 
          /* unbind any vertex pointers bound to this buffer */
-         for (j = 0; j < Elements(vao->VertexBinding); j++) {
+         for (j = 0; j < ARRAY_SIZE(vao->VertexBinding); j++) {
             unbind(ctx, &vao->VertexBinding[j].BufferObj, bufObj);
          }
 
@@ -1387,7 +1387,7 @@ _mesa_BufferStorage(GLenum target, GLsizeiptr size, const GLvoid *data,
    bufObj->Written = GL_TRUE;
    bufObj->Immutable = GL_TRUE;
 
-   ASSERT(ctx->Driver.BufferData);
+   assert(ctx->Driver.BufferData);
    if (!ctx->Driver.BufferData(ctx, target, size, data, GL_DYNAMIC_DRAW,
                                flags, bufObj)) {
       if (target == GL_EXTERNAL_VIRTUAL_MEMORY_BUFFER_AMD) {
@@ -1477,7 +1477,7 @@ _mesa_BufferData(GLenum target, GLsizeiptrARB size,
    size += 100;
 #endif
 
-   ASSERT(ctx->Driver.BufferData);
+   assert(ctx->Driver.BufferData);
    if (!ctx->Driver.BufferData(ctx, target, size, data, usage,
                                GL_MAP_READ_BIT |
                                GL_MAP_WRITE_BIT |
@@ -1525,7 +1525,7 @@ _mesa_BufferSubData(GLenum target, GLintptrARB offset,
 
    bufObj->Written = GL_TRUE;
 
-   ASSERT(ctx->Driver.BufferSubData);
+   assert(ctx->Driver.BufferSubData);
    ctx->Driver.BufferSubData( ctx, offset, size, data, bufObj );
 }
 
@@ -1545,7 +1545,7 @@ _mesa_GetBufferSubData(GLenum target, GLintptrARB offset,
       return;
    }
 
-   ASSERT(ctx->Driver.GetBufferSubData);
+   assert(ctx->Driver.GetBufferSubData);
    ctx->Driver.GetBufferSubData( ctx, offset, size, data, bufObj );
 }
 
@@ -1722,7 +1722,7 @@ _mesa_MapBuffer(GLenum target, GLenum access)
       return NULL;
    }
 
-   ASSERT(ctx->Driver.MapBufferRange);
+   assert(ctx->Driver.MapBufferRange);
    map = ctx->Driver.MapBufferRange(ctx, 0, bufObj->Size, accessFlags, bufObj,
                                     MAP_USER);
    if (!map) {
@@ -1734,9 +1734,9 @@ _mesa_MapBuffer(GLenum target, GLenum access)
        * This is important because other modules (like VBO) might call
        * the driver function directly.
        */
-      ASSERT(bufObj->Mappings[MAP_USER].Pointer == map);
-      ASSERT(bufObj->Mappings[MAP_USER].Length == bufObj->Size);
-      ASSERT(bufObj->Mappings[MAP_USER].Offset == 0);
+      assert(bufObj->Mappings[MAP_USER].Pointer == map);
+      assert(bufObj->Mappings[MAP_USER].Length == bufObj->Size);
+      assert(bufObj->Mappings[MAP_USER].Offset == 0);
       bufObj->Mappings[MAP_USER].AccessFlags = accessFlags;
    }
 
@@ -1824,9 +1824,9 @@ _mesa_UnmapBuffer(GLenum target)
 
    status = ctx->Driver.UnmapBuffer(ctx, bufObj, MAP_USER);
    bufObj->Mappings[MAP_USER].AccessFlags = 0;
-   ASSERT(bufObj->Mappings[MAP_USER].Pointer == NULL);
-   ASSERT(bufObj->Mappings[MAP_USER].Offset == 0);
-   ASSERT(bufObj->Mappings[MAP_USER].Length == 0);
+   assert(bufObj->Mappings[MAP_USER].Pointer == NULL);
+   assert(bufObj->Mappings[MAP_USER].Offset == 0);
+   assert(bufObj->Mappings[MAP_USER].Length == 0);
 
    return status;
 }
@@ -2203,7 +2203,7 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
       return bufObj->Mappings[MAP_USER].Pointer;
    }
 
-   ASSERT(ctx->Driver.MapBufferRange);
+   assert(ctx->Driver.MapBufferRange);
    map = ctx->Driver.MapBufferRange(ctx, offset, length, access, bufObj,
                                     MAP_USER);
    if (!map) {
@@ -2214,10 +2214,10 @@ _mesa_MapBufferRange(GLenum target, GLintptr offset, GLsizeiptr length,
        * This is important because other modules (like VBO) might call
        * the driver function directly.
        */
-      ASSERT(bufObj->Mappings[MAP_USER].Pointer == map);
-      ASSERT(bufObj->Mappings[MAP_USER].Length == length);
-      ASSERT(bufObj->Mappings[MAP_USER].Offset == offset);
-      ASSERT(bufObj->Mappings[MAP_USER].AccessFlags == access);
+      assert(bufObj->Mappings[MAP_USER].Pointer == map);
+      assert(bufObj->Mappings[MAP_USER].Length == length);
+      assert(bufObj->Mappings[MAP_USER].Offset == offset);
+      assert(bufObj->Mappings[MAP_USER].AccessFlags == access);
    }
 
    return map;
@@ -2278,7 +2278,7 @@ _mesa_FlushMappedBufferRange(GLenum target, GLintptr offset, GLsizeiptr length)
       return;
    }
 
-   ASSERT(bufObj->Mappings[MAP_USER].AccessFlags & GL_MAP_WRITE_BIT);
+   assert(bufObj->Mappings[MAP_USER].AccessFlags & GL_MAP_WRITE_BIT);
 
    if (ctx->Driver.FlushMappedBufferRange)
       ctx->Driver.FlushMappedBufferRange(ctx, offset, length, bufObj,
diff --git a/mesalib/src/mesa/main/buffers.c b/mesalib/src/mesa/main/buffers.c
index e5076e9bb..37a979092 100644
--- a/mesalib/src/mesa/main/buffers.c
+++ b/mesalib/src/mesa/main/buffers.c
@@ -502,7 +502,7 @@ _mesa_drawbuffers(struct gl_context *ctx, GLuint n, const GLenum *buffers,
       GLuint output;
       for (output = 0; output < n; output++) {
          mask[output] = draw_buffer_enum_to_bitmask(ctx, buffers[output]);
-         ASSERT(mask[output] != BAD_MASK);
+         assert(mask[output] != BAD_MASK);
          mask[output] &= supportedMask;
       }
       destMask = mask;
@@ -533,7 +533,7 @@ _mesa_drawbuffers(struct gl_context *ctx, GLuint n, const GLenum *buffers,
          if (destMask[buf]) {
             GLint bufIndex = ffs(destMask[buf]) - 1;
             /* only one bit should be set in the destMask[buf] field */
-            ASSERT(_mesa_bitcount(destMask[buf]) == 1);
+            assert(_mesa_bitcount(destMask[buf]) == 1);
             if (fb->_ColorDrawBufferIndexes[buf] != bufIndex) {
 	       updated_drawbuffers(ctx);
                fb->_ColorDrawBufferIndexes[buf] = bufIndex;
diff --git a/mesalib/src/mesa/main/clear.c b/mesalib/src/mesa/main/clear.c
index 3c4ced8ed..8d707bc34 100644
--- a/mesalib/src/mesa/main/clear.c
+++ b/mesalib/src/mesa/main/clear.c
@@ -218,7 +218,7 @@ _mesa_Clear( GLbitfield mask )
          bufferMask |= BUFFER_BIT_ACCUM;
       }
 
-      ASSERT(ctx->Driver.Clear);
+      assert(ctx->Driver.Clear);
       ctx->Driver.Clear(ctx, bufferMask);
    }
 }
diff --git a/mesalib/src/mesa/main/colortab.h b/mesalib/src/mesa/main/colortab.h
index 55c3b59fa..e1165d702 100644
--- a/mesalib/src/mesa/main/colortab.h
+++ b/mesalib/src/mesa/main/colortab.h
@@ -26,8 +26,6 @@
 #ifndef COLORTAB_H
 #define COLORTAB_H
 
-
-#include "compiler.h"
 #include "glheader.h"
 
 struct _glapi_table;
diff --git a/mesalib/src/mesa/main/compiler.h b/mesalib/src/mesa/main/compiler.h
index cdc843db2..95581fb49 100644
--- a/mesalib/src/mesa/main/compiler.h
+++ b/mesalib/src/mesa/main/compiler.h
@@ -35,14 +35,6 @@
 
 
 #include <assert.h>
-#include <ctype.h>
-#include <math.h>
-#include <limits.h>
-#include <stdlib.h>
-#include <stdio.h>
-#include <string.h>
-#include <float.h>
-#include <stdarg.h>
 
 #include "util/macros.h"
 
@@ -55,12 +47,6 @@ extern "C" {
 
 
 /**
- * Get standard integer types
- */
-#include <stdint.h>
-
-
-/**
   * Sun compilers define __i386 instead of the gcc-style __i386__
  */
 #ifdef __SUNPRO_C
@@ -78,14 +64,6 @@ extern "C" {
 
 
 /**
- * finite macro.
- */
-#if defined(_MSC_VER)
-#  define finite _finite
-#endif
-
-
-/**
  * Disable assorted warnings
  */
 #if defined(_WIN32) && !defined(__CYGWIN__)
@@ -105,33 +83,6 @@ extern "C" {
 #endif
 
 
-
-/* XXX: Use standard `inline` keyword instead */
-#ifndef INLINE
-#  define INLINE inline
-#endif
-
-
-/**
- * PUBLIC/USED macros
- *
- * If we build the library with gcc's -fvisibility=hidden flag, we'll
- * use the PUBLIC macro to mark functions that are to be exported.
- *
- * We also need to define a USED attribute, so the optimizer doesn't 
- * inline a static function that we later use in an alias. - ajax
- */
-#ifndef PUBLIC
-#  if defined(__GNUC__) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x590))
-#    define PUBLIC __attribute__((visibility("default")))
-#    define USED __attribute__((used))
-#  else
-#    define PUBLIC
-#    define USED
-#  endif
-#endif
-
-
 /* XXX: Use standard `__func__` instead */
 #ifndef __FUNCTION__
 #  define __FUNCTION__ __func__
@@ -174,11 +125,6 @@ static inline GLuint CPU_TO_LE32(GLuint x)
 
 
 
-#if !defined(CAPI) && defined(_WIN32)
-#define CAPI _cdecl
-#endif
-
-
 /**
  * Create a macro so that asm functions can be linked into compilers other
  * than GNU C
@@ -196,41 +142,6 @@ static inline GLuint CPU_TO_LE32(GLuint x)
 #endif
 #endif
 
-#ifdef USE_X86_ASM
-#define _NORMAPI _ASMAPI
-#define _NORMAPIP _ASMAPIP
-#else
-#define _NORMAPI
-#define _NORMAPIP *
-#endif
-
-
-/* Turn off macro checking systems used by other libraries */
-#ifdef CHECK
-#undef CHECK
-#endif
-
-
-/**
- * ASSERT macro
- */
-#if defined(DEBUG)
-#  define ASSERT(X)   assert(X)
-#else
-#  define ASSERT(X)
-#endif
-
-
-/*
- * A trick to suppress uninitialized variable warning without generating any
- * code
- */
-#define uninitialized_var(x) x = x
-
-#ifndef NULL
-#define NULL 0
-#endif
-
 
 /**
  * LONGSTRING macro
@@ -242,111 +153,8 @@ static inline GLuint CPU_TO_LE32(GLuint x)
 # define LONGSTRING __extension__
 #endif
 
-
-#ifndef M_PI
-#define M_PI (3.14159265358979323846)
-#endif
-
-#ifndef M_E
-#define M_E (2.7182818284590452354)
-#endif
-
-#ifndef M_LOG2E
-#define M_LOG2E     (1.4426950408889634074)
-#endif
-
-#ifndef ONE_DIV_SQRT_LN2
-#define ONE_DIV_SQRT_LN2 (1.201122408786449815)
-#endif
-
-#ifndef FLT_MAX_EXP
-#define FLT_MAX_EXP 128
-#endif
-
 #define IEEE_ONE 0x3f800000
 
-/**
- * START/END_FAST_MATH macros:
- *
- * START_FAST_MATH: Set x86 FPU to faster, 32-bit precision mode (and save
- *                  original mode to a temporary).
- * END_FAST_MATH: Restore x86 FPU to original mode.
- */
-#if defined(__GNUC__) && defined(__i386__)
-/*
- * Set the x86 FPU control word to guarentee only 32 bits of precision
- * are stored in registers.  Allowing the FPU to store more introduces
- * differences between situations where numbers are pulled out of memory
- * vs. situations where the compiler is able to optimize register usage.
- *
- * In the worst case, we force the compiler to use a memory access to
- * truncate the float, by specifying the 'volatile' keyword.
- */
-/* Hardware default: All exceptions masked, extended double precision,
- * round to nearest (IEEE compliant):
- */
-#define DEFAULT_X86_FPU		0x037f
-/* All exceptions masked, single precision, round to nearest:
- */
-#define FAST_X86_FPU		0x003f
-/* The fldcw instruction will cause any pending FP exceptions to be
- * raised prior to entering the block, and we clear any pending
- * exceptions before exiting the block.  Hence, asm code has free
- * reign over the FPU while in the fast math block.
- */
-#if defined(NO_FAST_MATH)
-#define START_FAST_MATH(x)						\
-do {									\
-   static GLuint mask = DEFAULT_X86_FPU;				\
-   __asm__ ( "fnstcw %0" : "=m" (*&(x)) );				\
-   __asm__ ( "fldcw %0" : : "m" (mask) );				\
-} while (0)
-#else
-#define START_FAST_MATH(x)						\
-do {									\
-   static GLuint mask = FAST_X86_FPU;					\
-   __asm__ ( "fnstcw %0" : "=m" (*&(x)) );				\
-   __asm__ ( "fldcw %0" : : "m" (mask) );				\
-} while (0)
-#endif
-/* Restore original FPU mode, and clear any exceptions that may have
- * occurred in the FAST_MATH block.
- */
-#define END_FAST_MATH(x)						\
-do {									\
-   __asm__ ( "fnclex ; fldcw %0" : : "m" (*&(x)) );			\
-} while (0)
-
-#elif defined(_MSC_VER) && defined(_M_IX86)
-#define DEFAULT_X86_FPU		0x037f /* See GCC comments above */
-#define FAST_X86_FPU		0x003f /* See GCC comments above */
-#if defined(NO_FAST_MATH)
-#define START_FAST_MATH(x) do {\
-	static GLuint mask = DEFAULT_X86_FPU;\
-	__asm fnstcw word ptr [x]\
-	__asm fldcw word ptr [mask]\
-} while(0)
-#else
-#define START_FAST_MATH(x) do {\
-	static GLuint mask = FAST_X86_FPU;\
-	__asm fnstcw word ptr [x]\
-	__asm fldcw word ptr [mask]\
-} while(0)
-#endif
-#define END_FAST_MATH(x) do {\
-	__asm fnclex\
-	__asm fldcw word ptr [x]\
-} while(0)
-
-#else
-#define START_FAST_MATH(x)  x = 0
-#define END_FAST_MATH(x)  (void)(x)
-#endif
-
-
-#ifndef Elements
-#define Elements(x) (sizeof(x)/sizeof(*(x)))
-#endif
 
 #ifdef __cplusplus
 }
diff --git a/mesalib/src/mesa/main/condrender.c b/mesalib/src/mesa/main/condrender.c
index 75f9d74bc..77e4b95ee 100644
--- a/mesalib/src/mesa/main/condrender.c
+++ b/mesalib/src/mesa/main/condrender.c
@@ -55,7 +55,7 @@ _mesa_BeginConditionalRender(GLuint queryId, GLenum mode)
       return;
    }
 
-   ASSERT(ctx->Query.CondRenderMode == GL_NONE);
+   assert(ctx->Query.CondRenderMode == GL_NONE);
 
    /* Section 2.14 (Conditional Rendering) of the OpenGL 3.0 spec says:
     *
@@ -70,7 +70,7 @@ _mesa_BeginConditionalRender(GLuint queryId, GLenum mode)
                   "glBeginConditionalRender(bad queryId=%u)", queryId);
       return;
    }
-   ASSERT(q->Id == queryId);
+   assert(q->Id == queryId);
 
    switch (mode) {
    case GL_QUERY_WAIT:
diff --git a/mesalib/src/mesa/main/context.c b/mesalib/src/mesa/main/context.c
index b186a1fad..22c2341d6 100644
--- a/mesalib/src/mesa/main/context.c
+++ b/mesalib/src/mesa/main/context.c
@@ -134,6 +134,7 @@
 #include "math/m_matrix.h"
 #include "main/dispatch.h" /* for _gloffset_COUNT */
 #include "uniforms.h"
+#include "macros.h"
 
 #ifdef USE_SPARC_ASM
 #include "sparc/sparc.h"
@@ -446,7 +447,7 @@ _mesa_init_current(struct gl_context *ctx)
    GLuint i;
 
    /* Init all to (0,0,0,1) */
-   for (i = 0; i < Elements(ctx->Current.Attrib); i++) {
+   for (i = 0; i < ARRAY_SIZE(ctx->Current.Attrib); i++) {
       ASSIGN_4V( ctx->Current.Attrib[i], 0.0, 0.0, 0.0, 1.0 );
    }
 
@@ -656,7 +657,7 @@ _mesa_init_constants(struct gl_constants *consts, gl_api api)
    consts->MaxSamples = 0;
 
    /* GLSL default if NativeIntegers == FALSE */
-   consts->UniformBooleanTrue = FLT_AS_UINT(1.0f);
+   consts->UniformBooleanTrue = FLOAT_AS_UNION(1.0f).u;
 
    /* GL_ARB_sync */
    consts->MaxServerWaitTimeout = 0x1fff7fffffffULL;
@@ -1644,7 +1645,7 @@ _mesa_make_current( struct gl_context *newCtx,
 
    /* We used to call _glapi_check_multithread() here.  Now do it in drivers */
    _glapi_set_context((void *) newCtx);
-   ASSERT(_mesa_get_current_context() == newCtx);
+   assert(_mesa_get_current_context() == newCtx);
 
    if (!newCtx) {
       _glapi_set_dispatch(NULL);  /* none current */
@@ -1653,8 +1654,8 @@ _mesa_make_current( struct gl_context *newCtx,
       _glapi_set_dispatch(newCtx->CurrentDispatch);
 
       if (drawBuffer && readBuffer) {
-         ASSERT(_mesa_is_winsys_fbo(drawBuffer));
-         ASSERT(_mesa_is_winsys_fbo(readBuffer));
+         assert(_mesa_is_winsys_fbo(drawBuffer));
+         assert(_mesa_is_winsys_fbo(readBuffer));
          _mesa_reference_framebuffer(&newCtx->WinSysDrawBuffer, drawBuffer);
          _mesa_reference_framebuffer(&newCtx->WinSysReadBuffer, readBuffer);
 
diff --git a/mesalib/src/mesa/main/convolve.h b/mesalib/src/mesa/main/convolve.h
index 0c0a9dd5e..e696ce497 100644
--- a/mesalib/src/mesa/main/convolve.h
+++ b/mesalib/src/mesa/main/convolve.h
@@ -28,7 +28,7 @@
 #define CONVOLVE_H
 
 
-#include "compiler.h"
+#include "glheader.h"
 
 struct _glapi_table;
 
diff --git a/mesalib/src/mesa/main/cpuinfo.c b/mesalib/src/mesa/main/cpuinfo.c
index 0755d6b21..67f22ab9c 100644
--- a/mesalib/src/mesa/main/cpuinfo.c
+++ b/mesalib/src/mesa/main/cpuinfo.c
@@ -23,7 +23,7 @@
  */
 
 
-#include "main/compiler.h"
+#include "main/imports.h"
 #include "main/cpuinfo.h"
 
 
diff --git a/mesalib/src/mesa/main/dd.h b/mesalib/src/mesa/main/dd.h
index ec8662b30..0c1a13fa4 100644
--- a/mesalib/src/mesa/main/dd.h
+++ b/mesalib/src/mesa/main/dd.h
@@ -824,7 +824,7 @@ struct dd_function_table {
    void (*BeginVertices)( struct gl_context *ctx );
 
    /**
-    * If inside glBegin()/glEnd(), it should ASSERT(0).  Otherwise, if
+    * If inside glBegin()/glEnd(), it should assert(0).  Otherwise, if
     * FLUSH_STORED_VERTICES bit in \p flags is set flushes any buffered
     * vertices, if FLUSH_UPDATE_CURRENT bit is set updates
     * __struct gl_contextRec::Current and gl_light_attrib::Material
diff --git a/mesalib/src/mesa/main/debug.c b/mesalib/src/mesa/main/debug.c
index 089ce8987..b50d79e0c 100644
--- a/mesalib/src/mesa/main/debug.c
+++ b/mesalib/src/mesa/main/debug.c
@@ -23,6 +23,7 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
+#include <stdio.h>
 #include "mtypes.h"
 #include "attrib.h"
 #include "colormac.h"
@@ -58,8 +59,8 @@ tex_target_name(GLenum tgt)
       { GL_TEXTURE_EXTERNAL_OES, "GL_TEXTURE_EXTERNAL_OES" }
    };
    GLuint i;
-   STATIC_ASSERT(Elements(tex_targets) == NUM_TEXTURE_TARGETS);
-   for (i = 0; i < Elements(tex_targets); i++) {
+   STATIC_ASSERT(ARRAY_SIZE(tex_targets) == NUM_TEXTURE_TARGETS);
+   for (i = 0; i < ARRAY_SIZE(tex_targets); i++) {
       if (tex_targets[i].target == tgt)
          return tex_targets[i].name;
    }
@@ -117,11 +118,6 @@ void _mesa_print_info( struct gl_context *ctx )
     */
    _mesa_debug(NULL, "Mesa GL_EXTENSIONS = %s\n", ctx->Extensions.String);
 
-#if defined(THREADS)
-   _mesa_debug(NULL, "Mesa thread-safe: YES\n");
-#else
-   _mesa_debug(NULL, "Mesa thread-safe: NO\n");
-#endif
 #if defined(USE_X86_ASM)
    _mesa_debug(NULL, "Mesa x86-optimized: YES\n");
 #else
@@ -168,7 +164,7 @@ set_verbose_flags(const char *str)
       return;
 
    MESA_VERBOSE = 0x0;
-   for (i = 0; i < Elements(opts); i++) {
+   for (i = 0; i < ARRAY_SIZE(opts); i++) {
       if (strstr(str, opts[i].name) || strcmp(str, "all") == 0)
          MESA_VERBOSE |= opts[i].flag;
    }
@@ -201,7 +197,7 @@ set_debug_flags(const char *str)
       return;
 
    MESA_DEBUG_FLAGS = 0x0;
-   for (i = 0; i < Elements(opts); i++) {
+   for (i = 0; i < ARRAY_SIZE(opts); i++) {
       if (strstr(str, opts[i].name))
          MESA_DEBUG_FLAGS |= opts[i].flag;
    }
diff --git a/mesalib/src/mesa/main/dlist.c b/mesalib/src/mesa/main/dlist.c
index 025f6abd2..431c4b48b 100644
--- a/mesalib/src/mesa/main/dlist.c
+++ b/mesalib/src/mesa/main/dlist.c
@@ -29,6 +29,7 @@
  * Display lists management functions.
  */
 
+#include "c99_math.h"
 #include "glheader.h"
 #include "imports.h"
 #include "api_arrayelt.h"
@@ -912,7 +913,7 @@ translate_id(GLsizei n, GLenum type, const GLvoid * list)
       return (GLint) uiptr[n];
    case GL_FLOAT:
       fptr = (GLfloat *) list;
-      return (GLint) FLOORF(fptr[n]);
+      return (GLint) floorf(fptr[n]);
    case GL_2_BYTES:
       ubptr = ((GLubyte *) list) + 2 * n;
       return (GLint) ubptr[0] * 256
@@ -1035,7 +1036,7 @@ dlist_alloc(struct gl_context *ctx, OpCode opcode, GLuint bytes, bool align8)
       }
       else {
          /* make sure instruction size agrees */
-         ASSERT(numNodes == InstSize[opcode]);
+         assert(numNodes == InstSize[opcode]);
       }
    }
 
@@ -5089,7 +5090,7 @@ save_Attr1fNV(GLenum attr, GLfloat x)
       n[2].f = x;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 1;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, 0, 0, 1);
 
@@ -5111,7 +5112,7 @@ save_Attr2fNV(GLenum attr, GLfloat x, GLfloat y)
       n[3].f = y;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 2;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, 0, 1);
 
@@ -5134,7 +5135,7 @@ save_Attr3fNV(GLenum attr, GLfloat x, GLfloat y, GLfloat z)
       n[4].f = z;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 3;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, 1);
 
@@ -5158,7 +5159,7 @@ save_Attr4fNV(GLenum attr, GLfloat x, GLfloat y, GLfloat z, GLfloat w)
       n[5].f = w;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 4;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, w);
 
@@ -5180,7 +5181,7 @@ save_Attr1fARB(GLenum attr, GLfloat x)
       n[2].f = x;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 1;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, 0, 0, 1);
 
@@ -5202,7 +5203,7 @@ save_Attr2fARB(GLenum attr, GLfloat x, GLfloat y)
       n[3].f = y;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 2;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, 0, 1);
 
@@ -5225,7 +5226,7 @@ save_Attr3fARB(GLenum attr, GLfloat x, GLfloat y, GLfloat z)
       n[4].f = z;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 3;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, 1);
 
@@ -5249,7 +5250,7 @@ save_Attr4fARB(GLenum attr, GLfloat x, GLfloat y, GLfloat z, GLfloat w)
       n[5].f = w;
    }
 
-   ASSERT(attr < MAX_VERTEX_GENERIC_ATTRIBS);
+   assert(attr < MAX_VERTEX_GENERIC_ATTRIBS);
    ctx->ListState.ActiveAttribSize[attr] = 4;
    ASSIGN_4V(ctx->ListState.CurrentAttrib[attr], x, y, z, w);
 
diff --git a/mesalib/src/mesa/main/dlist.h b/mesalib/src/mesa/main/dlist.h
index 6189632d4..a1214674c 100644
--- a/mesalib/src/mesa/main/dlist.h
+++ b/mesalib/src/mesa/main/dlist.h
@@ -32,7 +32,7 @@
 #ifndef DLIST_H
 #define DLIST_H
 
-
+#include <stdio.h>
 #include "main/mtypes.h"
 
 
diff --git a/mesalib/src/mesa/main/drawpix.c b/mesalib/src/mesa/main/drawpix.c
index 1865a66b9..55035f214 100644
--- a/mesalib/src/mesa/main/drawpix.c
+++ b/mesalib/src/mesa/main/drawpix.c
@@ -173,7 +173,7 @@ _mesa_DrawPixels( GLsizei width, GLsizei height,
                              ctx->Current.RasterTexCoords[0] );
    }
    else {
-      ASSERT(ctx->RenderMode == GL_SELECT);
+      assert(ctx->RenderMode == GL_SELECT);
       /* Do nothing.  See OpenGL Spec, Appendix B, Corollary 6. */
    }
 
@@ -279,7 +279,7 @@ _mesa_CopyPixels( GLint srcx, GLint srcy, GLsizei width, GLsizei height,
                              ctx->Current.RasterTexCoords[0] );
    }
    else {
-      ASSERT(ctx->RenderMode == GL_SELECT);
+      assert(ctx->RenderMode == GL_SELECT);
       /* Do nothing.  See OpenGL Spec, Appendix B, Corollary 6. */
    }
 
@@ -355,7 +355,7 @@ _mesa_Bitmap( GLsizei width, GLsizei height,
                              ctx->Current.RasterTexCoords[0] );
    }
    else {
-      ASSERT(ctx->RenderMode == GL_SELECT);
+      assert(ctx->RenderMode == GL_SELECT);
       /* Do nothing.  See OpenGL Spec, Appendix B, Corollary 6. */
    }
 
diff --git a/mesalib/src/mesa/main/drawpix.h b/mesalib/src/mesa/main/drawpix.h
index eb2cb8922..181c05ca5 100644
--- a/mesalib/src/mesa/main/drawpix.h
+++ b/mesalib/src/mesa/main/drawpix.h
@@ -26,9 +26,7 @@
 #define DRAWPIX_H
 
 
-#include "compiler.h"
-
-struct _glapi_table;
+#include "glheader.h"
 
 
 void GLAPIENTRY
diff --git a/mesalib/src/mesa/main/drawtex.c b/mesalib/src/mesa/main/drawtex.c
index 7c0ec832b..9c4fdf90e 100644
--- a/mesalib/src/mesa/main/drawtex.c
+++ b/mesalib/src/mesa/main/drawtex.c
@@ -46,7 +46,7 @@ draw_texture(struct gl_context *ctx, GLfloat x, GLfloat y, GLfloat z,
    if (ctx->NewState)
       _mesa_update_state(ctx);
 
-   ASSERT(ctx->Driver.DrawTex);
+   assert(ctx->Driver.DrawTex);
    ctx->Driver.DrawTex(ctx, x, y, z, width, height);
 
    _mesa_set_vp_override(ctx, GL_FALSE);
diff --git a/mesalib/src/mesa/main/enable.c b/mesalib/src/mesa/main/enable.c
index 11365ecc4..9008a3863 100644
--- a/mesalib/src/mesa/main/enable.c
+++ b/mesalib/src/mesa/main/enable.c
@@ -1060,7 +1060,7 @@ void
 _mesa_set_enablei(struct gl_context *ctx, GLenum cap,
                   GLuint index, GLboolean state)
 {
-   ASSERT(state == 0 || state == 1);
+   assert(state == 0 || state == 1);
    switch (cap) {
    case GL_BLEND:
       if (!ctx->Extensions.EXT_draw_buffers2) {
diff --git a/mesalib/src/mesa/main/errors.c b/mesalib/src/mesa/main/errors.c
index 478e4ed33..8ffbf413c 100644
--- a/mesalib/src/mesa/main/errors.c
+++ b/mesalib/src/mesa/main/errors.c
@@ -28,6 +28,8 @@
  */
 
 
+#include <stdarg.h>
+#include <stdio.h>
 #include "errors.h"
 #include "enums.h"
 #include "imports.h"
@@ -136,7 +138,7 @@ gl_enum_to_debug_source(GLenum e)
 {
    unsigned i;
 
-   for (i = 0; i < Elements(debug_source_enums); i++) {
+   for (i = 0; i < ARRAY_SIZE(debug_source_enums); i++) {
       if (debug_source_enums[i] == e)
          break;
    }
@@ -148,7 +150,7 @@ gl_enum_to_debug_type(GLenum e)
 {
    unsigned i;
 
-   for (i = 0; i < Elements(debug_type_enums); i++) {
+   for (i = 0; i < ARRAY_SIZE(debug_type_enums); i++) {
       if (debug_type_enums[i] == e)
          break;
    }
@@ -160,7 +162,7 @@ gl_enum_to_debug_severity(GLenum e)
 {
    unsigned i;
 
-   for (i = 0; i < Elements(debug_severity_enums); i++) {
+   for (i = 0; i < ARRAY_SIZE(debug_severity_enums); i++) {
       if (debug_severity_enums[i] == e)
          break;
    }
@@ -1465,7 +1467,7 @@ _mesa_error( struct gl_context *ctx, GLenum error, const char *fmtString, ... )
          /* Too long error message. Whoever calls _mesa_error should use
           * shorter strings.
           */
-         ASSERT(0);
+         assert(0);
          return;
       }
 
@@ -1473,7 +1475,7 @@ _mesa_error( struct gl_context *ctx, GLenum error, const char *fmtString, ... )
                            _mesa_lookup_enum_by_nr(error), s);
       if (len >= MAX_DEBUG_MESSAGE_LENGTH) {
          /* Same as above. */
-         ASSERT(0);
+         assert(0);
          return;
       }
 
diff --git a/mesalib/src/mesa/main/es1_conversion.c b/mesalib/src/mesa/main/es1_conversion.c
index 60a60e287..b254a6ef1 100644
--- a/mesalib/src/mesa/main/es1_conversion.c
+++ b/mesalib/src/mesa/main/es1_conversion.c
@@ -55,7 +55,7 @@ _mesa_ClipPlanef(GLenum plane, const GLfloat *equation)
    unsigned int i;
    GLdouble converted_equation[4];
 
-   for (i = 0; i < Elements(converted_equation); i++) {
+   for (i = 0; i < ARRAY_SIZE(converted_equation); i++) {
       converted_equation[i] = (GLdouble) (equation[i]);
    }
 
@@ -68,7 +68,7 @@ _mesa_ClipPlanex(GLenum plane, const GLfixed *equation)
    unsigned int i;
    GLdouble converted_equation[4];
 
-   for (i = 0; i < Elements(converted_equation); i++) {
+   for (i = 0; i < ARRAY_SIZE(converted_equation); i++) {
       converted_equation[i] = (GLdouble) (equation[i] / 65536.0);
    }
 
@@ -117,7 +117,7 @@ _mesa_DrawTexxvOES(const GLfixed *coords)
     unsigned int i;
     GLfloat converted_coords[5];
 
-    for (i = 0; i < Elements(converted_coords); i++) {
+    for (i = 0; i < ARRAY_SIZE(converted_coords); i++) {
         converted_coords[i] = (GLfloat) (coords[i] / 65536.0f);
     }
 
@@ -206,7 +206,7 @@ _mesa_GetClipPlanef(GLenum plane, GLfloat *equation)
    GLdouble converted_equation[4];
 
    _mesa_GetClipPlane(plane, converted_equation);
-   for (i = 0; i < Elements(converted_equation); i++) {
+   for (i = 0; i < ARRAY_SIZE(converted_equation); i++) {
       equation[i] = (GLfloat) (converted_equation[i]);
    }
 }
@@ -218,7 +218,7 @@ _mesa_GetClipPlanex(GLenum plane, GLfixed *equation)
    GLdouble converted_equation[4];
 
    _mesa_GetClipPlane(plane, converted_equation);
-   for (i = 0; i < Elements(converted_equation); i++) {
+   for (i = 0; i < ARRAY_SIZE(converted_equation); i++) {
       equation[i] = (GLfixed) (converted_equation[i] * 65536);
    }
 }
@@ -549,7 +549,7 @@ _mesa_LoadMatrixx(const GLfixed *m)
    unsigned int i;
    GLfloat converted_m[16];
 
-   for (i = 0; i < Elements(converted_m); i++) {
+   for (i = 0; i < ARRAY_SIZE(converted_m); i++) {
       converted_m[i] = (GLfloat) (m[i] / 65536.0f);
    }
 
@@ -617,7 +617,7 @@ _mesa_MultMatrixx(const GLfixed *m)
    unsigned int i;
    GLfloat converted_m[16];
 
-   for (i = 0; i < Elements(converted_m); i++) {
+   for (i = 0; i < ARRAY_SIZE(converted_m); i++) {
       converted_m[i] = (GLfloat) (m[i] / 65536.0f);
    }
 
@@ -825,7 +825,7 @@ _mesa_TexEnvxv(GLenum target, GLenum pname, const GLfixed *params)
       unsigned int i;
       GLfloat converted_params[4];
 
-      for (i = 0; i < Elements(converted_params); i++) {
+      for (i = 0; i < ARRAY_SIZE(converted_params); i++) {
          converted_params[i] = (GLfloat) (params[i] / 65536.0f);
       }
 
diff --git a/mesalib/src/mesa/main/eval.c b/mesalib/src/mesa/main/eval.c
index 84cd7563b..86c8f75a1 100644
--- a/mesalib/src/mesa/main/eval.c
+++ b/mesalib/src/mesa/main/eval.c
@@ -309,7 +309,7 @@ map1(GLenum target, GLfloat u1, GLfloat u2, GLint ustride,
    GLfloat *pnts;
    struct gl_1d_map *map = NULL;
 
-   ASSERT(type == GL_FLOAT || type == GL_DOUBLE);
+   assert(type == GL_FLOAT || type == GL_DOUBLE);
 
    if (u1 == u2) {
       _mesa_error( ctx, GL_INVALID_VALUE, "glMap1(u1,u2)" );
@@ -391,7 +391,7 @@ map2( GLenum target, GLfloat u1, GLfloat u2, GLint ustride, GLint uorder,
    GLfloat *pnts;
    struct gl_2d_map *map = NULL;
 
-   ASSERT(type == GL_FLOAT || type == GL_DOUBLE);
+   assert(type == GL_FLOAT || type == GL_DOUBLE);
 
    if (u1==u2) {
       _mesa_error( ctx, GL_INVALID_VALUE, "glMap2(u1,u2)" );
@@ -505,7 +505,7 @@ _mesa_GetnMapdvARB( GLenum target, GLenum query, GLsizei bufSize, GLdouble *v )
 
    map1d = get_1d_map(ctx, target);
    map2d = get_2d_map(ctx, target);
-   ASSERT(map1d || map2d);
+   assert(map1d || map2d);
 
    switch (query) {
       case GL_COEFF:
@@ -595,7 +595,7 @@ _mesa_GetnMapfvARB( GLenum target, GLenum query, GLsizei bufSize, GLfloat *v )
 
    map1d = get_1d_map(ctx, target);
    map2d = get_2d_map(ctx, target);
-   ASSERT(map1d || map2d);
+   assert(map1d || map2d);
 
    switch (query) {
       case GL_COEFF:
@@ -687,7 +687,7 @@ _mesa_GetnMapivARB( GLenum target, GLenum query, GLsizei bufSize, GLint *v )
 
    map1d = get_1d_map(ctx, target);
    map2d = get_2d_map(ctx, target);
-   ASSERT(map1d || map2d);
+   assert(map1d || map2d);
 
    switch (query) {
       case GL_COEFF:
diff --git a/mesalib/src/mesa/main/execmem.c b/mesalib/src/mesa/main/execmem.c
index 7267cf85f..3a1338515 100644
--- a/mesalib/src/mesa/main/execmem.c
+++ b/mesalib/src/mesa/main/execmem.c
@@ -31,6 +31,7 @@
  */
 
 
+#include <stdio.h>
 #include "imports.h"
 
 
diff --git a/mesalib/src/mesa/main/fbobject.c b/mesalib/src/mesa/main/fbobject.c
index 305362297..f8d0d92bf 100644
--- a/mesalib/src/mesa/main/fbobject.c
+++ b/mesalib/src/mesa/main/fbobject.c
@@ -324,14 +324,14 @@ remove_attachment(struct gl_context *ctx,
       ctx->Driver.FinishRenderTexture(ctx, rb);
 
    if (att->Type == GL_TEXTURE) {
-      ASSERT(att->Texture);
+      assert(att->Texture);
       _mesa_reference_texobj(&att->Texture, NULL); /* unbind */
-      ASSERT(!att->Texture);
+      assert(!att->Texture);
    }
    if (att->Type == GL_TEXTURE || att->Type == GL_RENDERBUFFER_EXT) {
-      ASSERT(!att->Texture);
+      assert(!att->Texture);
       _mesa_reference_renderbuffer(&att->Renderbuffer, NULL); /* unbind */
-      ASSERT(!att->Renderbuffer);
+      assert(!att->Renderbuffer);
    }
    att->Type = GL_NONE;
    att->Complete = GL_TRUE;
@@ -433,7 +433,7 @@ set_texture_attachment(struct gl_context *ctx,
 
    if (att->Texture == texObj) {
       /* re-attaching same texture */
-      ASSERT(att->Type == GL_TEXTURE);
+      assert(att->Type == GL_TEXTURE);
    }
    else {
       /* new attachment */
@@ -488,7 +488,7 @@ _mesa_framebuffer_renderbuffer(struct gl_context *ctx,
    mtx_lock(&fb->Mutex);
 
    att = get_attachment(ctx, fb, attachment);
-   ASSERT(att);
+   assert(att);
    if (rb) {
       set_renderbuffer_attachment(ctx, att, rb);
       if (attachment == GL_DEPTH_STENCIL_ATTACHMENT) {
@@ -809,7 +809,7 @@ test_attachment_completeness(const struct gl_context *ctx, GLenum format,
          }
       }
       else {
-         ASSERT(format == GL_STENCIL);
+         assert(format == GL_STENCIL);
          if (ctx->Extensions.ARB_depth_texture &&
              baseFormat == GL_DEPTH_STENCIL) {
             /* OK */
@@ -826,7 +826,7 @@ test_attachment_completeness(const struct gl_context *ctx, GLenum format,
       const GLenum baseFormat =
          _mesa_get_format_base_format(att->Renderbuffer->Format);
 
-      ASSERT(att->Renderbuffer);
+      assert(att->Renderbuffer);
       if (!att->Renderbuffer->InternalFormat ||
           att->Renderbuffer->Width < 1 ||
           att->Renderbuffer->Height < 1) {
@@ -868,7 +868,7 @@ test_attachment_completeness(const struct gl_context *ctx, GLenum format,
       }
    }
    else {
-      ASSERT(att->Type == GL_NONE);
+      assert(att->Type == GL_NONE);
       /* complete */
       return;
    }
@@ -1239,7 +1239,7 @@ bind_renderbuffer(GLenum target, GLuint renderbuffer, bool allow_user_names)
 	    _mesa_error(ctx, GL_OUT_OF_MEMORY, "glBindRenderbufferEXT");
 	    return;
 	 }
-         ASSERT(newRb->AllocStorage);
+         assert(newRb->AllocStorage);
          _mesa_HashInsert(ctx->Shared->RenderBuffers, renderbuffer, newRb);
          newRb->RefCount = 1; /* referenced by hash table */
       }
@@ -1248,7 +1248,7 @@ bind_renderbuffer(GLenum target, GLuint renderbuffer, bool allow_user_names)
       newRb = NULL;
    }
 
-   ASSERT(newRb != &DummyRenderbuffer);
+   assert(newRb != &DummyRenderbuffer);
 
    _mesa_reference_renderbuffer(&ctx->CurrentRenderbuffer, newRb);
 }
@@ -1339,7 +1339,7 @@ _mesa_DeleteRenderbuffers(GLsizei n, const GLuint *renderbuffers)
             /* check if deleting currently bound renderbuffer object */
             if (rb == ctx->CurrentRenderbuffer) {
                /* bind default */
-               ASSERT(rb->RefCount >= 2);
+               assert(rb->RefCount >= 2);
                _mesa_BindRenderbuffer(GL_RENDERBUFFER_EXT, 0);
             }
 
@@ -1838,7 +1838,7 @@ renderbuffer_storage(GLenum target, GLenum internalFormat,
    rb->NumSamples = samples;
 
    /* Now allocate the storage */
-   ASSERT(rb->AllocStorage);
+   assert(rb->AllocStorage);
    if (rb->AllocStorage(ctx, rb, internalFormat, width, height)) {
       /* No error - check/set fields now */
       /* If rb->Format == MESA_FORMAT_NONE, the format is unsupported. */
@@ -2036,7 +2036,7 @@ static void
 check_begin_texture_render(struct gl_context *ctx, struct gl_framebuffer *fb)
 {
    GLuint i;
-   ASSERT(ctx->Driver.RenderTexture);
+   assert(ctx->Driver.RenderTexture);
 
    if (_mesa_is_winsys_fbo(fb))
       return; /* can't render to texture with winsys framebuffers */
@@ -2134,8 +2134,8 @@ bind_framebuffer(GLenum target, GLuint framebuffer, bool allow_user_names)
       newReadFb = ctx->WinSysReadBuffer;
    }
 
-   ASSERT(newDrawFb);
-   ASSERT(newDrawFb != &DummyFramebuffer);
+   assert(newDrawFb);
+   assert(newDrawFb != &DummyFramebuffer);
 
    /* save pointers to current/old framebuffers */
    oldDrawFb = ctx->DrawBuffer;
@@ -2227,17 +2227,17 @@ _mesa_DeleteFramebuffers(GLsizei n, const GLuint *framebuffers)
 	 struct gl_framebuffer *fb;
 	 fb = _mesa_lookup_framebuffer(ctx, framebuffers[i]);
 	 if (fb) {
-            ASSERT(fb == &DummyFramebuffer || fb->Name == framebuffers[i]);
+            assert(fb == &DummyFramebuffer || fb->Name == framebuffers[i]);
 
             /* check if deleting currently bound framebuffer object */
             if (fb == ctx->DrawBuffer) {
                /* bind default */
-               ASSERT(fb->RefCount >= 2);
+               assert(fb->RefCount >= 2);
                _mesa_BindFramebuffer(GL_DRAW_FRAMEBUFFER, 0);
             }
             if (fb == ctx->ReadBuffer) {
                /* bind default */
-               ASSERT(fb->RefCount >= 2);
+               assert(fb->RefCount >= 2);
                _mesa_BindFramebuffer(GL_READ_FRAMEBUFFER, 0);
             }
 
diff --git a/mesalib/src/mesa/main/ffvertex_prog.c b/mesalib/src/mesa/main/ffvertex_prog.c
index c51c20ddf..395b00e15 100644
--- a/mesalib/src/mesa/main/ffvertex_prog.c
+++ b/mesalib/src/mesa/main/ffvertex_prog.c
@@ -479,7 +479,7 @@ static struct ureg register_const4f( struct tnl_program *p,
    values[3].f = s3;
    idx = _mesa_add_unnamed_constant( p->program->Base.Parameters, values, 4,
                                      &swizzle );
-   ASSERT(swizzle == SWIZZLE_NOOP);
+   assert(swizzle == SWIZZLE_NOOP);
    return make_ureg(PROGRAM_CONSTANT, idx);
 }
 
@@ -530,7 +530,7 @@ static void emit_arg( struct prog_src_register *src,
    src->Abs = 0;
    src->RelAddr = 0;
    /* Check that bitfield sizes aren't exceeded */
-   ASSERT(src->Index == reg.idx);
+   assert(src->Index == reg.idx);
 }
 
 
@@ -544,7 +544,7 @@ static void emit_dst( struct prog_dst_register *dst,
    dst->CondMask = COND_TR;  /* always pass cond test */
    dst->CondSwizzle = SWIZZLE_NOOP;
    /* Check that bitfield sizes aren't exceeded */
-   ASSERT(dst->Index == reg.idx);
+   assert(dst->Index == reg.idx);
 }
 
 
diff --git a/mesalib/src/mesa/main/format_unpack.py b/mesalib/src/mesa/main/format_unpack.py
index 2276a1063..53bdf641d 100644
--- a/mesalib/src/mesa/main/format_unpack.py
+++ b/mesalib/src/mesa/main/format_unpack.py
@@ -463,8 +463,8 @@ unpack_float_z_X8_UINT_Z24_UNORM(GLuint n, const void *src, GLfloat *dst)
    GLuint i;
    for (i = 0; i < n; i++) {
       dst[i] = (GLfloat) ((s[i] >> 8) * scale);
-      ASSERT(dst[i] >= 0.0F);
-      ASSERT(dst[i] <= 1.0F);
+      assert(dst[i] >= 0.0F);
+      assert(dst[i] <= 1.0F);
    }
 }
 
@@ -477,8 +477,8 @@ unpack_float_z_Z24_UNORM_X8_UINT(GLuint n, const void *src, GLfloat *dst)
    GLuint i;
    for (i = 0; i < n; i++) {
       dst[i] = (GLfloat) ((s[i] & 0x00ffffff) * scale);
-      ASSERT(dst[i] >= 0.0F);
-      ASSERT(dst[i] <= 1.0F);
+      assert(dst[i] >= 0.0F);
+      assert(dst[i] <= 1.0F);
    }
 }
 
diff --git a/mesalib/src/mesa/main/formats.c b/mesalib/src/mesa/main/formats.c
index 958d6f245..422c9dc46 100644
--- a/mesalib/src/mesa/main/formats.c
+++ b/mesalib/src/mesa/main/formats.c
@@ -75,7 +75,7 @@ struct gl_format_info
    mesa_array_format ArrayFormat;
 };
 
-#include "format_info.c"
+#include "format_info.h"
 
 static const struct gl_format_info *
 _mesa_get_format_info(mesa_format format)
@@ -107,8 +107,8 @@ GLint
 _mesa_get_format_bytes(mesa_format format)
 {
    const struct gl_format_info *info = _mesa_get_format_info(format);
-   ASSERT(info->BytesPerBlock);
-   ASSERT(info->BytesPerBlock <= MAX_PIXEL_BYTES ||
+   assert(info->BytesPerBlock);
+   assert(info->BytesPerBlock <= MAX_PIXEL_BYTES ||
           _mesa_is_format_compressed(format));
    return info->BytesPerBlock;
 }
@@ -887,7 +887,7 @@ _mesa_test_formats(void)
 {
    GLuint i;
 
-   STATIC_ASSERT(Elements(format_info) == MESA_FORMAT_COUNT);
+   STATIC_ASSERT(ARRAY_SIZE(format_info) == MESA_FORMAT_COUNT);
 
    for (i = 0; i < MESA_FORMAT_COUNT; i++) {
       const struct gl_format_info *info = _mesa_get_format_info(i);
diff --git a/mesalib/src/mesa/main/framebuffer.c b/mesalib/src/mesa/main/framebuffer.c
index dc0386d23..4f7736a64 100644
--- a/mesalib/src/mesa/main/framebuffer.c
+++ b/mesalib/src/mesa/main/framebuffer.c
@@ -29,7 +29,7 @@
  * ushorts, uints, etc.
  */
 
-
+#include <stdio.h>
 #include "glheader.h"
 #include "imports.h"
 #include "blend.h"
@@ -223,8 +223,8 @@ _mesa_free_framebuffer_data(struct gl_framebuffer *fb)
       if (att->Texture) {
          _mesa_reference_texobj(&att->Texture, NULL);
       }
-      ASSERT(!att->Renderbuffer);
-      ASSERT(!att->Texture);
+      assert(!att->Renderbuffer);
+      assert(!att->Texture);
       att->Type = GL_NONE;
    }
 }
@@ -245,7 +245,7 @@ _mesa_reference_framebuffer_(struct gl_framebuffer **ptr,
       struct gl_framebuffer *oldFb = *ptr;
 
       mtx_lock(&oldFb->Mutex);
-      ASSERT(oldFb->RefCount > 0);
+      assert(oldFb->RefCount > 0);
       oldFb->RefCount--;
       deleteFlag = (oldFb->RefCount == 0);
       mtx_unlock(&oldFb->Mutex);
@@ -296,8 +296,8 @@ _mesa_resize_framebuffer(struct gl_context *ctx, struct gl_framebuffer *fb,
          /* only resize if size is changing */
          if (rb->Width != width || rb->Height != height) {
             if (rb->AllocStorage(ctx, rb, rb->InternalFormat, width, height)) {
-               ASSERT(rb->Width == width);
-               ASSERT(rb->Height == height);
+               assert(rb->Width == width);
+               assert(rb->Height == height);
             }
             else {
                _mesa_error(ctx, GL_OUT_OF_MEMORY, "Resizing framebuffer");
@@ -402,8 +402,8 @@ _mesa_scissor_bounding_box(const struct gl_context *ctx,
       }
    }
 
-   ASSERT(bbox[0] <= bbox[1]);
-   ASSERT(bbox[2] <= bbox[3]);
+   assert(bbox[0] <= bbox[1]);
+   assert(bbox[2] <= bbox[3]);
 }
 
 /**
@@ -621,8 +621,8 @@ update_color_read_buffer(struct gl_context *ctx, struct gl_framebuffer *fb)
       fb->_ColorReadBuffer = NULL; /* legal! */
    }
    else {
-      ASSERT(fb->_ColorReadBufferIndex >= 0);
-      ASSERT(fb->_ColorReadBufferIndex < BUFFER_COUNT);
+      assert(fb->_ColorReadBufferIndex >= 0);
+      assert(fb->_ColorReadBufferIndex < BUFFER_COUNT);
       fb->_ColorReadBuffer
          = fb->Attachment[fb->_ColorReadBufferIndex].Renderbuffer;
    }
@@ -756,7 +756,7 @@ renderbuffer_exists(struct gl_context *ctx,
          if (!readBuf) {
             return GL_FALSE;
          }
-         ASSERT(_mesa_get_format_bits(readBuf->Format, GL_RED_BITS) > 0 ||
+         assert(_mesa_get_format_bits(readBuf->Format, GL_RED_BITS) > 0 ||
                 _mesa_get_format_bits(readBuf->Format, GL_ALPHA_BITS) > 0 ||
                 _mesa_get_format_bits(readBuf->Format, GL_TEXTURE_LUMINANCE_SIZE) > 0 ||
                 _mesa_get_format_bits(readBuf->Format, GL_TEXTURE_INTENSITY_SIZE) > 0 ||
diff --git a/mesalib/src/mesa/main/get.c b/mesalib/src/mesa/main/get.c
index 3f9d74516..a881bc589 100644
--- a/mesalib/src/mesa/main/get.c
+++ b/mesalib/src/mesa/main/get.c
@@ -497,12 +497,12 @@ print_table_stats(int api)
    };
    const char *api_name;
 
-   api_name = api < Elements(api_names) ? api_names[api] : "N/A";
+   api_name = api < ARRAY_SIZE(api_names) ? api_names[api] : "N/A";
    count = 0;
-   mask = Elements(table(api)) - 1;
+   mask = ARRAY_SIZE(table(api)) - 1;
    memset(collisions, 0, sizeof collisions);
 
-   for (i = 0; i < Elements(table(api)); i++) {
+   for (i = 0; i < ARRAY_SIZE(table(api)); i++) {
       if (!table(api)[i])
          continue;
       count++;
@@ -523,8 +523,8 @@ print_table_stats(int api)
    }
 
    printf("number of enums for %s: %d (total %ld)\n",
-         api_name, count, Elements(values));
-   for (i = 0; i < Elements(collisions) - 1; i++)
+         api_name, count, ARRAY_SIZE(values));
+   for (i = 0; i < ARRAY_SIZE(collisions) - 1; i++)
       if (collisions[i] > 0)
          printf("  %d enums with %d %scollisions\n",
                collisions[i], i, i == 10 ? "or more " : "");
@@ -758,7 +758,7 @@ find_custom_value(struct gl_context *ctx, const struct value_desc *d, union valu
    case GL_COMPRESSED_TEXTURE_FORMATS_ARB:
       v->value_int_n.n = 
 	 _mesa_get_compressed_formats(ctx, v->value_int_n.ints);
-      ASSERT(v->value_int_n.n <= (int) ARRAY_SIZE(v->value_int_n.ints));
+      assert(v->value_int_n.n <= (int) ARRAY_SIZE(v->value_int_n.ints));
       break;
 
    case GL_MAX_VARYING_FLOATS_ARB:
@@ -1195,11 +1195,11 @@ find_value(const char *func, GLenum pname, void **p, union value *v)
     * value since it's compatible with GLES2 its entry in table_set[] is at the
     * end.
     */
-   STATIC_ASSERT(Elements(table_set) == API_OPENGL_LAST + 2);
+   STATIC_ASSERT(ARRAY_SIZE(table_set) == API_OPENGL_LAST + 2);
    if (_mesa_is_gles3(ctx)) {
       api = API_OPENGL_LAST + 1;
    }
-   mask = Elements(table(api)) - 1;
+   mask = ARRAY_SIZE(table(api)) - 1;
    hash = (pname * prime_factor);
    while (1) {
       int idx = table(api)[hash & mask];
diff --git a/mesalib/src/mesa/main/hash.c b/mesalib/src/mesa/main/hash.c
index 1a152ec34..d04cccd94 100644
--- a/mesalib/src/mesa/main/hash.c
+++ b/mesalib/src/mesa/main/hash.c
@@ -371,8 +371,8 @@ _mesa_HashDeleteAll(struct _mesa_HashTable *table,
 {
    struct hash_entry *entry;
 
-   ASSERT(table);
-   ASSERT(callback);
+   assert(table);
+   assert(callback);
    mtx_lock(&table->Mutex);
    table->InDeleteAll = GL_TRUE;
    hash_table_foreach(table->ht, entry) {
@@ -401,7 +401,7 @@ _mesa_HashClone(const struct _mesa_HashTable *table)
    struct hash_entry *entry;
    struct _mesa_HashTable *clonetable;
 
-   ASSERT(table);
+   assert(table);
    mtx_lock(&table2->Mutex);
 
    clonetable = _mesa_NewHashTable();
@@ -437,8 +437,8 @@ _mesa_HashWalk(const struct _mesa_HashTable *table,
    struct _mesa_HashTable *table2 = (struct _mesa_HashTable *) table;
    struct hash_entry *entry;
 
-   ASSERT(table);
-   ASSERT(callback);
+   assert(table);
+   assert(callback);
    mtx_lock(&table2->WalkMutex);
    hash_table_foreach(table->ht, entry) {
       callback((uintptr_t)entry->key, entry->data, userData);
diff --git a/mesalib/src/mesa/main/histogram.h b/mesalib/src/mesa/main/histogram.h
index 1d145a03b..47a2bf0e5 100644
--- a/mesalib/src/mesa/main/histogram.h
+++ b/mesalib/src/mesa/main/histogram.h
@@ -36,9 +36,8 @@
 #ifndef HISTOGRAM_H
 #define HISTOGRAM_H
 
-#include "compiler.h"
+#include "glheader.h"
 
-struct _glapi_table;
 
 void GLAPIENTRY
 _mesa_GetnMinmaxARB(GLenum target, GLboolean reset, GLenum format,
diff --git a/mesalib/src/mesa/main/image.c b/mesalib/src/mesa/main/image.c
index e97b006e0..711a19062 100644
--- a/mesalib/src/mesa/main/image.c
+++ b/mesalib/src/mesa/main/image.c
@@ -118,7 +118,7 @@ _mesa_image_offset( GLuint dimensions,
    GLint skipimages;       /* for 3-D volume images */
    GLintptr offset;
 
-   ASSERT(dimensions >= 1 && dimensions <= 3);
+   assert(dimensions >= 1 && dimensions <= 3);
 
    alignment = packing->Alignment;
    if (packing->RowLength > 0) {
@@ -174,7 +174,7 @@ _mesa_image_offset( GLuint dimensions,
       if (remainder > 0)
          bytes_per_row += (alignment - remainder);
 
-      ASSERT(bytes_per_row % alignment == 0);
+      assert(bytes_per_row % alignment == 0);
 
       bytes_per_image = bytes_per_row * rows_per_image;
 
@@ -288,7 +288,7 @@ _mesa_image_row_stride( const struct gl_pixelstore_attrib *packing,
 {
    GLint bytesPerRow, remainder;
 
-   ASSERT(packing);
+   assert(packing);
 
    if (type == GL_BITMAP) {
       if (packing->RowLength == 0) {
@@ -336,7 +336,7 @@ _mesa_image_image_stride( const struct gl_pixelstore_attrib *packing,
 {
    GLint bytesPerRow, bytesPerImage, remainder;
 
-   ASSERT(packing);
+   assert(packing);
 
    if (type == GL_BITMAP) {
       if (packing->RowLength == 0) {
@@ -476,7 +476,7 @@ _mesa_convert_colors(GLenum srcType, const GLvoid *src,
    if (!tempBuffer)
       return;
 
-   ASSERT(srcType != dstType);
+   assert(srcType != dstType);
 
    switch (srcType) {
    case GL_UNSIGNED_BYTE:
@@ -499,7 +499,7 @@ _mesa_convert_colors(GLenum srcType, const GLvoid *src,
          const GLubyte (*src1)[4] = (const GLubyte (*)[4]) src;
          GLfloat (*dst4)[4] = (GLfloat (*)[4]) (useTemp ? tempBuffer : dst);
          GLuint i;
-         ASSERT(dstType == GL_FLOAT);
+         assert(dstType == GL_FLOAT);
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                dst4[i][RCOMP] = UBYTE_TO_FLOAT(src1[i][RCOMP]);
@@ -532,7 +532,7 @@ _mesa_convert_colors(GLenum srcType, const GLvoid *src,
          const GLushort (*src2)[4] = (const GLushort (*)[4]) src;
          GLfloat (*dst4)[4] = (GLfloat (*)[4]) (useTemp ? tempBuffer : dst);
          GLuint i;
-         ASSERT(dstType == GL_FLOAT);
+         assert(dstType == GL_FLOAT);
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                dst4[i][RCOMP] = USHORT_TO_FLOAT(src2[i][RCOMP]);
@@ -561,7 +561,7 @@ _mesa_convert_colors(GLenum srcType, const GLvoid *src,
          const GLfloat (*src4)[4] = (const GLfloat (*)[4]) src;
          GLushort (*dst2)[4] = (GLushort (*)[4]) (useTemp ? tempBuffer : dst);
          GLuint i;
-         ASSERT(dstType == GL_UNSIGNED_SHORT);
+         assert(dstType == GL_UNSIGNED_SHORT);
          for (i = 0; i < count; i++) {
             if (!mask || mask[i]) {
                UNCLAMPED_FLOAT_TO_USHORT(dst2[i][RCOMP], src4[i][RCOMP]);
@@ -607,8 +607,8 @@ _mesa_clip_drawpixels(const struct gl_context *ctx,
       unpack->RowLength = *width;
    }
 
-   ASSERT(ctx->Pixel.ZoomX == 1.0F);
-   ASSERT(ctx->Pixel.ZoomY == 1.0F || ctx->Pixel.ZoomY == -1.0F);
+   assert(ctx->Pixel.ZoomX == 1.0F);
+   assert(ctx->Pixel.ZoomY == 1.0F || ctx->Pixel.ZoomY == -1.0F);
 
    /* left clipping */
    if (*destX < buffer->_Xmin) {
@@ -792,20 +792,20 @@ clip_right_or_top(GLint *srcX0, GLint *srcX1,
 
    if (*dstX1 > maxValue) {
       /* X1 outside right edge */
-      ASSERT(*dstX0 < maxValue); /* X0 should be inside right edge */
+      assert(*dstX0 < maxValue); /* X0 should be inside right edge */
       t = (GLfloat) (maxValue - *dstX0) / (GLfloat) (*dstX1 - *dstX0);
       /* chop off [t, 1] part */
-      ASSERT(t >= 0.0 && t <= 1.0);
+      assert(t >= 0.0 && t <= 1.0);
       *dstX1 = maxValue;
       bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F;
       *srcX1 = *srcX0 + (GLint) (t * (*srcX1 - *srcX0) + bias);
    }
    else if (*dstX0 > maxValue) {
       /* X0 outside right edge */
-      ASSERT(*dstX1 < maxValue); /* X1 should be inside right edge */
+      assert(*dstX1 < maxValue); /* X1 should be inside right edge */
       t = (GLfloat) (maxValue - *dstX1) / (GLfloat) (*dstX0 - *dstX1);
       /* chop off [t, 1] part */
-      ASSERT(t >= 0.0 && t <= 1.0);
+      assert(t >= 0.0 && t <= 1.0);
       *dstX0 = maxValue;
       bias = (*srcX0 < *srcX1) ? -0.5F : 0.5F;
       *srcX0 = *srcX1 + (GLint) (t * (*srcX0 - *srcX1) + bias);
@@ -825,20 +825,20 @@ clip_left_or_bottom(GLint *srcX0, GLint *srcX1,
 
    if (*dstX0 < minValue) {
       /* X0 outside left edge */
-      ASSERT(*dstX1 > minValue); /* X1 should be inside left edge */
+      assert(*dstX1 > minValue); /* X1 should be inside left edge */
       t = (GLfloat) (minValue - *dstX0) / (GLfloat) (*dstX1 - *dstX0);
       /* chop off [0, t] part */
-      ASSERT(t >= 0.0 && t <= 1.0);
+      assert(t >= 0.0 && t <= 1.0);
       *dstX0 = minValue;
       bias = (*srcX0 < *srcX1) ? 0.5F : -0.5F;
       *srcX0 = *srcX0 + (GLint) (t * (*srcX1 - *srcX0) + bias);
    }
    else if (*dstX1 < minValue) {
       /* X1 outside left edge */
-      ASSERT(*dstX0 > minValue); /* X0 should be inside left edge */
+      assert(*dstX0 > minValue); /* X0 should be inside left edge */
       t = (GLfloat) (minValue - *dstX1) / (GLfloat) (*dstX0 - *dstX1);
       /* chop off [0, t] part */
-      ASSERT(t >= 0.0 && t <= 1.0);
+      assert(t >= 0.0 && t <= 1.0);
       *dstX1 = minValue;
       bias = (*srcX0 < *srcX1) ? -0.5F : 0.5F;
       *srcX1 = *srcX1 + (GLint) (t * (*srcX0 - *srcX1) + bias);
@@ -936,25 +936,25 @@ _mesa_clip_blit(struct gl_context *ctx,
           *srcY0, *srcY1, *dstY0, *dstY1);
    */
 
-   ASSERT(*dstX0 >= dstXmin);
-   ASSERT(*dstX0 <= dstXmax);
-   ASSERT(*dstX1 >= dstXmin);
-   ASSERT(*dstX1 <= dstXmax);
-
-   ASSERT(*dstY0 >= dstYmin);
-   ASSERT(*dstY0 <= dstYmax);
-   ASSERT(*dstY1 >= dstYmin);
-   ASSERT(*dstY1 <= dstYmax);
-
-   ASSERT(*srcX0 >= srcXmin);
-   ASSERT(*srcX0 <= srcXmax);
-   ASSERT(*srcX1 >= srcXmin);
-   ASSERT(*srcX1 <= srcXmax);
-
-   ASSERT(*srcY0 >= srcYmin);
-   ASSERT(*srcY0 <= srcYmax);
-   ASSERT(*srcY1 >= srcYmin);
-   ASSERT(*srcY1 <= srcYmax);
+   assert(*dstX0 >= dstXmin);
+   assert(*dstX0 <= dstXmax);
+   assert(*dstX1 >= dstXmin);
+   assert(*dstX1 <= dstXmax);
+
+   assert(*dstY0 >= dstYmin);
+   assert(*dstY0 <= dstYmax);
+   assert(*dstY1 >= dstYmin);
+   assert(*dstY1 <= dstYmax);
+
+   assert(*srcX0 >= srcXmin);
+   assert(*srcX0 <= srcXmax);
+   assert(*srcX1 >= srcXmin);
+   assert(*srcX1 <= srcXmax);
+
+   assert(*srcY0 >= srcYmin);
+   assert(*srcY0 <= srcYmax);
+   assert(*srcY1 >= srcYmin);
+   assert(*srcY1 <= srcYmax);
 
    return GL_TRUE;
 }
diff --git a/mesalib/src/mesa/main/imports.c b/mesalib/src/mesa/main/imports.c
index 4f5a2d11f..a7ffe2296 100644
--- a/mesalib/src/mesa/main/imports.c
+++ b/mesalib/src/mesa/main/imports.c
@@ -42,8 +42,9 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-
-
+#include <stdio.h>
+#include <stdarg.h>
+#include "c99_math.h"
 #include "imports.h"
 #include "context.h"
 #include "mtypes.h"
@@ -92,7 +93,7 @@ _mesa_align_malloc(size_t bytes, unsigned long alignment)
 #else
    uintptr_t ptr, buf;
 
-   ASSERT( alignment > 0 );
+   assert( alignment > 0 );
 
    ptr = (uintptr_t)malloc(bytes + alignment + sizeof(void *));
    if (!ptr)
@@ -141,7 +142,7 @@ _mesa_align_calloc(size_t bytes, unsigned long alignment)
 #else
    uintptr_t ptr, buf;
 
-   ASSERT( alignment > 0 );
+   assert( alignment > 0 );
 
    ptr = (uintptr_t)calloc(1, bytes + alignment + sizeof(void *));
    if (!ptr)
diff --git a/mesalib/src/mesa/main/imports.h b/mesalib/src/mesa/main/imports.h
index 0fcba4fb9..792100061 100644
--- a/mesalib/src/mesa/main/imports.h
+++ b/mesalib/src/mesa/main/imports.h
@@ -36,6 +36,9 @@
 #define IMPORTS_H
 
 
+#include <stdlib.h>
+#include <stdarg.h>
+#include <string.h>
 #include "compiler.h"
 #include "glheader.h"
 #include "errors.h"
@@ -79,62 +82,8 @@ typedef union { GLfloat f; GLint i; GLuint u; } fi_type;
 
 
 
-/**********************************************************************
- * Math macros
- */
-
-#define MAX_GLUSHORT	0xffff
-#define MAX_GLUINT	0xffffffff
-
-/* Degrees to radians conversion: */
-#define DEG2RAD (M_PI/180.0)
-
-
-/**
- * \name Work-arounds for platforms that lack C99 math functions
- */
-/*@{*/
-#if (!defined(_XOPEN_SOURCE) || (_XOPEN_SOURCE < 600)) && !defined(_ISOC99_SOURCE) \
-   && (!defined(__STDC_VERSION__) || (__STDC_VERSION__ < 199901L)) \
-   && (!defined(_MSC_VER) || (_MSC_VER < 1400))
-#define acosf(f) ((float) acos(f))
-#define asinf(f) ((float) asin(f))
-#define atan2f(x,y) ((float) atan2(x,y))
-#define atanf(f) ((float) atan(f))
-#define ceilf(f) ((float) ceil(f))
-#define cosf(f) ((float) cos(f))
-#define coshf(f) ((float) cosh(f))
-#define expf(f) ((float) exp(f))
-#define exp2f(f) ((float) exp2(f))
-#define floorf(f) ((float) floor(f))
-#define logf(f) ((float) log(f))
-
-#ifdef ANDROID
-#define log2f(f) (logf(f) * (float) (1.0 / M_LN2))
-#else
-#define log2f(f) ((float) log2(f))
-#endif
-
-#define powf(x,y) ((float) pow(x,y))
-#define sinf(f) ((float) sin(f))
-#define sinhf(f) ((float) sinh(f))
-#define sqrtf(f) ((float) sqrt(f))
-#define tanf(f) ((float) tan(f))
-#define tanhf(f) ((float) tanh(f))
-#define acoshf(f) ((float) acosh(f))
-#define asinhf(f) ((float) asinh(f))
-#define atanhf(f) ((float) atanh(f))
-#endif
-
 #if defined(_MSC_VER)
 #if _MSC_VER < 1800  /* Not req'd on VS2013 and above */
-static inline float truncf(float x) { return x < 0.0f ? ceilf(x) : floorf(x); }
-static inline float exp2f(float x) { return powf(2.0f, x); }
-static inline float log2f(float x) { return logf(x) * 1.442695041f; }
-static inline float asinhf(float x) { return logf(x + sqrtf(x * x + 1.0f)); }
-static inline float acoshf(float x) { return logf(x + sqrtf(x * x - 1.0f)); }
-static inline float atanhf(float x) { return (logf(1.0f + x) - logf(1.0f - x)) / 2.0f; }
-static inline int isblank(int ch) { return ch == ' ' || ch == '\t'; }
 #define strtoll(p, e, b) _strtoi64(p, e, b)
 #endif /* _MSC_VER < 1800 */
 #define strcasecmp(s1, s2) _stricmp(s1, s2)
@@ -142,23 +91,6 @@ static inline int isblank(int ch) { return ch == ' ' || ch == '\t'; }
 /*@}*/
 
 
-/*
- * signbit() is a macro on Linux.  Not available on Windows.
- */
-#ifndef signbit
-#define signbit(x) ((x) < 0.0f)
-#endif
-
-
-/** single-precision inverse square root */
-static inline float
-INV_SQRTF(float x)
-{
-   /* XXX we could try Quake's fast inverse square root function here */
-   return 1.0F / sqrtf(x);
-}
-
-
 /***
  *** LOG2: Log base 2 of float
  ***/
@@ -189,6 +121,14 @@ static inline GLfloat LOG2(GLfloat x)
 
 
 
+/**
+ * finite macro.
+ */
+#if defined(_MSC_VER)
+#  define finite _finite
+#endif
+
+
 /***
  *** IS_INF_OR_NAN: test if float is infinite or NaN
  ***/
@@ -203,35 +143,6 @@ static inline GLfloat LOG2(GLfloat x)
 #endif
 
 
-/***
- *** CEILF: ceiling of float
- *** FLOORF: floor of float
- *** FABSF: absolute value of float
- *** LOGF: the natural logarithm (base e) of the value
- *** EXPF: raise e to the value
- *** LDEXPF: multiply value by an integral power of two
- *** FREXPF: extract mantissa and exponent from value
- ***/
-#if defined(__gnu_linux__)
-/* C99 functions */
-#define CEILF(x)   ceilf(x)
-#define FLOORF(x)  floorf(x)
-#define FABSF(x)   fabsf(x)
-#define LOGF(x)    logf(x)
-#define EXPF(x)    expf(x)
-#define LDEXPF(x,y)  ldexpf(x,y)
-#define FREXPF(x,y)  frexpf(x,y)
-#else
-#define CEILF(x)   ((GLfloat) ceil(x))
-#define FLOORF(x)  ((GLfloat) floor(x))
-#define FABSF(x)   ((GLfloat) fabs(x))
-#define LOGF(x)    ((GLfloat) log(x))
-#define EXPF(x)    ((GLfloat) exp(x))
-#define LDEXPF(x,y)  ((GLfloat) ldexp(x,y))
-#define FREXPF(x,y)  ((GLfloat) frexp(x,y))
-#endif
-
-
 /**
  * Convert float to int by rounding to nearest integer, away from zero.
  */
diff --git a/mesalib/src/mesa/main/light.c b/mesalib/src/mesa/main/light.c
index e483b826e..4021dbef9 100644
--- a/mesalib/src/mesa/main/light.c
+++ b/mesalib/src/mesa/main/light.c
@@ -24,6 +24,7 @@
  */
 
 
+#include "c99_math.h"
 #include "glheader.h"
 #include "imports.h"
 #include "context.h"
@@ -101,7 +102,7 @@ _mesa_light(struct gl_context *ctx, GLuint lnum, GLenum pname, const GLfloat *pa
 {
    struct gl_light *light;
 
-   ASSERT(lnum < MAX_LIGHTS);
+   assert(lnum < MAX_LIGHTS);
    light = &ctx->Light.Light[lnum];
 
    switch (pname) {
@@ -142,20 +143,20 @@ _mesa_light(struct gl_context *ctx, GLuint lnum, GLenum pname, const GLfloat *pa
       COPY_3V(light->SpotDirection, params);
       break;
    case GL_SPOT_EXPONENT:
-      ASSERT(params[0] >= 0.0);
-      ASSERT(params[0] <= ctx->Const.MaxSpotExponent);
+      assert(params[0] >= 0.0);
+      assert(params[0] <= ctx->Const.MaxSpotExponent);
       if (light->SpotExponent == params[0])
 	 return;
       FLUSH_VERTICES(ctx, _NEW_LIGHT);
       light->SpotExponent = params[0];
       break;
    case GL_SPOT_CUTOFF:
-      ASSERT(params[0] == 180.0 || (params[0] >= 0.0 && params[0] <= 90.0));
+      assert(params[0] == 180.0 || (params[0] >= 0.0 && params[0] <= 90.0));
       if (light->SpotCutoff == params[0])
          return;
       FLUSH_VERTICES(ctx, _NEW_LIGHT);
       light->SpotCutoff = params[0];
-      light->_CosCutoff = (GLfloat) (cos(light->SpotCutoff * DEG2RAD));
+      light->_CosCutoff = (GLfloat) (cos(light->SpotCutoff * M_PI / 180.0));
       if (light->_CosCutoff < 0)
          light->_CosCutoff = 0;
       if (light->SpotCutoff != 180.0F)
@@ -164,21 +165,21 @@ _mesa_light(struct gl_context *ctx, GLuint lnum, GLenum pname, const GLfloat *pa
          light->_Flags &= ~LIGHT_SPOT;
       break;
    case GL_CONSTANT_ATTENUATION:
-      ASSERT(params[0] >= 0.0);
+      assert(params[0] >= 0.0);
       if (light->ConstantAttenuation == params[0])
 	 return;
       FLUSH_VERTICES(ctx, _NEW_LIGHT);
       light->ConstantAttenuation = params[0];
       break;
    case GL_LINEAR_ATTENUATION:
-      ASSERT(params[0] >= 0.0);
+      assert(params[0] >= 0.0);
       if (light->LinearAttenuation == params[0])
 	 return;
       FLUSH_VERTICES(ctx, _NEW_LIGHT);
       light->LinearAttenuation = params[0];
       break;
    case GL_QUADRATIC_ATTENUATION:
-      ASSERT(params[0] >= 0.0);
+      assert(params[0] >= 0.0);
       if (light->QuadraticAttenuation == params[0])
 	 return;
       FLUSH_VERTICES(ctx, _NEW_LIGHT);
@@ -808,7 +809,7 @@ _mesa_GetMaterialiv( GLenum face, GLenum pname, GLint *params )
    GLuint f;
    GLfloat (*mat)[4] = ctx->Light.Material.Attrib;
 
-   ASSERT(ctx->API == API_OPENGL_COMPAT);
+   assert(ctx->API == API_OPENGL_COMPAT);
 
    FLUSH_VERTICES(ctx, 0); /* update materials */
    FLUSH_CURRENT(ctx, 0); /* update ctx->Light.Material from vertex buffer */
@@ -1025,9 +1026,9 @@ update_modelview_scale( struct gl_context *ctx )
       GLfloat f = m[2] * m[2] + m[6] * m[6] + m[10] * m[10];
       if (f < 1e-12) f = 1.0;
       if (ctx->_NeedEyeCoords)
-	 ctx->_ModelViewInvScale = (GLfloat) INV_SQRTF(f);
+	 ctx->_ModelViewInvScale = 1.0f / sqrtf(f);
       else
-	 ctx->_ModelViewInvScale = (GLfloat) sqrtf(f);
+	 ctx->_ModelViewInvScale = sqrtf(f);
    }
 }
 
diff --git a/mesalib/src/mesa/main/macros.h b/mesalib/src/mesa/main/macros.h
index cf1f0e9c9..2d7a6a107 100644
--- a/mesalib/src/mesa/main/macros.h
+++ b/mesalib/src/mesa/main/macros.h
@@ -171,25 +171,25 @@ extern GLfloat _mesa_ubyte_to_float_color_tab[256];
 	ub = ((GLubyte) F_TO_I((f) * 255.0F))
 #endif
 
-static inline GLfloat INT_AS_FLT(GLint i)
+static fi_type UINT_AS_UNION(GLuint u)
 {
    fi_type tmp;
-   tmp.i = i;
-   return tmp.f;
+   tmp.u = u;
+   return tmp;
 }
 
-static inline GLfloat UINT_AS_FLT(GLuint u)
+static inline fi_type INT_AS_UNION(GLint i)
 {
    fi_type tmp;
-   tmp.u = u;
-   return tmp.f;
+   tmp.i = i;
+   return tmp;
 }
 
-static inline unsigned FLT_AS_UINT(float f)
+static inline fi_type FLOAT_AS_UNION(GLfloat f)
 {
    fi_type tmp;
    tmp.f = f;
-   return tmp.u;
+   return tmp;
 }
 
 /**
@@ -604,24 +604,26 @@ do {				\
  * The default values are chosen based on \p type.
  */
 static inline void
-COPY_CLEAN_4V_TYPE_AS_FLOAT(GLfloat dst[4], int sz, const GLfloat src[4],
+COPY_CLEAN_4V_TYPE_AS_UNION(fi_type dst[4], int sz, const fi_type src[4],
                             GLenum type)
 {
    switch (type) {
    case GL_FLOAT:
-      ASSIGN_4V(dst, 0, 0, 0, 1);
+      ASSIGN_4V(dst, FLOAT_AS_UNION(0), FLOAT_AS_UNION(0),
+                FLOAT_AS_UNION(0), FLOAT_AS_UNION(1));
       break;
    case GL_INT:
-      ASSIGN_4V(dst, INT_AS_FLT(0), INT_AS_FLT(0),
-                     INT_AS_FLT(0), INT_AS_FLT(1));
+      ASSIGN_4V(dst, INT_AS_UNION(0), INT_AS_UNION(0),
+                INT_AS_UNION(0), INT_AS_UNION(1));
       break;
    case GL_UNSIGNED_INT:
-      ASSIGN_4V(dst, UINT_AS_FLT(0), UINT_AS_FLT(0),
-                     UINT_AS_FLT(0), UINT_AS_FLT(1));
+      ASSIGN_4V(dst, UINT_AS_UNION(0), UINT_AS_UNION(0),
+                UINT_AS_UNION(0), UINT_AS_UNION(1));
       break;
    default:
-      ASSIGN_4V(dst, 0.0f, 0.0f, 0.0f, 1.0f); /* silence warnings */
-      ASSERT(!"Unexpected type in COPY_CLEAN_4V_TYPE_AS_FLOAT macro");
+      ASSIGN_4V(dst, FLOAT_AS_UNION(0), FLOAT_AS_UNION(0),
+                FLOAT_AS_UNION(0), FLOAT_AS_UNION(1)); /* silence warnings */
+      assert(!"Unexpected type in COPY_CLEAN_4V_TYPE_AS_UNION macro");
    }
    COPY_SZ_4V(dst, sz, src);
 }
@@ -775,7 +777,7 @@ NORMALIZE_3FV(GLfloat v[3])
 {
    GLfloat len = (GLfloat) LEN_SQUARED_3FV(v);
    if (len) {
-      len = INV_SQRTF(len);
+      len = 1.0f / sqrtf(len);
       v[0] *= len;
       v[1] *= len;
       v[2] *= len;
@@ -801,10 +803,6 @@ DIFFERENT_SIGNS(GLfloat x, GLfloat y)
 #define ENUM_TO_DOUBLE(E)  ((GLdouble)(GLint)(E))
 #define ENUM_TO_BOOLEAN(E) ((E) ? GL_TRUE : GL_FALSE)
 
-/* Compute the size of an array */
-#ifndef ARRAY_SIZE
-#  define ARRAY_SIZE(x) (sizeof(x) / sizeof(x[0]))
-#endif
 
 /* Stringify */
 #define STRINGIFY(x) #x
diff --git a/mesalib/src/mesa/main/matrix.c b/mesalib/src/mesa/main/matrix.c
index 0539caa47..80c8a248c 100644
--- a/mesalib/src/mesa/main/matrix.c
+++ b/mesalib/src/mesa/main/matrix.c
@@ -176,7 +176,7 @@ _mesa_MatrixMode( GLenum mode )
          return;
       }
 #endif
-      ASSERT(ctx->Texture.CurrentUnit < Elements(ctx->TextureMatrixStack));
+      assert(ctx->Texture.CurrentUnit < ARRAY_SIZE(ctx->TextureMatrixStack));
       ctx->CurrentStack = &ctx->TextureMatrixStack[ctx->Texture.CurrentUnit];
       break;
    case GL_MATRIX0_ARB:
@@ -697,10 +697,10 @@ void _mesa_init_matrix( struct gl_context * ctx )
                      _NEW_MODELVIEW);
    init_matrix_stack(&ctx->ProjectionMatrixStack, MAX_PROJECTION_STACK_DEPTH,
                      _NEW_PROJECTION);
-   for (i = 0; i < Elements(ctx->TextureMatrixStack); i++)
+   for (i = 0; i < ARRAY_SIZE(ctx->TextureMatrixStack); i++)
       init_matrix_stack(&ctx->TextureMatrixStack[i], MAX_TEXTURE_STACK_DEPTH,
                         _NEW_TEXTURE_MATRIX);
-   for (i = 0; i < Elements(ctx->ProgramMatrixStack); i++)
+   for (i = 0; i < ARRAY_SIZE(ctx->ProgramMatrixStack); i++)
       init_matrix_stack(&ctx->ProgramMatrixStack[i],
 		        MAX_PROGRAM_MATRIX_STACK_DEPTH, _NEW_TRACK_MATRIX);
    ctx->CurrentStack = &ctx->ModelviewMatrixStack;
@@ -724,9 +724,9 @@ void _mesa_free_matrix_data( struct gl_context *ctx )
 
    free_matrix_stack(&ctx->ModelviewMatrixStack);
    free_matrix_stack(&ctx->ProjectionMatrixStack);
-   for (i = 0; i < Elements(ctx->TextureMatrixStack); i++)
+   for (i = 0; i < ARRAY_SIZE(ctx->TextureMatrixStack); i++)
       free_matrix_stack(&ctx->TextureMatrixStack[i]);
-   for (i = 0; i < Elements(ctx->ProgramMatrixStack); i++)
+   for (i = 0; i < ARRAY_SIZE(ctx->ProgramMatrixStack); i++)
       free_matrix_stack(&ctx->ProgramMatrixStack[i]);
    /* combined Modelview*Projection matrix */
    _math_matrix_dtr( &ctx->_ModelProjectMatrix );
diff --git a/mesalib/src/mesa/main/mipmap.c b/mesalib/src/mesa/main/mipmap.c
index 75a12cd16..7732d09b2 100644
--- a/mesalib/src/mesa/main/mipmap.c
+++ b/mesalib/src/mesa/main/mipmap.c
@@ -153,8 +153,8 @@ do_row(GLenum datatype, GLuint comps, GLint srcWidth,
    const GLuint k0 = (srcWidth == dstWidth) ? 0 : 1;
    const GLuint colStride = (srcWidth == dstWidth) ? 1 : 2;
 
-   ASSERT(comps >= 1);
-   ASSERT(comps <= 4);
+   assert(comps >= 1);
+   assert(comps <= 4);
 
    /* This assertion is no longer valid with non-power-of-2 textures
    assert(srcWidth == dstWidth || srcWidth == 2 * dstWidth);
@@ -788,8 +788,8 @@ do_row_3D(GLenum datatype, GLuint comps, GLint srcWidth,
    const GLuint colStride = (srcWidth == dstWidth) ? 1 : 2;
    GLuint i, j, k;
 
-   ASSERT(comps >= 1);
-   ASSERT(comps <= 4);
+   assert(comps >= 1);
+   assert(comps <= 4);
 
    if ((datatype == GL_UNSIGNED_BYTE) && (comps == 4)) {
       DECLARE_ROW_POINTERS(GLubyte, 4);
@@ -1650,7 +1650,7 @@ make_3d_mipmap(GLenum datatype, GLuint comps, GLint border,
       }
       else {
          /* average border pixels from adjacent src image pairs */
-         ASSERT(srcDepthNB == 2 * dstDepthNB);
+         assert(srcDepthNB == 2 * dstDepthNB);
          for (img = 0; img < dstDepthNB; img++) {
             const GLubyte *srcA, *srcB;
             GLubyte *dst;
@@ -1902,7 +1902,7 @@ generate_mipmap_uncompressed(struct gl_context *ctx, GLenum target,
 
       /* get src image parameters */
       srcImage = _mesa_select_tex_image(texObj, target, level);
-      ASSERT(srcImage);
+      assert(srcImage);
       srcWidth = srcImage->Width;
       srcHeight = srcImage->Height;
       srcDepth = srcImage->Depth;
@@ -2097,7 +2097,7 @@ generate_mipmap_compressed(struct gl_context *ctx, GLenum target,
 
       /* get src image parameters */
       srcImage = _mesa_select_tex_image(texObj, target, level);
-      ASSERT(srcImage);
+      assert(srcImage);
       srcWidth = srcImage->Width;
       srcHeight = srcImage->Height;
       srcDepth = srcImage->Depth;
@@ -2192,12 +2192,12 @@ _mesa_generate_mipmap(struct gl_context *ctx, GLenum target,
    struct gl_texture_image *srcImage;
    GLint maxLevel;
 
-   ASSERT(texObj);
+   assert(texObj);
    srcImage = _mesa_select_tex_image(texObj, target, texObj->BaseLevel);
-   ASSERT(srcImage);
+   assert(srcImage);
 
    maxLevel = _mesa_max_texture_levels(ctx, texObj->Target) - 1;
-   ASSERT(maxLevel >= 0);  /* bad target */
+   assert(maxLevel >= 0);  /* bad target */
 
    maxLevel = MIN2(maxLevel, texObj->MaxLevel);
 
diff --git a/mesalib/src/mesa/main/mtypes.h b/mesalib/src/mesa/main/mtypes.h
index 6e9977309..efeee8bff 100644
--- a/mesalib/src/mesa/main/mtypes.h
+++ b/mesalib/src/mesa/main/mtypes.h
@@ -36,6 +36,7 @@
 
 #include <stdint.h>             /* uint32_t */
 #include <stdbool.h>
+#include "c11/threads.h"
 
 #include "main/glheader.h"
 #include "main/config.h"
diff --git a/mesalib/src/mesa/main/multisample.c b/mesalib/src/mesa/main/multisample.c
index b696de9f2..0e9207bc1 100644
--- a/mesalib/src/mesa/main/multisample.c
+++ b/mesalib/src/mesa/main/multisample.c
@@ -150,16 +150,17 @@ GLenum
 _mesa_check_sample_count(struct gl_context *ctx, GLenum target,
                          GLenum internalFormat, GLsizei samples)
 {
-   /* Section 4.4 (Framebuffer objects) of the OpenGL 3.0 specification says:
+   /* Section 4.4 (Framebuffer objects), page 198 of the OpenGL ES 3.0.0
+    * specification says:
     *
     *     "If internalformat is a signed or unsigned integer format and samples
     *     is greater than zero, then the error INVALID_OPERATION is generated."
     */
-   if (_mesa_is_gles3(ctx) && _mesa_is_enum_format_integer(internalFormat)) {
+   if (_mesa_is_gles3(ctx) && _mesa_is_enum_format_integer(internalFormat)
+       && samples > 0) {
       return GL_INVALID_OPERATION;
    }
 
-
    /* If ARB_internalformat_query is supported, then treat its highest
     * returned sample count as the absolute maximum for this format; it is
     * allowed to exceed MAX_SAMPLES.
diff --git a/mesalib/src/mesa/main/pack.c b/mesalib/src/mesa/main/pack.c
index 2111a7604..f72360817 100644
--- a/mesalib/src/mesa/main/pack.c
+++ b/mesalib/src/mesa/main/pack.c
@@ -257,9 +257,9 @@ extract_uint_indexes(GLuint n, GLuint indexes[],
                      GLenum srcFormat, GLenum srcType, const GLvoid *src,
                      const struct gl_pixelstore_attrib *unpack )
 {
-   ASSERT(srcFormat == GL_COLOR_INDEX || srcFormat == GL_STENCIL_INDEX);
+   assert(srcFormat == GL_COLOR_INDEX || srcFormat == GL_STENCIL_INDEX);
 
-   ASSERT(srcType == GL_BITMAP ||
+   assert(srcType == GL_BITMAP ||
           srcType == GL_UNSIGNED_BYTE ||
           srcType == GL_BYTE ||
           srcType == GL_UNSIGNED_SHORT ||
@@ -503,7 +503,7 @@ _mesa_unpack_stencil_span( struct gl_context *ctx, GLuint n,
                            const struct gl_pixelstore_attrib *srcPacking,
                            GLbitfield transferOps )
 {
-   ASSERT(srcType == GL_BITMAP ||
+   assert(srcType == GL_BITMAP ||
           srcType == GL_UNSIGNED_BYTE ||
           srcType == GL_BYTE ||
           srcType == GL_UNSIGNED_SHORT ||
@@ -516,7 +516,7 @@ _mesa_unpack_stencil_span( struct gl_context *ctx, GLuint n,
           srcType == GL_FLOAT ||
           srcType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
-   ASSERT(dstType == GL_UNSIGNED_BYTE ||
+   assert(dstType == GL_UNSIGNED_BYTE ||
           dstType == GL_UNSIGNED_SHORT ||
           dstType == GL_UNSIGNED_INT ||
           dstType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
@@ -988,7 +988,7 @@ _mesa_unpack_depth_span( struct gl_context *ctx, GLuint n,
    else if (dstType == GL_UNSIGNED_SHORT) {
       GLushort *zValues = (GLushort *) dest;
       GLuint i;
-      ASSERT(depthMax <= 0xffff);
+      assert(depthMax <= 0xffff);
       for (i = 0; i < n; i++) {
          zValues[i] = (GLushort) (depthValues[i] * (GLfloat) depthMax);
       }
@@ -1004,7 +1004,7 @@ _mesa_unpack_depth_span( struct gl_context *ctx, GLuint n,
       }
    }
    else {
-      ASSERT(0);
+      assert(0);
    }
 
    free(depthTemp);
@@ -1487,20 +1487,20 @@ _mesa_pack_luminance_from_rgba_integer(GLuint n,
          case GL_UNSIGNED_BYTE: {
             GLbyte *dst = (GLbyte *) dstAddr;
             dst[i] = lum32;
+            break;
          }
-         break;
          case GL_SHORT:
          case GL_UNSIGNED_SHORT: {
             GLshort *dst = (GLshort *) dstAddr;
             dst[i] = lum32;
+            break;
          }
-         break;
          case GL_INT:
          case GL_UNSIGNED_INT: {
             GLint *dst = (GLint *) dstAddr;
             dst[i] = lum32;
+            break;
          }
-         break;
          }
       }
       return;
@@ -1525,21 +1525,22 @@ _mesa_pack_luminance_from_rgba_integer(GLuint n,
             GLbyte *dst = (GLbyte *) dstAddr;
             dst[2*i] = lum32;
             dst[2*i+1] = alpha;
+            break;
          }
          case GL_SHORT:
          case GL_UNSIGNED_SHORT: {
             GLshort *dst = (GLshort *) dstAddr;
             dst[i] = lum32;
             dst[2*i+1] = alpha;
+            break;
          }
-         break;
          case GL_INT:
          case GL_UNSIGNED_INT: {
             GLint *dst = (GLint *) dstAddr;
             dst[i] = lum32;
             dst[2*i+1] = alpha;
+            break;
          }
-         break;
          }
       }
       return;
diff --git a/mesalib/src/mesa/main/pbo.c b/mesalib/src/mesa/main/pbo.c
index a0d61a643..5c906ed74 100644
--- a/mesalib/src/mesa/main/pbo.c
+++ b/mesalib/src/mesa/main/pbo.c
@@ -182,7 +182,7 @@ _mesa_map_validate_pbo_source(struct gl_context *ctx,
                               GLsizei clientMemSize,
                               const GLvoid *ptr, const char *where)
 {
-   ASSERT(dimensions == 1 || dimensions == 2 || dimensions == 3);
+   assert(dimensions == 1 || dimensions == 2 || dimensions == 3);
 
    if (!_mesa_validate_pbo_access(dimensions, unpack, width, height, depth,
                                   format, type, clientMemSize, ptr)) {
@@ -220,7 +220,7 @@ void
 _mesa_unmap_pbo_source(struct gl_context *ctx,
                        const struct gl_pixelstore_attrib *unpack)
 {
-   ASSERT(unpack != &ctx->Pack); /* catch pack/unpack mismatch */
+   assert(unpack != &ctx->Pack); /* catch pack/unpack mismatch */
    if (_mesa_is_bufferobj(unpack->BufferObj)) {
       ctx->Driver.UnmapBuffer(ctx, unpack->BufferObj, MAP_INTERNAL);
    }
@@ -279,7 +279,7 @@ _mesa_map_validate_pbo_dest(struct gl_context *ctx,
                             GLenum format, GLenum type, GLsizei clientMemSize,
                             GLvoid *ptr, const char *where)
 {
-   ASSERT(dimensions == 1 || dimensions == 2 || dimensions == 3);
+   assert(dimensions == 1 || dimensions == 2 || dimensions == 3);
 
    if (!_mesa_validate_pbo_access(dimensions, unpack, width, height, depth,
                                   format, type, clientMemSize, ptr)) {
@@ -317,7 +317,7 @@ void
 _mesa_unmap_pbo_dest(struct gl_context *ctx,
                      const struct gl_pixelstore_attrib *pack)
 {
-   ASSERT(pack != &ctx->Unpack); /* catch pack/unpack mismatch */
+   assert(pack != &ctx->Unpack); /* catch pack/unpack mismatch */
    if (_mesa_is_bufferobj(pack->BufferObj)) {
       ctx->Driver.UnmapBuffer(ctx, pack->BufferObj, MAP_INTERNAL);
    }
diff --git a/mesalib/src/mesa/main/pipelineobj.c b/mesalib/src/mesa/main/pipelineobj.c
index b713d956f..fb241af1e 100644
--- a/mesalib/src/mesa/main/pipelineobj.c
+++ b/mesalib/src/mesa/main/pipelineobj.c
@@ -187,7 +187,7 @@ _mesa_reference_pipeline_object_(struct gl_context *ctx,
       struct gl_pipeline_object *oldObj = *ptr;
 
       mtx_lock(&oldObj->Mutex);
-      ASSERT(oldObj->RefCount > 0);
+      assert(oldObj->RefCount > 0);
       oldObj->RefCount--;
       deleteFlag = (oldObj->RefCount == 0);
       mtx_unlock(&oldObj->Mutex);
@@ -198,7 +198,7 @@ _mesa_reference_pipeline_object_(struct gl_context *ctx,
 
       *ptr = NULL;
    }
-   ASSERT(!*ptr);
+   assert(!*ptr);
 
    if (obj) {
       /* reference new pipeline object */
@@ -471,7 +471,7 @@ _mesa_DeleteProgramPipelines(GLsizei n, const GLuint *pipelines)
          lookup_pipeline_object(ctx, pipelines[i]);
 
       if (obj) {
-         ASSERT(obj->Name == pipelines[i]);
+         assert(obj->Name == pipelines[i]);
 
          /* If the pipeline object is currently bound, the spec says "If an
           * object that is currently bound is deleted, the binding for that
diff --git a/mesalib/src/mesa/main/pixel.h b/mesalib/src/mesa/main/pixel.h
index c12ef97b8..fd1782e1b 100644
--- a/mesalib/src/mesa/main/pixel.h
+++ b/mesalib/src/mesa/main/pixel.h
@@ -33,10 +33,8 @@
 #define PIXEL_H
 
 
-#include "compiler.h"
 #include "glheader.h"
 
-struct _glapi_table;
 struct gl_context;
 
 
diff --git a/mesalib/src/mesa/main/points.c b/mesalib/src/mesa/main/points.c
index de26bf961..5ad1f38f3 100644
--- a/mesalib/src/mesa/main/points.c
+++ b/mesalib/src/mesa/main/points.c
@@ -103,7 +103,7 @@ _mesa_PointParameterfv( GLenum pname, const GLfloat *params)
     * If point parameters aren't supported, then this function shouldn't even
     * exist.
     */
-   ASSERT(!(ctx->Extensions.ARB_point_sprite
+   assert(!(ctx->Extensions.ARB_point_sprite
             || ctx->Extensions.NV_point_sprite)
           || ctx->Extensions.EXT_point_parameters);
 
@@ -253,7 +253,7 @@ _mesa_init_point(struct gl_context *ctx)
 
    ctx->Point.SpriteRMode = GL_ZERO; /* GL_NV_point_sprite (only!) */
    ctx->Point.SpriteOrigin = GL_UPPER_LEFT; /* GL_ARB_point_sprite */
-   for (i = 0; i < Elements(ctx->Point.CoordReplace); i++) {
+   for (i = 0; i < ARRAY_SIZE(ctx->Point.CoordReplace); i++) {
       ctx->Point.CoordReplace[i] = GL_FALSE; /* GL_ARB/NV_point_sprite */
    }
 }
diff --git a/mesalib/src/mesa/main/queryobj.c b/mesalib/src/mesa/main/queryobj.c
index 1b19afe4b..0842b540d 100644
--- a/mesalib/src/mesa/main/queryobj.c
+++ b/mesalib/src/mesa/main/queryobj.c
@@ -146,12 +146,13 @@ static struct gl_query_object **
 get_pipe_stats_binding_point(struct gl_context *ctx,
                              GLenum target)
 {
+   const int which = target - GL_VERTICES_SUBMITTED_ARB;
+   assert(which < MAX_PIPELINE_STATISTICS);
+
    if (!_mesa_is_desktop_gl(ctx) ||
        !ctx->Extensions.ARB_pipeline_statistics_query)
       return NULL;
 
-   const int which = target - GL_VERTICES_SUBMITTED_ARB;
-   assert(which < MAX_PIPELINE_STATISTICS);
    return &ctx->Query.pipeline_stats[which];
 }
 
@@ -405,6 +406,22 @@ _mesa_BeginQueryIndexed(GLenum target, GLuint index, GLuint id)
                      "glBeginQuery{Indexed}(query already active)");
          return;
       }
+
+      /* Section 2.14 Asynchronous Queries, page 84 of the OpenGL ES 3.0.4
+       * spec states:
+       *
+       *     "BeginQuery generates an INVALID_OPERATION error if any of the
+       *      following conditions hold: [...] id is the name of an
+       *      existing query object whose type does not match target; [...]
+       *
+       * Similar wording exists in the OpenGL 4.5 spec, section 4.2. QUERY
+       * OBJECTS AND ASYNCHRONOUS QUERIES, page 43.
+       */
+      if (q->EverBound && q->Target != target) {
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glBeginQuery{Indexed}(target mismatch)");
+         return;
+      }
    }
 
    q->Target = target;
@@ -666,7 +683,7 @@ _mesa_GetQueryObjectiv(GLuint id, GLenum pname, GLint *params)
    if (id)
       q = _mesa_lookup_query_object(ctx, id);
 
-   if (!q || q->Active) {
+   if (!q || q->Active || !q->EverBound) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glGetQueryObjectivARB(id=%d is invalid or active)", id);
       return;
@@ -717,7 +734,7 @@ _mesa_GetQueryObjectuiv(GLuint id, GLenum pname, GLuint *params)
    if (id)
       q = _mesa_lookup_query_object(ctx, id);
 
-   if (!q || q->Active) {
+   if (!q || q->Active || !q->EverBound) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glGetQueryObjectuivARB(id=%d is invalid or active)", id);
       return;
@@ -771,7 +788,7 @@ _mesa_GetQueryObjecti64v(GLuint id, GLenum pname, GLint64EXT *params)
    if (id)
       q = _mesa_lookup_query_object(ctx, id);
 
-   if (!q || q->Active) {
+   if (!q || q->Active || !q->EverBound) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glGetQueryObjectui64vARB(id=%d is invalid or active)", id);
       return;
@@ -811,7 +828,7 @@ _mesa_GetQueryObjectui64v(GLuint id, GLenum pname, GLuint64EXT *params)
    if (id)
       q = _mesa_lookup_query_object(ctx, id);
 
-   if (!q || q->Active) {
+   if (!q || q->Active || !q->EverBound) {
       _mesa_error(ctx, GL_INVALID_OPERATION,
                   "glGetQueryObjectuui64vARB(id=%d is invalid or active)", id);
       return;
diff --git a/mesalib/src/mesa/main/rastpos.c b/mesalib/src/mesa/main/rastpos.c
index 2027a9bd0..54b2125a8 100644
--- a/mesalib/src/mesa/main/rastpos.c
+++ b/mesalib/src/mesa/main/rastpos.c
@@ -266,7 +266,7 @@ window_pos3f(GLfloat x, GLfloat y, GLfloat z)
    {
       GLuint texSet;
       for (texSet = 0; texSet < ctx->Const.MaxTextureCoordUnits; texSet++) {
-         assert(texSet < Elements(ctx->Current.RasterTexCoords));
+         assert(texSet < ARRAY_SIZE(ctx->Current.RasterTexCoords));
          COPY_4FV( ctx->Current.RasterTexCoords[texSet],
                   ctx->Current.Attrib[VERT_ATTRIB_TEX0 + texSet] );
       }
@@ -496,7 +496,7 @@ void _mesa_init_rastpos( struct gl_context * ctx )
    ctx->Current.RasterDistance = 0.0;
    ASSIGN_4V( ctx->Current.RasterColor, 1.0, 1.0, 1.0, 1.0 );
    ASSIGN_4V( ctx->Current.RasterSecondaryColor, 0.0, 0.0, 0.0, 1.0 );
-   for (i = 0; i < Elements(ctx->Current.RasterTexCoords); i++)
+   for (i = 0; i < ARRAY_SIZE(ctx->Current.RasterTexCoords); i++)
       ASSIGN_4V( ctx->Current.RasterTexCoords[i], 0.0, 0.0, 0.0, 1.0 );
    ctx->Current.RasterPosValid = GL_TRUE;
 }
diff --git a/mesalib/src/mesa/main/rastpos.h b/mesalib/src/mesa/main/rastpos.h
index 2d36936de..dc28c68d4 100644
--- a/mesalib/src/mesa/main/rastpos.h
+++ b/mesalib/src/mesa/main/rastpos.h
@@ -32,7 +32,8 @@
 #define RASTPOS_H
 
 
-#include "compiler.h"
+#include "glheader.h"
+
 
 struct _glapi_table;
 struct gl_context;
diff --git a/mesalib/src/mesa/main/readpix.c b/mesalib/src/mesa/main/readpix.c
index ca4b9431b..ed0104c9e 100644
--- a/mesalib/src/mesa/main/readpix.c
+++ b/mesalib/src/mesa/main/readpix.c
@@ -127,7 +127,7 @@ _mesa_readpixels_needs_slow_path(const struct gl_context *ctx, GLenum format,
          _mesa_get_read_renderbuffer_for_format(ctx, format);
    GLenum srcType;
 
-   ASSERT(rb);
+   assert(rb);
 
    /* There are different rules depending on the base format. */
    switch (format) {
@@ -180,7 +180,7 @@ readpixels_can_use_memcpy(const struct gl_context *ctx, GLenum format, GLenum ty
    struct gl_renderbuffer *rb =
          _mesa_get_read_renderbuffer_for_format(ctx, format);
 
-   ASSERT(rb);
+   assert(rb);
 
    if (_mesa_readpixels_needs_slow_path(ctx, format, type, GL_FALSE)) {
       return GL_FALSE;
@@ -313,10 +313,10 @@ read_depth_pixels( struct gl_context *ctx,
       return;
 
    /* clipping should have been done already */
-   ASSERT(x >= 0);
-   ASSERT(y >= 0);
-   ASSERT(x + width <= (GLint) rb->Width);
-   ASSERT(y + height <= (GLint) rb->Height);
+   assert(x >= 0);
+   assert(y >= 0);
+   assert(x + width <= (GLint) rb->Width);
+   assert(y + height <= (GLint) rb->Height);
 
    if (type == GL_UNSIGNED_INT &&
        read_uint_depth_pixels(ctx, x, y, width, height, type, pixels, packing)) {
@@ -594,6 +594,7 @@ read_rgba_pixels( struct gl_context *ctx,
       _mesa_format_convert(dst, dst_format, dst_stride,
                            luminance, luminance_format, luminance_stride,
                            width, height, NULL);
+      free(luminance);
    } else {
       _mesa_pack_luminance_from_rgba_integer(width * height, src, !src_is_uint,
                                              dst, format, type);
diff --git a/mesalib/src/mesa/main/remap.c b/mesalib/src/mesa/main/remap.c
index 4a881865a..5a9538772 100644
--- a/mesalib/src/mesa/main/remap.c
+++ b/mesalib/src/mesa/main/remap.c
@@ -60,7 +60,7 @@ int driDispatchRemapTable[driDispatchRemapTable_size];
 const char *
 _mesa_get_function_spec(GLint func_index)
 {
-   if (func_index < Elements(_mesa_function_pool))
+   if (func_index < ARRAY_SIZE(_mesa_function_pool))
       return _mesa_function_pool + func_index;
    else
       return NULL;
@@ -197,7 +197,7 @@ _mesa_do_init_remap_table(const char *pool,
       const char *spec;
 
       /* sanity check */
-      ASSERT(i == remap[i].remap_index);
+      assert(i == remap[i].remap_index);
       spec = _mesa_function_pool + remap[i].pool_index;
 
       offset = _mesa_map_function_spec(spec);
diff --git a/mesalib/src/mesa/main/remap.h b/mesalib/src/mesa/main/remap.h
index 7199169d3..9d91b83c1 100644
--- a/mesalib/src/mesa/main/remap.h
+++ b/mesalib/src/mesa/main/remap.h
@@ -27,8 +27,6 @@
 #define REMAP_H
 
 
-#include "main/compiler.h"
-
 struct gl_function_pool_remap {
    int pool_index;
    int remap_index;
diff --git a/mesalib/src/mesa/main/renderbuffer.c b/mesalib/src/mesa/main/renderbuffer.c
index 98f3c13b5..e9d129a1d 100644
--- a/mesalib/src/mesa/main/renderbuffer.c
+++ b/mesalib/src/mesa/main/renderbuffer.c
@@ -172,7 +172,7 @@ _mesa_reference_renderbuffer_(struct gl_renderbuffer **ptr,
       struct gl_renderbuffer *oldRb = *ptr;
 
       mtx_lock(&oldRb->Mutex);
-      ASSERT(oldRb->RefCount > 0);
+      assert(oldRb->RefCount > 0);
       oldRb->RefCount--;
       /*printf("RB DECR %p (%d) to %d\n", (void*) oldRb, oldRb->Name, oldRb->RefCount);*/
       deleteFlag = (oldRb->RefCount == 0);
diff --git a/mesalib/src/mesa/main/samplerobj.c b/mesalib/src/mesa/main/samplerobj.c
index d62a06bf1..d66b0b5e8 100644
--- a/mesalib/src/mesa/main/samplerobj.c
+++ b/mesalib/src/mesa/main/samplerobj.c
@@ -89,7 +89,7 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
       struct gl_sampler_object *oldSamp = *ptr;
 
       /*mtx_lock(&oldSamp->Mutex);*/
-      ASSERT(oldSamp->RefCount > 0);
+      assert(oldSamp->RefCount > 0);
       oldSamp->RefCount--;
 #if 0
       printf("SamplerObj %p %d DECR to %d\n",
@@ -99,13 +99,13 @@ _mesa_reference_sampler_object_(struct gl_context *ctx,
       /*mtx_unlock(&oldSamp->Mutex);*/
 
       if (deleteFlag) {
-	 ASSERT(ctx->Driver.DeleteSamplerObject);
+	 assert(ctx->Driver.DeleteSamplerObject);
          ctx->Driver.DeleteSamplerObject(ctx, oldSamp);
       }
 
       *ptr = NULL;
    }
-   ASSERT(!*ptr);
+   assert(!*ptr);
 
    if (samp) {
       /* reference new sampler */
diff --git a/mesalib/src/mesa/main/shaderapi.c b/mesalib/src/mesa/main/shaderapi.c
index 52eab4655..5731d581a 100644
--- a/mesalib/src/mesa/main/shaderapi.c
+++ b/mesalib/src/mesa/main/shaderapi.c
@@ -523,7 +523,7 @@ get_programiv(struct gl_context *ctx, GLuint program, GLenum pname,
               GLint *params)
 {
    struct gl_shader_program *shProg
-      = _mesa_lookup_shader_program(ctx, program);
+      = _mesa_lookup_shader_program_err(ctx, program, "glGetProgramiv(program)");
 
    /* Is transform feedback available in this context?
     */
@@ -546,7 +546,6 @@ get_programiv(struct gl_context *ctx, GLuint program, GLenum pname,
       || _mesa_is_gles3(ctx);
 
    if (!shProg) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGetProgramiv(program)");
       return;
    }
 
@@ -764,11 +763,25 @@ static void
 get_program_info_log(struct gl_context *ctx, GLuint program, GLsizei bufSize,
                      GLsizei *length, GLchar *infoLog)
 {
-   struct gl_shader_program *shProg = _mesa_lookup_shader_program(ctx, program);
+   struct gl_shader_program *shProg;
+
+   /* Section 2.5 GL Errors (page 18) of the OpenGL ES 3.0.4 spec and
+    * section 2.3.1 (Errors) of the OpenGL 4.5 spec say:
+    *
+    *     "If a negative number is provided where an argument of type sizei or
+    *     sizeiptr is specified, an INVALID_VALUE error is generated."
+    */
+   if (bufSize < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glGetProgramInfoLog(bufSize < 0)");
+      return;
+   }
+
+   shProg = _mesa_lookup_shader_program_err(ctx, program,
+                                            "glGetProgramInfoLog(program)");
    if (!shProg) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGetProgramInfoLog(program)");
       return;
    }
+
    _mesa_copy_string(infoLog, bufSize, length, shProg->InfoLog);
 }
 
@@ -777,11 +790,24 @@ static void
 get_shader_info_log(struct gl_context *ctx, GLuint shader, GLsizei bufSize,
                     GLsizei *length, GLchar *infoLog)
 {
-   struct gl_shader *sh = _mesa_lookup_shader(ctx, shader);
+   struct gl_shader *sh;
+
+   /* Section 2.5 GL Errors (page 18) of the OpenGL ES 3.0.4 spec and
+    * section 2.3.1 (Errors) of the OpenGL 4.5 spec say:
+    *
+    *     "If a negative number is provided where an argument of type sizei or
+    *     sizeiptr is specified, an INVALID_VALUE error is generated."
+    */
+   if (bufSize < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glGetShaderInfoLog(bufSize < 0)");
+      return;
+   }
+
+   sh = _mesa_lookup_shader_err(ctx, shader, "glGetShaderInfoLog(shader)");
    if (!sh) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "glGetShaderInfoLog(shader)");
       return;
    }
+
    _mesa_copy_string(infoLog, bufSize, length, sh->InfoLog);
 }
 
@@ -1713,7 +1739,7 @@ _mesa_GetProgramBinary(GLuint program, GLsizei bufSize, GLsizei *length,
     * Ensure that length always points to valid storage to avoid multiple NULL
     * pointer checks below.
     */
-   if (length != NULL)
+   if (length == NULL)
       length = &length_dummy;
 
 
diff --git a/mesalib/src/mesa/main/shaderobj.c b/mesalib/src/mesa/main/shaderobj.c
index 02ccf450b..d7620c8ef 100644
--- a/mesalib/src/mesa/main/shaderobj.c
+++ b/mesalib/src/mesa/main/shaderobj.c
@@ -67,7 +67,7 @@ _mesa_reference_shader(struct gl_context *ctx, struct gl_shader **ptr,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_shader *old = *ptr;
 
-      ASSERT(old->RefCount > 0);
+      assert(old->RefCount > 0);
       old->RefCount--;
       /*printf("SHADER DECR %p (%d) to %d\n",
         (void*) old, old->Name, old->RefCount);*/
@@ -207,7 +207,7 @@ _mesa_reference_shader_program_(struct gl_context *ctx,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_shader_program *old = *ptr;
 
-      ASSERT(old->RefCount > 0);
+      assert(old->RefCount > 0);
       old->RefCount--;
 #if 0
       printf("ShaderProgram %p ID=%u  RefCount-- to %d\n",
diff --git a/mesalib/src/mesa/main/shaderobj.h b/mesalib/src/mesa/main/shaderobj.h
index 92f7a33ee..3d696a188 100644
--- a/mesalib/src/mesa/main/shaderobj.h
+++ b/mesalib/src/mesa/main/shaderobj.h
@@ -114,7 +114,7 @@ _mesa_shader_enum_to_shader_stage(GLenum v)
    case GL_COMPUTE_SHADER:
       return MESA_SHADER_COMPUTE;
    default:
-      ASSERT(0 && "bad value in _mesa_shader_enum_to_shader_stage()");
+      assert(0 && "bad value in _mesa_shader_enum_to_shader_stage()");
       return MESA_SHADER_VERTEX;
    }
 }
diff --git a/mesalib/src/mesa/main/shared.c b/mesalib/src/mesa/main/shared.c
index ccf5355f3..0b76cc012 100644
--- a/mesalib/src/mesa/main/shared.c
+++ b/mesalib/src/mesa/main/shared.c
@@ -105,7 +105,7 @@ _mesa_alloc_shared_state(struct gl_context *ctx)
          GL_TEXTURE_2D,
          GL_TEXTURE_1D
       };
-      STATIC_ASSERT(Elements(targets) == NUM_TEXTURE_TARGETS);
+      STATIC_ASSERT(ARRAY_SIZE(targets) == NUM_TEXTURE_TARGETS);
       shared->DefaultTex[i] = ctx->Driver.NewTextureObject(ctx, 0, targets[i]);
    }
 
@@ -159,7 +159,7 @@ delete_program_cb(GLuint id, void *data, void *userData)
    struct gl_program *prog = (struct gl_program *) data;
    struct gl_context *ctx = (struct gl_context *) userData;
    if(prog != &_mesa_DummyProgram) {
-      ASSERT(prog->RefCount == 1); /* should only be referenced by hash table */
+      assert(prog->RefCount == 1); /* should only be referenced by hash table */
       prog->RefCount = 0;  /* now going away */
       ctx->Driver.DeleteProgram(ctx, prog);
    }
@@ -223,7 +223,7 @@ delete_shader_cb(GLuint id, void *data, void *userData)
    }
    else {
       struct gl_shader_program *shProg = (struct gl_shader_program *) data;
-      ASSERT(shProg->Type == GL_SHADER_PROGRAM_MESA);
+      assert(shProg->Type == GL_SHADER_PROGRAM_MESA);
       ctx->Driver.DeleteShaderProgram(ctx, shProg);
    }
 }
@@ -346,7 +346,7 @@ free_shared_state(struct gl_context *ctx, struct gl_shared_state *shared)
     * Free texture objects (after FBOs since some textures might have
     * been bound to FBOs).
     */
-   ASSERT(ctx->Driver.DeleteTexture);
+   assert(ctx->Driver.DeleteTexture);
    /* the default textures */
    for (i = 0; i < NUM_TEXTURE_TARGETS; i++) {
       ctx->Driver.DeleteTexture(ctx, shared->DefaultTex[i]);
diff --git a/mesalib/src/mesa/main/state.c b/mesalib/src/mesa/main/state.c
index 45bce78fd..dadfb3c8c 100644
--- a/mesalib/src/mesa/main/state.c
+++ b/mesalib/src/mesa/main/state.c
@@ -275,7 +275,7 @@ update_viewport_matrix(struct gl_context *ctx)
    const GLfloat depthMax = ctx->DrawBuffer->_DepthMaxF;
    unsigned i;
 
-   ASSERT(depthMax > 0);
+   assert(depthMax > 0);
 
    /* Compute scale and bias values. This is really driver-specific
     * and should be maintained elsewhere if at all.
diff --git a/mesalib/src/mesa/main/stencil.h b/mesalib/src/mesa/main/stencil.h
index 6480b2adc..3302cb947 100644
--- a/mesalib/src/mesa/main/stencil.h
+++ b/mesalib/src/mesa/main/stencil.h
@@ -33,6 +33,7 @@
 
 
 #include "glheader.h"
+#include "macros.h"
 
 struct gl_context;
 
diff --git a/mesalib/src/mesa/main/texcompress.c b/mesalib/src/mesa/main/texcompress.c
index b4efeee3b..00234d4cc 100644
--- a/mesalib/src/mesa/main/texcompress.c
+++ b/mesalib/src/mesa/main/texcompress.c
@@ -570,8 +570,8 @@ _mesa_compressed_image_address(GLint col, GLint row, GLint img,
 
    _mesa_get_format_block_size(mesaFormat, &bw, &bh);
 
-   ASSERT(col % bw == 0);
-   ASSERT(row % bh == 0);
+   assert(col % bw == 0);
+   assert(row % bh == 0);
 
    offset = ((width + bw - 1) / bw) * (row / bh) + col / bw;
    offset *= blockSize;
diff --git a/mesalib/src/mesa/main/texcompress_bptc.c b/mesalib/src/mesa/main/texcompress_bptc.c
index c944ac26f..a600180e2 100644
--- a/mesalib/src/mesa/main/texcompress_bptc.c
+++ b/mesalib/src/mesa/main/texcompress_bptc.c
@@ -1630,7 +1630,7 @@ texstore_bptc_rgb_float(TEXSTORE_PARAMS,
 GLboolean
 _mesa_texstore_bptc_rgb_signed_float(TEXSTORE_PARAMS)
 {
-   ASSERT(dstFormat == MESA_FORMAT_BPTC_RGB_SIGNED_FLOAT);
+   assert(dstFormat == MESA_FORMAT_BPTC_RGB_SIGNED_FLOAT);
 
    return texstore_bptc_rgb_float(ctx, dims, baseInternalFormat,
                                   dstFormat, dstRowStride, dstSlices,
@@ -1643,7 +1643,7 @@ _mesa_texstore_bptc_rgb_signed_float(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_bptc_rgb_unsigned_float(TEXSTORE_PARAMS)
 {
-   ASSERT(dstFormat == MESA_FORMAT_BPTC_RGB_UNSIGNED_FLOAT);
+   assert(dstFormat == MESA_FORMAT_BPTC_RGB_UNSIGNED_FLOAT);
 
    return texstore_bptc_rgb_float(ctx, dims, baseInternalFormat,
                                   dstFormat, dstRowStride, dstSlices,
diff --git a/mesalib/src/mesa/main/texcompress_cpal.c b/mesalib/src/mesa/main/texcompress_cpal.c
index 54e9af41d..c46ef8c53 100644
--- a/mesalib/src/mesa/main/texcompress_cpal.c
+++ b/mesalib/src/mesa/main/texcompress_cpal.c
@@ -32,7 +32,6 @@
 
 
 #include "glheader.h"
-#include "compiler.h" /* for ASSERT */
 #include "context.h"
 #include "mtypes.h"
 #include "imports.h"
@@ -120,7 +119,7 @@ _mesa_cpal_compressed_size(int level, GLenum internalFormat,
    }
 
    info = &formats[internalFormat - GL_PALETTE4_RGB8_OES];
-   ASSERT(info->cpal_format == internalFormat);
+   assert(info->cpal_format == internalFormat);
 
    expect_size = info->palette_size * info->size;
    for (lvl = 0; lvl < num_levels; lvl++) {
diff --git a/mesalib/src/mesa/main/texcompress_etc.c b/mesalib/src/mesa/main/texcompress_etc.c
index 98b4fe201..f29561cc6 100644
--- a/mesalib/src/mesa/main/texcompress_etc.c
+++ b/mesalib/src/mesa/main/texcompress_etc.c
@@ -108,7 +108,7 @@ GLboolean
 _mesa_texstore_etc1_rgb8(TEXSTORE_PARAMS)
 {
    /* GL_ETC1_RGB8_OES is only valid in glCompressedTexImage2D */
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1099,7 +1099,7 @@ etc2_unpack_srgb8_punchthrough_alpha1(uint8_t *dst_row,
 GLboolean
 _mesa_texstore_etc2_rgb8(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1107,7 +1107,7 @@ _mesa_texstore_etc2_rgb8(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_srgb8(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1115,7 +1115,7 @@ _mesa_texstore_etc2_srgb8(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_rgba8_eac(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1123,7 +1123,7 @@ _mesa_texstore_etc2_rgba8_eac(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_srgb8_alpha8_eac(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1131,7 +1131,7 @@ _mesa_texstore_etc2_srgb8_alpha8_eac(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_r11_eac(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1139,7 +1139,7 @@ _mesa_texstore_etc2_r11_eac(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_signed_r11_eac(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1147,7 +1147,7 @@ _mesa_texstore_etc2_signed_r11_eac(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_rg11_eac(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1155,7 +1155,7 @@ _mesa_texstore_etc2_rg11_eac(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_signed_rg11_eac(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1163,7 +1163,7 @@ _mesa_texstore_etc2_signed_rg11_eac(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_rgb8_punchthrough_alpha1(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
@@ -1171,7 +1171,7 @@ _mesa_texstore_etc2_rgb8_punchthrough_alpha1(TEXSTORE_PARAMS)
 GLboolean
 _mesa_texstore_etc2_srgb8_punchthrough_alpha1(TEXSTORE_PARAMS)
 {
-   ASSERT(0);
+   assert(0);
 
    return GL_FALSE;
 }
diff --git a/mesalib/src/mesa/main/texcompress_fxt1.c b/mesalib/src/mesa/main/texcompress_fxt1.c
index 7b25e1039..562359410 100644
--- a/mesalib/src/mesa/main/texcompress_fxt1.c
+++ b/mesalib/src/mesa/main/texcompress_fxt1.c
@@ -61,7 +61,7 @@ _mesa_texstore_rgb_fxt1(TEXSTORE_PARAMS)
    GLubyte *dst;
    const GLubyte *tempImage = NULL;
 
-   ASSERT(dstFormat == MESA_FORMAT_RGB_FXT1);
+   assert(dstFormat == MESA_FORMAT_RGB_FXT1);
 
    if (srcFormat != GL_RGB ||
        srcType != GL_UNSIGNED_BYTE ||
@@ -116,7 +116,7 @@ _mesa_texstore_rgba_fxt1(TEXSTORE_PARAMS)
    GLubyte *dst;
    const GLubyte *tempImage = NULL;
 
-   ASSERT(dstFormat == MESA_FORMAT_RGBA_FXT1);
+   assert(dstFormat == MESA_FORMAT_RGBA_FXT1);
 
    if (srcFormat != GL_RGBA ||
        srcType != GL_UNSIGNED_BYTE ||
@@ -1264,12 +1264,12 @@ upscale_teximage2d(GLsizei inWidth, GLsizei inHeight,
 {
    GLint i, j, k;
 
-   ASSERT(outWidth >= inWidth);
-   ASSERT(outHeight >= inHeight);
+   assert(outWidth >= inWidth);
+   assert(outHeight >= inHeight);
 #if 0
-   ASSERT(inWidth == 1 || inWidth == 2 || inHeight == 1 || inHeight == 2);
-   ASSERT((outWidth & 3) == 0);
-   ASSERT((outHeight & 3) == 0);
+   assert(inWidth == 1 || inWidth == 2 || inHeight == 1 || inHeight == 2);
+   assert((outWidth & 3) == 0);
+   assert((outHeight & 3) == 0);
 #endif
 
    for (i = 0; i < outHeight; i++) {
diff --git a/mesalib/src/mesa/main/texcompress_rgtc.c b/mesalib/src/mesa/main/texcompress_rgtc.c
index e3042011a..f40e4e6c0 100644
--- a/mesalib/src/mesa/main/texcompress_rgtc.c
+++ b/mesalib/src/mesa/main/texcompress_rgtc.c
@@ -86,7 +86,7 @@ _mesa_texstore_red_rgtc1(TEXSTORE_PARAMS)
    GLint dstRowDiff, redRowStride;
    GLubyte *tempImageSlices[1];
 
-   ASSERT(dstFormat == MESA_FORMAT_R_RGTC1_UNORM ||
+   assert(dstFormat == MESA_FORMAT_R_RGTC1_UNORM ||
           dstFormat == MESA_FORMAT_L_LATC1_UNORM);
 
    tempImage = malloc(srcWidth * srcHeight * 1 * sizeof(GLubyte));
@@ -139,7 +139,7 @@ _mesa_texstore_signed_red_rgtc1(TEXSTORE_PARAMS)
    GLint dstRowDiff, redRowStride;
    GLfloat *tempImageSlices[1];
 
-   ASSERT(dstFormat == MESA_FORMAT_R_RGTC1_SNORM ||
+   assert(dstFormat == MESA_FORMAT_R_RGTC1_SNORM ||
           dstFormat == MESA_FORMAT_L_LATC1_SNORM);
 
    redRowStride = 1 * srcWidth * sizeof(GLfloat);
@@ -193,7 +193,7 @@ _mesa_texstore_rg_rgtc2(TEXSTORE_PARAMS)
    mesa_format tempFormat;
    GLubyte *tempImageSlices[1];
 
-   ASSERT(dstFormat == MESA_FORMAT_RG_RGTC2_UNORM ||
+   assert(dstFormat == MESA_FORMAT_RG_RGTC2_UNORM ||
           dstFormat == MESA_FORMAT_LA_LATC2_UNORM);
 
    if (baseInternalFormat == GL_RG)
@@ -258,7 +258,7 @@ _mesa_texstore_signed_rg_rgtc2(TEXSTORE_PARAMS)
    mesa_format tempFormat;
    GLfloat *tempImageSlices[1];
 
-   ASSERT(dstFormat == MESA_FORMAT_RG_RGTC2_SNORM ||
+   assert(dstFormat == MESA_FORMAT_RG_RGTC2_SNORM ||
           dstFormat == MESA_FORMAT_LA_LATC2_SNORM);
 
    if (baseInternalFormat == GL_RG)
diff --git a/mesalib/src/mesa/main/texcompress_s3tc.c b/mesalib/src/mesa/main/texcompress_s3tc.c
index bfb53dce4..41d00d4e9 100644
--- a/mesalib/src/mesa/main/texcompress_s3tc.c
+++ b/mesalib/src/mesa/main/texcompress_s3tc.c
@@ -133,7 +133,7 @@ _mesa_texstore_rgb_dxt1(TEXSTORE_PARAMS)
    GLubyte *dst;
    const GLubyte *tempImage = NULL;
 
-   ASSERT(dstFormat == MESA_FORMAT_RGB_DXT1 ||
+   assert(dstFormat == MESA_FORMAT_RGB_DXT1 ||
           dstFormat == MESA_FORMAT_SRGB_DXT1);
 
    if (srcFormat != GL_RGB ||
@@ -190,7 +190,7 @@ _mesa_texstore_rgba_dxt1(TEXSTORE_PARAMS)
    GLubyte *dst;
    const GLubyte *tempImage = NULL;
 
-   ASSERT(dstFormat == MESA_FORMAT_RGBA_DXT1 ||
+   assert(dstFormat == MESA_FORMAT_RGBA_DXT1 ||
           dstFormat == MESA_FORMAT_SRGBA_DXT1);
 
    if (srcFormat != GL_RGBA ||
@@ -247,7 +247,7 @@ _mesa_texstore_rgba_dxt3(TEXSTORE_PARAMS)
    GLubyte *dst;
    const GLubyte *tempImage = NULL;
 
-   ASSERT(dstFormat == MESA_FORMAT_RGBA_DXT3 ||
+   assert(dstFormat == MESA_FORMAT_RGBA_DXT3 ||
           dstFormat == MESA_FORMAT_SRGBA_DXT3);
 
    if (srcFormat != GL_RGBA ||
@@ -303,7 +303,7 @@ _mesa_texstore_rgba_dxt5(TEXSTORE_PARAMS)
    GLubyte *dst;
    const GLubyte *tempImage = NULL;
 
-   ASSERT(dstFormat == MESA_FORMAT_RGBA_DXT5 ||
+   assert(dstFormat == MESA_FORMAT_RGBA_DXT5 ||
           dstFormat == MESA_FORMAT_SRGBA_DXT5);
 
    if (srcFormat != GL_RGBA ||
diff --git a/mesalib/src/mesa/main/texcompress_s3tc.h b/mesalib/src/mesa/main/texcompress_s3tc.h
index 2734210a5..438b71fe3 100644
--- a/mesalib/src/mesa/main/texcompress_s3tc.h
+++ b/mesalib/src/mesa/main/texcompress_s3tc.h
@@ -25,7 +25,6 @@
 #ifndef TEXCOMPRESS_S3TC_H
 #define TEXCOMPRESS_S3TC_H
 
-#include "compiler.h"
 #include "glheader.h"
 #include "texstore.h"
 #include "texcompress.h"
diff --git a/mesalib/src/mesa/main/texformat.c b/mesalib/src/mesa/main/texformat.c
index ec16af9d2..3c4baca70 100644
--- a/mesalib/src/mesa/main/texformat.c
+++ b/mesalib/src/mesa/main/texformat.c
@@ -430,12 +430,12 @@ _mesa_choose_tex_format(struct gl_context *ctx, GLenum target,
 
    case GL_RGB9_E5:
       /* GL_EXT_texture_shared_exponent -- just one format to support */
-      ASSERT(ctx->TextureFormatSupported[MESA_FORMAT_R9G9B9E5_FLOAT]);
+      assert(ctx->TextureFormatSupported[MESA_FORMAT_R9G9B9E5_FLOAT]);
       return MESA_FORMAT_R9G9B9E5_FLOAT;
 
    case GL_R11F_G11F_B10F:
       /* GL_EXT_texture_packed_float -- just one format to support */
-      ASSERT(ctx->TextureFormatSupported[MESA_FORMAT_R11G11B10_FLOAT]);
+      assert(ctx->TextureFormatSupported[MESA_FORMAT_R11G11B10_FLOAT]);
       return MESA_FORMAT_R11G11B10_FLOAT;
 
    case GL_DEPTH_STENCIL_EXT:
@@ -445,10 +445,10 @@ _mesa_choose_tex_format(struct gl_context *ctx, GLenum target,
       break;
 
    case GL_DEPTH_COMPONENT32F:
-      ASSERT(ctx->TextureFormatSupported[MESA_FORMAT_Z_FLOAT32]);
+      assert(ctx->TextureFormatSupported[MESA_FORMAT_Z_FLOAT32]);
       return MESA_FORMAT_Z_FLOAT32;
    case GL_DEPTH32F_STENCIL8:
-      ASSERT(ctx->TextureFormatSupported[MESA_FORMAT_Z32_FLOAT_S8X24_UINT]);
+      assert(ctx->TextureFormatSupported[MESA_FORMAT_Z32_FLOAT_S8X24_UINT]);
       return MESA_FORMAT_Z32_FLOAT_S8X24_UINT;
 
    case GL_RED_SNORM:
diff --git a/mesalib/src/mesa/main/texgen.h b/mesalib/src/mesa/main/texgen.h
index aad5d2ecf..84adfc061 100644
--- a/mesalib/src/mesa/main/texgen.h
+++ b/mesalib/src/mesa/main/texgen.h
@@ -27,7 +27,6 @@
 #define TEXGEN_H
 
 
-#include "compiler.h"
 #include "glheader.h"
 
 struct _glapi_table;
diff --git a/mesalib/src/mesa/main/texgetimage.c b/mesalib/src/mesa/main/texgetimage.c
index 24df5b6f8..f975c160e 100644
--- a/mesalib/src/mesa/main/texgetimage.c
+++ b/mesalib/src/mesa/main/texgetimage.c
@@ -375,7 +375,7 @@ get_tex_rgba_uncompressed(struct gl_context *ctx, GLuint dimensions,
    GLuint height = texImage->Height;
    GLuint depth = texImage->Depth;
    GLuint img;
-   GLboolean dst_is_integer = _mesa_is_enum_format_integer(format);
+   GLboolean dst_is_integer;
    uint32_t dst_format;
    int dst_stride;
    uint8_t rebaseSwizzle[4];
@@ -423,6 +423,7 @@ get_tex_rgba_uncompressed(struct gl_context *ctx, GLuint dimensions,
     * integer then transferOps do not apply.
     */
    assert(!transferOps || (transferOps && !dst_is_integer));
+   (void) dst_is_integer; /* silence unused var warning */
 
    for (img = 0; img < depth; img++) {
       GLubyte *srcMap;
diff --git a/mesalib/src/mesa/main/teximage.c b/mesalib/src/mesa/main/teximage.c
index 29c325bf2..611d664b6 100644
--- a/mesalib/src/mesa/main/teximage.c
+++ b/mesalib/src/mesa/main/teximage.c
@@ -224,7 +224,10 @@ _mesa_base_tex_format( struct gl_context *ctx, GLint internalFormat )
    if (ctx->Extensions.ARB_stencil_texturing) {
       switch (internalFormat) {
       case GL_STENCIL_INDEX:
+      case GL_STENCIL_INDEX1:
+      case GL_STENCIL_INDEX4:
       case GL_STENCIL_INDEX8:
+      case GL_STENCIL_INDEX16:
          return GL_STENCIL_INDEX;
       default:
          ; /* fallthrough */
@@ -650,8 +653,8 @@ set_tex_image(struct gl_texture_object *tObj,
 {
    const GLuint face = _mesa_tex_target_to_face(target);
 
-   ASSERT(tObj);
-   ASSERT(texImage);
+   assert(tObj);
+   assert(texImage);
    if (target == GL_TEXTURE_RECTANGLE_NV || target == GL_TEXTURE_EXTERNAL_OES)
       assert(level == 0);
 
@@ -696,7 +699,7 @@ _mesa_delete_texture_image(struct gl_context *ctx,
    /* Free texImage->Data and/or any other driver-specific texture
     * image storage.
     */
-   ASSERT(ctx->Driver.FreeTextureImageBuffer);
+   assert(ctx->Driver.FreeTextureImageBuffer);
    ctx->Driver.FreeTextureImageBuffer( ctx, texImage );
    free(texImage);
 }
@@ -729,9 +732,9 @@ _mesa_is_proxy_texture(GLenum target)
     * NUM_TEXTURE_TARGETS should match number of terms above, except there's no
     * proxy for GL_TEXTURE_BUFFER and GL_TEXTURE_EXTERNAL_OES.
     */
-   STATIC_ASSERT(NUM_TEXTURE_TARGETS == Elements(targets) + 2);
+   STATIC_ASSERT(NUM_TEXTURE_TARGETS == ARRAY_SIZE(targets) + 2);
 
-   for (i = 0; i < Elements(targets); ++i)
+   for (i = 0; i < ARRAY_SIZE(targets); ++i)
       if (target == targets[i])
          return GL_TRUE;
    return GL_FALSE;
@@ -739,6 +742,28 @@ _mesa_is_proxy_texture(GLenum target)
 
 
 /**
+ * Test if a target is an array target.
+ *
+ * \param target texture target.
+ *
+ * \return true if the target is an array target, false otherwise.
+ */
+bool
+_mesa_is_array_texture(GLenum target)
+{
+   switch (target) {
+   case GL_TEXTURE_1D_ARRAY:
+   case GL_TEXTURE_2D_ARRAY:
+   case GL_TEXTURE_CUBE_MAP_ARRAY:
+   case GL_TEXTURE_2D_MULTISAMPLE_ARRAY:
+      return true;
+   default:
+      return false;
+   };
+}
+
+
+/**
  * Return the proxy target which corresponds to the given texture target
  */
 static GLenum
@@ -807,9 +832,9 @@ _mesa_select_tex_image(const struct gl_texture_object *texObj,
 {
    const GLuint face = _mesa_tex_target_to_face(target);
 
-   ASSERT(texObj);
-   ASSERT(level >= 0);
-   ASSERT(level < MAX_TEXTURE_LEVELS);
+   assert(texObj);
+   assert(level >= 0);
+   assert(level < MAX_TEXTURE_LEVELS);
 
    return texObj->Image[face][level];
 }
@@ -1239,7 +1264,7 @@ make_null_texture(GLint width, GLint height, GLint depth, GLenum format)
 static void
 clear_teximage_fields(struct gl_texture_image *img)
 {
-   ASSERT(img);
+   assert(img);
    img->_BaseFormat = 0;
    img->InternalFormat = 0;
    img->Border = 0;
@@ -1284,14 +1309,14 @@ init_teximage_fields_ms(struct gl_context *ctx,
                         GLuint numSamples, GLboolean fixedSampleLocations)
 {
    GLenum target;
-   ASSERT(img);
-   ASSERT(width >= 0);
-   ASSERT(height >= 0);
-   ASSERT(depth >= 0);
+   assert(img);
+   assert(width >= 0);
+   assert(height >= 0);
+   assert(depth >= 0);
 
    target = img->TexObject->Target;
    img->_BaseFormat = _mesa_base_tex_format( ctx, internalFormat );
-   ASSERT(img->_BaseFormat != -1);
+   assert(img->_BaseFormat != -1);
    img->InternalFormat = internalFormat;
    img->Border = border;
    img->Width = width;
@@ -2184,7 +2209,7 @@ texture_error_check( struct gl_context *ctx,
 
    /* additional checks for ycbcr textures */
    if (internalFormat == GL_YCBCR_MESA) {
-      ASSERT(ctx->Extensions.MESA_ycbcr_texture);
+      assert(ctx->Extensions.MESA_ycbcr_texture);
       if (type != GL_UNSIGNED_SHORT_8_8_MESA &&
           type != GL_UNSIGNED_SHORT_8_8_REV_MESA) {
          char message[100];
@@ -2491,8 +2516,8 @@ texsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
    }
 
    if (error_check_subtexture_dimensions(ctx, dimensions,
-                                         texImage, xoffset, yoffset, 0,
-                                         width, height, 1,
+                                         texImage, xoffset, yoffset, zoffset,
+                                         width, height, depth,
                                          dsa ? "glTextureSubImage" :
                                          "glTexSubImage")) {
       return GL_TRUE;
@@ -2815,14 +2840,6 @@ copytexsubimage_error_check(struct gl_context *ctx, GLuint dimensions,
       }
    }
 
-   /* check target (proxies not allowed) */
-   if (!legal_texsubimage_target(ctx, dimensions, target, dsa)) {
-      _mesa_error(ctx, GL_INVALID_ENUM, "glCopyTex%sSubImage%uD(target=%s)",
-                  suffix, dimensions,
-                  _mesa_lookup_enum_by_nr(target));
-      return GL_TRUE;
-   }
-
    /* Check level */
    if (level < 0 || level >= _mesa_max_texture_levels(ctx, target)) {
       _mesa_error(ctx, GL_INVALID_VALUE,
@@ -2933,7 +2950,7 @@ check_rtt_cb(GLuint key, void *data, void *userData)
              att->TextureLevel == level &&
              att->CubeMapFace == face) {
             _mesa_update_texture_renderbuffer(ctx, ctx->DrawBuffer, att);
-            ASSERT(att->Renderbuffer->TexImage);
+            assert(att->Renderbuffer->TexImage);
             /* Mark fb status as indeterminate to force re-validation */
             fb->_Status = 0;
          }
@@ -2977,7 +2994,7 @@ check_gen_mipmap(struct gl_context *ctx, GLenum target,
    if (texObj->GenerateMipmap &&
        level == texObj->BaseLevel &&
        level < texObj->MaxLevel) {
-      ASSERT(ctx->Driver.GenerateMipmap);
+      assert(ctx->Driver.GenerateMipmap);
       ctx->Driver.GenerateMipmap(ctx, target, texObj);
    }
 }
@@ -3054,7 +3071,7 @@ _mesa_choose_texture_format(struct gl_context *ctx,
           prevImage->Width > 0 &&
           prevImage->InternalFormat == internalFormat) {
          /* use the same format */
-         ASSERT(prevImage->TexFormat != MESA_FORMAT_NONE);
+         assert(prevImage->TexFormat != MESA_FORMAT_NONE);
          return prevImage->TexFormat;
       }
    }
@@ -3093,7 +3110,7 @@ _mesa_choose_texture_format(struct gl_context *ctx,
    /* choose format from scratch */
    f = ctx->Driver.ChooseTextureFormat(ctx, target, internalFormat,
                                        format, type);
-   ASSERT(f != MESA_FORMAT_NONE);
+   assert(f != MESA_FORMAT_NONE);
    return f;
 }
 
@@ -4088,6 +4105,16 @@ _mesa_CopyTexSubImage1D( GLenum target, GLint level,
    struct gl_texture_object* texObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   /* Check target (proxies not allowed). Target must be checked prior to
+    * calling _mesa_get_current_tex_object.
+    */
+   if (!legal_texsubimage_target(ctx, 1, target, false)) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glCopyTexSubImage1D(invalid target %s)",
+                  _mesa_lookup_enum_by_nr(target));
+      return;
+   }
+
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
@@ -4106,6 +4133,16 @@ _mesa_CopyTexSubImage2D( GLenum target, GLint level,
    struct gl_texture_object* texObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   /* Check target (proxies not allowed). Target must be checked prior to
+    * calling _mesa_get_current_tex_object.
+    */
+   if (!legal_texsubimage_target(ctx, 2, target, false)) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glCopyTexSubImage2D(invalid target %s)",
+                  _mesa_lookup_enum_by_nr(target));
+      return;
+   }
+
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
@@ -4125,6 +4162,16 @@ _mesa_CopyTexSubImage3D( GLenum target, GLint level,
    struct gl_texture_object* texObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   /* Check target (proxies not allowed). Target must be checked prior to
+    * calling _mesa_get_current_tex_object.
+    */
+   if (!legal_texsubimage_target(ctx, 3, target, false)) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glCopyTexSubImage3D(invalid target %s)",
+                  _mesa_lookup_enum_by_nr(target));
+      return;
+   }
+
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
@@ -4145,6 +4192,14 @@ _mesa_CopyTextureSubImage1D(GLuint texture, GLint level,
    if (!texObj)
       return;
 
+   /* Check target (proxies not allowed). */
+   if (!legal_texsubimage_target(ctx, 1, texObj->Target, true)) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glCopyTextureSubImage1D(invalid target %s)",
+                  _mesa_lookup_enum_by_nr(texObj->Target));
+      return;
+   }
+
    _mesa_copy_texture_sub_image(ctx, 1, texObj, texObj->Target, level,
                                 xoffset, 0, 0, x, y, width, 1, true);
 }
@@ -4161,6 +4216,14 @@ _mesa_CopyTextureSubImage2D(GLuint texture, GLint level,
    if (!texObj)
       return;
 
+   /* Check target (proxies not allowed). */
+   if (!legal_texsubimage_target(ctx, 2, texObj->Target, true)) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glCopyTextureSubImage2D(invalid target %s)",
+                  _mesa_lookup_enum_by_nr(texObj->Target));
+      return;
+   }
+
    _mesa_copy_texture_sub_image(ctx, 2, texObj, texObj->Target, level,
                                 xoffset, yoffset, 0,
                                 x, y, width, height, true);
@@ -4180,6 +4243,14 @@ _mesa_CopyTextureSubImage3D(GLuint texture, GLint level,
    if (!texObj)
       return;
 
+   /* Check target (proxies not allowed). */
+   if (!legal_texsubimage_target(ctx, 3, texObj->Target, true)) {
+      _mesa_error(ctx, GL_INVALID_ENUM,
+                  "glCopyTextureSubImage3D(invalid target %s)",
+                  _mesa_lookup_enum_by_nr(texObj->Target));
+      return;
+   }
+
    _mesa_copy_texture_sub_image(ctx, 3, texObj, texObj->Target, level,
                                 xoffset, yoffset, zoffset,
                                 x, y, width, height, true);
@@ -4454,25 +4525,21 @@ out:
 
 
 /**
- * Error checking for glCompressedTexSubImage[123]D().
+ * Target checking for glCompressedTexSubImage[123]D().
  * \return GL_TRUE if error, GL_FALSE if no error
+ * Must come before other error checking so that the texture object can
+ * be correctly retrieved using _mesa_get_current_tex_object.
  */
 static GLboolean
-compressed_subtexture_error_check(struct gl_context *ctx, GLint dims,
-                                  const struct gl_texture_object *texObj,
-                                  GLenum target, GLint level,
-                                  GLint xoffset, GLint yoffset, GLint zoffset,
-                                  GLsizei width, GLsizei height, GLsizei depth,
-                                  GLenum format, GLsizei imageSize, bool dsa)
+compressed_subtexture_target_check(struct gl_context *ctx, GLenum target,
+                                   GLint dims, GLenum format, bool dsa,
+                                   const char *caller)
 {
-   struct gl_texture_image *texImage;
-   GLint expectedSize;
    GLboolean targetOK;
-   const char *suffix = dsa ? "ture" : "";
 
    if (dsa && target == GL_TEXTURE_RECTANGLE) {
-      _mesa_error(ctx, GL_INVALID_OPERATION,
-                  "glCompressedSubTexture%dD(target)", dims);
+      _mesa_error(ctx, GL_INVALID_OPERATION, "%s(invalid target %s)", caller,
+                  _mesa_lookup_enum_by_nr(target));
       return GL_TRUE;
    }
 
@@ -4535,7 +4602,9 @@ compressed_subtexture_error_check(struct gl_context *ctx, GLint dims,
          }
          if (invalidformat) {
             _mesa_error(ctx, GL_INVALID_OPERATION,
-                        "glCompressedTex%sSubImage%uD(target)", suffix, dims);
+                        "%s(invalid target %s for format %s)", caller,
+                        _mesa_lookup_enum_by_nr(target),
+                        _mesa_lookup_enum_by_nr(format));
             return GL_TRUE;
          }
       }
@@ -4549,11 +4618,30 @@ compressed_subtexture_error_check(struct gl_context *ctx, GLint dims,
    }
 
    if (!targetOK) {
-      _mesa_error(ctx, GL_INVALID_ENUM,
-                  "glCompressedTex%sSubImage%uD(target)", suffix, dims);
+      _mesa_error(ctx, GL_INVALID_ENUM, "%s(invalid target %s)", caller,
+                  _mesa_lookup_enum_by_nr(target));
       return GL_TRUE;
    }
 
+   return GL_FALSE;
+}
+
+/**
+ * Error checking for glCompressedTexSubImage[123]D().
+ * \return GL_TRUE if error, GL_FALSE if no error
+ */
+static GLboolean
+compressed_subtexture_error_check(struct gl_context *ctx, GLint dims,
+                                  const struct gl_texture_object *texObj,
+                                  GLenum target, GLint level,
+                                  GLint xoffset, GLint yoffset, GLint zoffset,
+                                  GLsizei width, GLsizei height, GLsizei depth,
+                                  GLenum format, GLsizei imageSize, bool dsa)
+{
+   struct gl_texture_image *texImage;
+   GLint expectedSize;
+   const char *suffix = dsa ? "ture" : "";
+
    /* this will catch any invalid compressed format token */
    if (!_mesa_is_compressed_format(ctx, format)) {
       _mesa_error(ctx, GL_INVALID_ENUM,
@@ -4709,6 +4797,11 @@ _mesa_CompressedTexSubImage1D(GLenum target, GLint level, GLint xoffset,
    struct gl_texture_object *texObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   if (compressed_subtexture_target_check(ctx, target, 1, format, false,
+                                          "glCompressedTexSubImage1D")) {
+      return;
+   }
+
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
@@ -4731,6 +4824,12 @@ _mesa_CompressedTextureSubImage1D(GLuint texture, GLint level, GLint xoffset,
    if (!texObj)
       return;
 
+   if (compressed_subtexture_target_check(ctx, texObj->Target, 1, format,
+                                          true,
+                                          "glCompressedTextureSubImage1D")) {
+      return;
+   }
+
    _mesa_compressed_texture_sub_image(ctx, 1, texObj, texObj->Target, level,
                                       xoffset, 0, 0, width, 1, 1,
                                       format, imageSize, data, true);
@@ -4746,6 +4845,11 @@ _mesa_CompressedTexSubImage2D(GLenum target, GLint level, GLint xoffset,
    struct gl_texture_object *texObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   if (compressed_subtexture_target_check(ctx, target, 2, format, false,
+                                          "glCompressedTexSubImage2D")) {
+      return;
+   }
+
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
@@ -4770,6 +4874,12 @@ _mesa_CompressedTextureSubImage2D(GLuint texture, GLint level, GLint xoffset,
    if (!texObj)
       return;
 
+   if (compressed_subtexture_target_check(ctx, texObj->Target, 2, format,
+                                          true,
+                                          "glCompressedTextureSubImage2D")) {
+      return;
+   }
+
    _mesa_compressed_texture_sub_image(ctx, 2, texObj, texObj->Target, level,
                                       xoffset, yoffset, 0, width, height, 1,
                                       format, imageSize, data, true);
@@ -4784,6 +4894,11 @@ _mesa_CompressedTexSubImage3D(GLenum target, GLint level, GLint xoffset,
    struct gl_texture_object *texObj;
    GET_CURRENT_CONTEXT(ctx);
 
+   if (compressed_subtexture_target_check(ctx, target, 3, format, false,
+                                          "glCompressedTexSubImage3D")) {
+      return;
+   }
+
    texObj = _mesa_get_current_tex_object(ctx, target);
    if (!texObj)
       return;
@@ -4809,6 +4924,12 @@ _mesa_CompressedTextureSubImage3D(GLuint texture, GLint level, GLint xoffset,
    if (!texObj)
       return;
 
+   if (compressed_subtexture_target_check(ctx, texObj->Target, 3, format,
+                                          true,
+                                          "glCompressedTextureSubImage3D")) {
+      return;
+   }
+
    _mesa_compressed_texture_sub_image(ctx, 3, texObj, texObj->Target, level,
                                       xoffset, yoffset, zoffset,
                                       width, height, depth,
@@ -5031,7 +5152,7 @@ _mesa_validate_texbuffer_format(const struct gl_context *ctx,
 
 void
 _mesa_texture_buffer_range(struct gl_context *ctx,
-                           struct gl_texture_object *texObj, GLenum target, 
+                           struct gl_texture_object *texObj, GLenum target,
                            GLenum internalFormat,
                            struct gl_buffer_object *bufObj,
                            GLintptr offset, GLsizeiptr size, bool range,
diff --git a/mesalib/src/mesa/main/teximage.h b/mesalib/src/mesa/main/teximage.h
index 02b0eda38..b7336bc6c 100644
--- a/mesalib/src/mesa/main/teximage.h
+++ b/mesalib/src/mesa/main/teximage.h
@@ -66,6 +66,9 @@ _mesa_base_tex_format( struct gl_context *ctx, GLint internalFormat );
 extern GLboolean
 _mesa_is_proxy_texture(GLenum target);
 
+extern bool
+_mesa_is_array_texture(GLenum target);
+
 extern struct gl_texture_image *
 _mesa_new_texture_image( struct gl_context *ctx );
 
diff --git a/mesalib/src/mesa/main/texobj.c b/mesalib/src/mesa/main/texobj.c
index 59090db4e..e018ab910 100644
--- a/mesalib/src/mesa/main/texobj.c
+++ b/mesalib/src/mesa/main/texobj.c
@@ -28,6 +28,7 @@
  */
 
 
+#include <stdio.h>
 #include "bufferobj.h"
 #include "colortab.h"
 #include "context.h"
@@ -265,7 +266,7 @@ _mesa_initialize_texture_object( struct gl_context *ctx,
                                  struct gl_texture_object *obj,
                                  GLuint name, GLenum target )
 {
-   ASSERT(target == 0 ||
+   assert(target == 0 ||
           target == GL_TEXTURE_1D ||
           target == GL_TEXTURE_2D ||
           target == GL_TEXTURE_3D ||
@@ -541,11 +542,11 @@ _mesa_reference_texobj_(struct gl_texture_object **ptr,
       GLboolean deleteFlag = GL_FALSE;
       struct gl_texture_object *oldTex = *ptr;
 
-      ASSERT(valid_texture_object(oldTex));
+      assert(valid_texture_object(oldTex));
       (void) valid_texture_object; /* silence warning in release builds */
 
       mtx_lock(&oldTex->Mutex);
-      ASSERT(oldTex->RefCount > 0);
+      assert(oldTex->RefCount > 0);
       oldTex->RefCount--;
 
       deleteFlag = (oldTex->RefCount == 0);
@@ -568,7 +569,7 @@ _mesa_reference_texobj_(struct gl_texture_object **ptr,
 
    if (tex) {
       /* reference new texture */
-      ASSERT(valid_texture_object(tex));
+      assert(valid_texture_object(tex));
       mtx_lock(&tex->Mutex);
       if (tex->RefCount == 0) {
          /* this texture's being deleted (look just above) */
@@ -730,7 +731,7 @@ _mesa_test_texobj_completeness( const struct gl_context *ctx,
       return;
    }
 
-   ASSERT(maxLevels > 0);
+   assert(maxLevels > 0);
 
    t->_MaxLevel = MIN3(t->MaxLevel,
                        /* 'p' in the GL spec */
@@ -1724,7 +1725,7 @@ _mesa_BindTexture( GLenum target, GLuint texName )
    _mesa_reference_texobj(&texUnit->CurrentTex[targetIndex], newTexObj);
    ctx->Texture.NumCurrentTexUsed = MAX2(ctx->Texture.NumCurrentTexUsed,
                                          ctx->Texture.CurrentUnit + 1);
-   ASSERT(texUnit->CurrentTex[targetIndex]);
+   assert(texUnit->CurrentTex[targetIndex]);
 
    if (texName != 0)
       texUnit->_BoundTextures |= (1 << targetIndex);
@@ -1773,7 +1774,7 @@ _mesa_bind_texture_unit(struct gl_context *ctx,
 
    _mesa_reference_texobj(&texUnit->CurrentTex[texObj->TargetIndex],
                           texObj);
-   ASSERT(texUnit->CurrentTex[texObj->TargetIndex]);
+   assert(texUnit->CurrentTex[texObj->TargetIndex]);
    ctx->Texture.NumCurrentTexUsed = MAX2(ctx->Texture.NumCurrentTexUsed,
                                          unit + 1);
    texUnit->_BoundTextures |= (1 << texObj->TargetIndex);
diff --git a/mesalib/src/mesa/main/texparam.c b/mesalib/src/mesa/main/texparam.c
index c4a5841c5..1ef728660 100644
--- a/mesalib/src/mesa/main/texparam.c
+++ b/mesalib/src/mesa/main/texparam.c
@@ -217,8 +217,8 @@ comp_to_swizzle(GLenum comp)
 static void
 set_swizzle_component(GLuint *swizzle, GLuint comp, GLuint swz)
 {
-   ASSERT(comp < 4);
-   ASSERT(swz <= SWIZZLE_NIL);
+   assert(comp < 4);
+   assert(swz <= SWIZZLE_NIL);
    {
       GLuint mask = 0x7 << (3 * comp);
       GLuint s = (*swizzle & ~mask) | (swz << (3 * comp));
@@ -537,7 +537,7 @@ set_tex_parameteri(struct gl_context *ctx,
                         "glTex%sParameter(swizzle 0x%x)", suffix, params[0]);
             return GL_FALSE;
          }
-         ASSERT(comp < 4);
+         assert(comp < 4);
 
          flush(ctx);
          texObj->Swizzle[comp] = params[0];
diff --git a/mesalib/src/mesa/main/texstate.c b/mesalib/src/mesa/main/texstate.c
index 99c7c8178..0a7f983fb 100644
--- a/mesalib/src/mesa/main/texstate.c
+++ b/mesalib/src/mesa/main/texstate.c
@@ -28,6 +28,7 @@
  * Texture state handling.
  */
 
+#include <stdio.h>
 #include "glheader.h"
 #include "bufferobj.h"
 #include "colormac.h"
@@ -68,8 +69,8 @@ _mesa_copy_texture_state( const struct gl_context *src, struct gl_context *dst )
 {
    GLuint u, tex;
 
-   ASSERT(src);
-   ASSERT(dst);
+   assert(src);
+   assert(dst);
 
    dst->Texture.CurrentUnit = src->Texture.CurrentUnit;
    dst->Texture._GenFlags = src->Texture._GenFlags;
@@ -292,7 +293,7 @@ _mesa_ActiveTexture(GLenum texture)
 
    k = _mesa_max_tex_unit(ctx);
 
-   ASSERT(k <= Elements(ctx->Texture.Unit));
+   assert(k <= ARRAY_SIZE(ctx->Texture.Unit));
 
    if (MESA_VERBOSE & (VERBOSE_API|VERBOSE_TEXTURE))
       _mesa_debug(ctx, "glActiveTexture %s\n",
@@ -363,7 +364,7 @@ update_texture_matrices( struct gl_context *ctx )
    ctx->Texture._TexMatEnabled = 0x0;
 
    for (u = 0; u < ctx->Const.MaxTextureCoordUnits; u++) {
-      ASSERT(u < Elements(ctx->TextureMatrixStack));
+      assert(u < ARRAY_SIZE(ctx->TextureMatrixStack));
       if (_math_matrix_is_dirty(ctx->TextureMatrixStack[u].Top)) {
 	 _math_matrix_analyse( ctx->TextureMatrixStack[u].Top );
 
@@ -501,7 +502,7 @@ update_texgen(struct gl_context *ctx)
 	 ctx->Texture._GenFlags |= texUnit->_GenFlags;
       }
 
-      ASSERT(unit < Elements(ctx->TextureMatrixStack));
+      assert(unit < ARRAY_SIZE(ctx->TextureMatrixStack));
       if (ctx->TextureMatrixStack[unit].Top->type != MATRIX_IDENTITY)
 	 ctx->Texture._TexMatEnabled |= ENABLE_TEXMAT(unit);
    }
@@ -797,7 +798,7 @@ alloc_proxy_textures( struct gl_context *ctx )
    };
    GLint tgt;
 
-   STATIC_ASSERT(Elements(targets) == NUM_TEXTURE_TARGETS);
+   STATIC_ASSERT(ARRAY_SIZE(targets) == NUM_TEXTURE_TARGETS);
    assert(targets[TEXTURE_2D_INDEX] == GL_TEXTURE_2D);
    assert(targets[TEXTURE_CUBE_INDEX] == GL_TEXTURE_CUBE_MAP);
 
@@ -882,10 +883,19 @@ _mesa_init_texture(struct gl_context *ctx)
     *     "OpenGL ES 3.0 requires that all cube map filtering be
     *     seamless. OpenGL ES 2.0 specified that a single cube map face be
     *     selected and used for filtering."
+    *
+    * Unfortunatley, a call to _mesa_is_gles3 below will only work if
+    * the driver has already computed and set ctx->Version, however drivers
+    * seem to call _mesa_initialize_context (which calls this) early
+    * in the CreateContext hook and _mesa_compute_version much later (since
+    * it needs information about available extensions). So, we will
+    * enable seamless cubemaps by default since GLES2. This should work
+    * for most implementations and drivers that don't support seamless
+    * cubemaps for GLES2 can still disable it.
     */
-   ctx->Texture.CubeMapSeamless = _mesa_is_gles3(ctx);
+   ctx->Texture.CubeMapSeamless = ctx->API == API_OPENGLES2;
 
-   for (u = 0; u < Elements(ctx->Texture.Unit); u++)
+   for (u = 0; u < ARRAY_SIZE(ctx->Texture.Unit); u++)
       init_texture_unit(ctx, u);
 
    /* After we're done initializing the context's texture state the default
@@ -918,7 +928,7 @@ _mesa_free_texture_data(struct gl_context *ctx)
    GLuint u, tgt;
 
    /* unreference current textures */
-   for (u = 0; u < Elements(ctx->Texture.Unit); u++) {
+   for (u = 0; u < ARRAY_SIZE(ctx->Texture.Unit); u++) {
       /* The _Current texture could account for another reference */
       _mesa_reference_texobj(&ctx->Texture.Unit[u]._Current, NULL);
 
@@ -934,7 +944,7 @@ _mesa_free_texture_data(struct gl_context *ctx)
    /* GL_ARB_texture_buffer_object */
    _mesa_reference_buffer_object(ctx, &ctx->Texture.BufferObject, NULL);
 
-   for (u = 0; u < Elements(ctx->Texture.Unit); u++) {
+   for (u = 0; u < ARRAY_SIZE(ctx->Texture.Unit); u++) {
       _mesa_reference_sampler_object(ctx, &ctx->Texture.Unit[u].Sampler, NULL);
    }
 }
@@ -950,7 +960,7 @@ _mesa_update_default_objects_texture(struct gl_context *ctx)
 {
    GLuint u, tex;
 
-   for (u = 0; u < Elements(ctx->Texture.Unit); u++) {
+   for (u = 0; u < ARRAY_SIZE(ctx->Texture.Unit); u++) {
       struct gl_texture_unit *texUnit = &ctx->Texture.Unit[u];
       for (tex = 0; tex < NUM_TEXTURE_TARGETS; tex++) {
          _mesa_reference_texobj(&texUnit->CurrentTex[tex],
diff --git a/mesalib/src/mesa/main/texstate.h b/mesalib/src/mesa/main/texstate.h
index abc07eafb..662435b47 100644
--- a/mesalib/src/mesa/main/texstate.h
+++ b/mesalib/src/mesa/main/texstate.h
@@ -41,7 +41,7 @@
 static inline struct gl_texture_unit *
 _mesa_get_tex_unit(struct gl_context *ctx, GLuint unit)
 {
-   ASSERT(unit < Elements(ctx->Texture.Unit));
+   assert(unit < ARRAY_SIZE(ctx->Texture.Unit));
    return &(ctx->Texture.Unit[unit]);
 }
 
diff --git a/mesalib/src/mesa/main/texstorage.c b/mesalib/src/mesa/main/texstorage.c
index 3ace5e8bb..53cb2c091 100644
--- a/mesalib/src/mesa/main/texstorage.c
+++ b/mesalib/src/mesa/main/texstorage.c
@@ -171,7 +171,7 @@ clear_texture_fields(struct gl_context *ctx,
    GLint level;
    GLuint face;
 
-   for (level = 0; level < Elements(texObj->Image[0]); level++) {
+   for (level = 0; level < ARRAY_SIZE(texObj->Image[0]); level++) {
       for (face = 0; face < numFaces; face++) {
          struct gl_texture_image *texImage =
             get_tex_image(ctx, texObj, face, level);
diff --git a/mesalib/src/mesa/main/texstore.c b/mesalib/src/mesa/main/texstore.c
index 7039cdf81..7ad9d2b68 100644
--- a/mesalib/src/mesa/main/texstore.c
+++ b/mesalib/src/mesa/main/texstore.c
@@ -154,9 +154,9 @@ _mesa_texstore_z32(TEXSTORE_PARAMS)
    const GLuint depthScale = 0xffffffff;
    GLenum dstType;
    (void) dims;
-   ASSERT(dstFormat == MESA_FORMAT_Z_UNORM32 ||
+   assert(dstFormat == MESA_FORMAT_Z_UNORM32 ||
           dstFormat == MESA_FORMAT_Z_FLOAT32);
-   ASSERT(_mesa_get_format_bytes(dstFormat) == sizeof(GLuint));
+   assert(_mesa_get_format_bytes(dstFormat) == sizeof(GLuint));
 
    if (dstFormat == MESA_FORMAT_Z_UNORM32)
       dstType = GL_UNSIGNED_INT;
@@ -191,7 +191,7 @@ _mesa_texstore_x8_z24(TEXSTORE_PARAMS)
    const GLuint depthScale = 0xffffff;
 
    (void) dims;
-   ASSERT(dstFormat == MESA_FORMAT_Z24_UNORM_X8_UINT);
+   assert(dstFormat == MESA_FORMAT_Z24_UNORM_X8_UINT);
 
    {
       /* general path */
@@ -221,7 +221,7 @@ _mesa_texstore_z24_x8(TEXSTORE_PARAMS)
    const GLuint depthScale = 0xffffff;
 
    (void) dims;
-   ASSERT(dstFormat == MESA_FORMAT_X8_UINT_Z24_UNORM);
+   assert(dstFormat == MESA_FORMAT_X8_UINT_Z24_UNORM);
 
    {
       /* general path */
@@ -254,8 +254,8 @@ _mesa_texstore_z16(TEXSTORE_PARAMS)
 {
    const GLuint depthScale = 0xffff;
    (void) dims;
-   ASSERT(dstFormat == MESA_FORMAT_Z_UNORM16);
-   ASSERT(_mesa_get_format_bytes(dstFormat) == sizeof(GLushort));
+   assert(dstFormat == MESA_FORMAT_Z_UNORM16);
+   assert(_mesa_get_format_bytes(dstFormat) == sizeof(GLushort));
 
    {
       /* general path */
@@ -287,14 +287,14 @@ _mesa_texstore_ycbcr(TEXSTORE_PARAMS)
 
    (void) ctx; (void) dims; (void) baseInternalFormat;
 
-   ASSERT((dstFormat == MESA_FORMAT_YCBCR) ||
+   assert((dstFormat == MESA_FORMAT_YCBCR) ||
           (dstFormat == MESA_FORMAT_YCBCR_REV));
-   ASSERT(_mesa_get_format_bytes(dstFormat) == 2);
-   ASSERT(ctx->Extensions.MESA_ycbcr_texture);
-   ASSERT(srcFormat == GL_YCBCR_MESA);
-   ASSERT((srcType == GL_UNSIGNED_SHORT_8_8_MESA) ||
+   assert(_mesa_get_format_bytes(dstFormat) == 2);
+   assert(ctx->Extensions.MESA_ycbcr_texture);
+   assert(srcFormat == GL_YCBCR_MESA);
+   assert((srcType == GL_UNSIGNED_SHORT_8_8_MESA) ||
           (srcType == GL_UNSIGNED_SHORT_8_8_REV_MESA));
-   ASSERT(baseInternalFormat == GL_YCBCR_MESA);
+   assert(baseInternalFormat == GL_YCBCR_MESA);
 
    /* always just memcpy since no pixel transfer ops apply */
    memcpy_texture(ctx, dims,
@@ -335,11 +335,11 @@ _mesa_texstore_z24_s8(TEXSTORE_PARAMS)
    GLuint *depth = malloc(srcWidth * sizeof(GLuint));
    GLubyte *stencil = malloc(srcWidth * sizeof(GLubyte));
 
-   ASSERT(dstFormat == MESA_FORMAT_S8_UINT_Z24_UNORM);
-   ASSERT(srcFormat == GL_DEPTH_STENCIL_EXT ||
+   assert(dstFormat == MESA_FORMAT_S8_UINT_Z24_UNORM);
+   assert(srcFormat == GL_DEPTH_STENCIL_EXT ||
           srcFormat == GL_DEPTH_COMPONENT ||
           srcFormat == GL_STENCIL_INDEX);
-   ASSERT(srcFormat != GL_DEPTH_STENCIL_EXT ||
+   assert(srcFormat != GL_DEPTH_STENCIL_EXT ||
           srcType == GL_UNSIGNED_INT_24_8_EXT ||
           srcType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
@@ -414,11 +414,11 @@ _mesa_texstore_s8_z24(TEXSTORE_PARAMS)
    GLuint *depth;
    GLubyte *stencil;
 
-   ASSERT(dstFormat == MESA_FORMAT_Z24_UNORM_S8_UINT);
-   ASSERT(srcFormat == GL_DEPTH_STENCIL_EXT ||
+   assert(dstFormat == MESA_FORMAT_Z24_UNORM_S8_UINT);
+   assert(srcFormat == GL_DEPTH_STENCIL_EXT ||
           srcFormat == GL_DEPTH_COMPONENT ||
           srcFormat == GL_STENCIL_INDEX);
-   ASSERT(srcFormat != GL_DEPTH_STENCIL_EXT ||
+   assert(srcFormat != GL_DEPTH_STENCIL_EXT ||
           srcType == GL_UNSIGNED_INT_24_8_EXT ||
           srcType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
@@ -491,8 +491,8 @@ _mesa_texstore_s8_z24(TEXSTORE_PARAMS)
 static GLboolean
 _mesa_texstore_s8(TEXSTORE_PARAMS)
 {
-   ASSERT(dstFormat == MESA_FORMAT_S_UINT8);
-   ASSERT(srcFormat == GL_STENCIL_INDEX);
+   assert(dstFormat == MESA_FORMAT_S_UINT8);
+   assert(srcFormat == GL_STENCIL_INDEX);
 
    {
       const GLint srcRowStride
@@ -543,11 +543,11 @@ _mesa_texstore_z32f_x24s8(TEXSTORE_PARAMS)
       = _mesa_image_row_stride(srcPacking, srcWidth, srcFormat, srcType)
       / sizeof(uint64_t);
 
-   ASSERT(dstFormat == MESA_FORMAT_Z32_FLOAT_S8X24_UINT);
-   ASSERT(srcFormat == GL_DEPTH_STENCIL ||
+   assert(dstFormat == MESA_FORMAT_Z32_FLOAT_S8X24_UINT);
+   assert(srcFormat == GL_DEPTH_STENCIL ||
           srcFormat == GL_DEPTH_COMPONENT ||
           srcFormat == GL_STENCIL_INDEX);
-   ASSERT(srcFormat != GL_DEPTH_STENCIL ||
+   assert(srcFormat != GL_DEPTH_STENCIL ||
           srcType == GL_UNSIGNED_INT_24_8 ||
           srcType == GL_FLOAT_32_UNSIGNED_INT_24_8_REV);
 
@@ -606,7 +606,7 @@ texstore_depth_stencil(TEXSTORE_PARAMS)
       initialized = GL_TRUE;
    }
 
-   ASSERT(table[dstFormat]);
+   assert(table[dstFormat]);
    return table[dstFormat](ctx, dims, baseInternalFormat,
                            dstFormat, dstRowStride, dstSlices,
                            srcWidth, srcHeight, srcDepth,
@@ -666,7 +666,7 @@ texstore_compressed(TEXSTORE_PARAMS)
       initialized = GL_TRUE;
    }
 
-   ASSERT(table[dstFormat]);
+   assert(table[dstFormat]);
    return table[dstFormat](ctx, dims, baseInternalFormat,
                            dstFormat, dstRowStride, dstSlices,
                            srcWidth, srcHeight, srcDepth,
@@ -1225,10 +1225,10 @@ _mesa_store_compressed_teximage(struct gl_context *ctx, GLuint dims,
     * have to worry about the usual image unpacking or image transfer
     * operations.
     */
-   ASSERT(texImage);
-   ASSERT(texImage->Width > 0);
-   ASSERT(texImage->Height > 0);
-   ASSERT(texImage->Depth > 0);
+   assert(texImage);
+   assert(texImage->Width > 0);
+   assert(texImage->Height > 0);
+   assert(texImage->Depth > 0);
 
    /* allocate storage for texture data */
    if (!ctx->Driver.AllocTextureImageBuffer(ctx, texImage)) {
diff --git a/mesalib/src/mesa/main/transformfeedback.c b/mesalib/src/mesa/main/transformfeedback.c
index a73746335..a3e23ced5 100644
--- a/mesalib/src/mesa/main/transformfeedback.c
+++ b/mesalib/src/mesa/main/transformfeedback.c
@@ -93,7 +93,7 @@ reference_transform_feedback_object(struct gl_transform_feedback_object **ptr,
       /* Unreference the old object */
       struct gl_transform_feedback_object *oldObj = *ptr;
 
-      ASSERT(oldObj->RefCount > 0);
+      assert(oldObj->RefCount > 0);
       oldObj->RefCount--;
 
       if (oldObj->RefCount == 0) {
@@ -104,7 +104,7 @@ reference_transform_feedback_object(struct gl_transform_feedback_object **ptr,
 
       *ptr = NULL;
    }
-   ASSERT(!*ptr);
+   assert(!*ptr);
 
    if (obj) {
       /* reference new object */
@@ -143,7 +143,7 @@ void
 _mesa_init_transform_feedback(struct gl_context *ctx)
 {
    /* core mesa expects this, even a dummy one, to be available */
-   ASSERT(ctx->Driver.NewTransformFeedback);
+   assert(ctx->Driver.NewTransformFeedback);
 
    ctx->TransformFeedback.DefaultObject =
       ctx->Driver.NewTransformFeedback(ctx, 0);
@@ -185,7 +185,7 @@ void
 _mesa_free_transform_feedback(struct gl_context *ctx)
 {
    /* core mesa expects this, even a dummy one, to be available */
-   ASSERT(ctx->Driver.NewTransformFeedback);
+   assert(ctx->Driver.NewTransformFeedback);
 
    _mesa_reference_buffer_object(ctx,
                                  &ctx->TransformFeedback.CurrentBuffer,
@@ -235,7 +235,7 @@ delete_transform_feedback(struct gl_context *ctx,
 {
    GLuint i;
 
-   for (i = 0; i < Elements(obj->Buffers); i++) {
+   for (i = 0; i < ARRAY_SIZE(obj->Buffers); i++) {
       _mesa_reference_buffer_object(ctx, &obj->Buffers[i], NULL);
    }
 
diff --git a/mesalib/src/mesa/main/uniform_query.cpp b/mesalib/src/mesa/main/uniform_query.cpp
index 40327fba4..9f82de952 100644
--- a/mesalib/src/mesa/main/uniform_query.cpp
+++ b/mesalib/src/mesa/main/uniform_query.cpp
@@ -762,7 +762,7 @@ _mesa_uniform(struct gl_context *ctx, struct gl_shader_program *shProg,
 	  * been modified.
 	  */
 	 bool changed = false;
-	 for (unsigned j = 0; j < Elements(prog->SamplerUnits); j++) {
+	 for (unsigned j = 0; j < ARRAY_SIZE(prog->SamplerUnits); j++) {
 	    if ((sh->active_samplers & (1U << j)) != 0
 		&& (prog->SamplerUnits[j] != sh->SamplerUnits[j])) {
 	       changed = true;
diff --git a/mesalib/src/mesa/main/uniforms.c b/mesalib/src/mesa/main/uniforms.c
index e471b878c..fb1482f9c 100644
--- a/mesalib/src/mesa/main/uniforms.c
+++ b/mesalib/src/mesa/main/uniforms.c
@@ -81,7 +81,7 @@ _mesa_update_shader_textures_used(struct gl_shader_program *shProg,
       if (prog->SamplersUsed & (1 << s)) {
          GLuint unit = shader->SamplerUnits[s];
          GLuint tgt = shader->SamplerTargets[s];
-         assert(unit < Elements(prog->TexturesUsed));
+         assert(unit < ARRAY_SIZE(prog->TexturesUsed));
          assert(tgt < NUM_TEXTURE_TARGETS);
 
          /* The types of the samplers associated with a particular texture
@@ -1142,6 +1142,8 @@ _mesa_GetActiveUniformBlockiv(GLuint program,
       return;
 
    case GL_UNIFORM_BLOCK_REFERENCED_BY_GEOMETRY_SHADER:
+      if (!_mesa_has_geometry_shaders(ctx))
+         break;
       params[0] = shProg->UniformBlockStageIndex[MESA_SHADER_GEOMETRY][uniformBlockIndex] != -1;
       return;
 
diff --git a/mesalib/src/mesa/main/varray.c b/mesalib/src/mesa/main/varray.c
index 978ec7b53..42e7f89b2 100644
--- a/mesalib/src/mesa/main/varray.c
+++ b/mesalib/src/mesa/main/varray.c
@@ -24,6 +24,7 @@
  */
 
 
+#include <stdio.h>
 #include <inttypes.h>  /* for PRId64 macro */
 
 #include "glheader.h"
@@ -356,7 +357,7 @@ update_array_format(struct gl_context *ctx,
       return false;
    }
 
-   ASSERT(size <= 4);
+   assert(size <= 4);
 
    elementSize = _mesa_bytes_per_vertex_attrib(size, type);
    assert(elementSize != -1);
@@ -714,7 +715,7 @@ _mesa_EnableVertexAttribArray(GLuint index)
 
    vao = ctx->Array.VAO;
 
-   ASSERT(VERT_ATTRIB_GENERIC(index) < Elements(vao->VertexAttrib));
+   assert(VERT_ATTRIB_GENERIC(index) < ARRAY_SIZE(vao->VertexAttrib));
 
    if (!vao->VertexAttrib[VERT_ATTRIB_GENERIC(index)].Enabled) {
       /* was disabled, now being enabled */
@@ -740,7 +741,7 @@ _mesa_DisableVertexAttribArray(GLuint index)
 
    vao = ctx->Array.VAO;
 
-   ASSERT(VERT_ATTRIB_GENERIC(index) < Elements(vao->VertexAttrib));
+   assert(VERT_ATTRIB_GENERIC(index) < ARRAY_SIZE(vao->VertexAttrib));
 
    if (vao->VertexAttrib[VERT_ATTRIB_GENERIC(index)].Enabled) {
       /* was enabled, now being disabled */
@@ -769,7 +770,7 @@ get_vertex_array_attrib(struct gl_context *ctx, GLuint index, GLenum pname,
       return 0;
    }
 
-   ASSERT(VERT_ATTRIB_GENERIC(index) < Elements(vao->VertexAttrib));
+   assert(VERT_ATTRIB_GENERIC(index) < ARRAY_SIZE(vao->VertexAttrib));
 
    array = &vao->VertexAttrib[VERT_ATTRIB_GENERIC(index)];
 
@@ -834,7 +835,7 @@ get_current_attrib(struct gl_context *ctx, GLuint index, const char *function)
       return NULL;
    }
 
-   ASSERT(VERT_ATTRIB_GENERIC(index) < Elements(ctx->Array.VAO->VertexAttrib));
+   assert(VERT_ATTRIB_GENERIC(index) < ARRAY_SIZE(ctx->Array.VAO->VertexAttrib));
 
    FLUSH_CURRENT(ctx, 0);
    return ctx->Current.Attrib[VERT_ATTRIB_GENERIC(index)];
@@ -956,7 +957,7 @@ _mesa_GetVertexAttribPointerv(GLuint index, GLenum pname, GLvoid **pointer)
       return;
    }
 
-   ASSERT(VERT_ATTRIB_GENERIC(index) < Elements(ctx->Array.VAO->VertexAttrib));
+   assert(VERT_ATTRIB_GENERIC(index) < ARRAY_SIZE(ctx->Array.VAO->VertexAttrib));
 
    *pointer = (GLvoid *) ctx->Array.VAO->VertexAttrib[VERT_ATTRIB_GENERIC(index)].Ptr;
 }
@@ -1348,7 +1349,7 @@ _mesa_VertexAttribDivisor(GLuint index, GLuint divisor)
       return;
    }
 
-   ASSERT(genericIndex < Elements(ctx->Array.VAO->VertexAttrib));
+   assert(genericIndex < ARRAY_SIZE(ctx->Array.VAO->VertexAttrib));
 
    /* The ARB_vertex_attrib_binding spec says:
     *
@@ -1802,8 +1803,8 @@ _mesa_VertexAttribBinding(GLuint attribIndex, GLuint bindingIndex)
       return;
    }
 
-   ASSERT(VERT_ATTRIB_GENERIC(attribIndex) <
-          Elements(ctx->Array.VAO->VertexAttrib));
+   assert(VERT_ATTRIB_GENERIC(attribIndex) <
+          ARRAY_SIZE(ctx->Array.VAO->VertexAttrib));
 
    vertex_attrib_binding(ctx, VERT_ATTRIB_GENERIC(attribIndex),
                          VERT_ATTRIB_GENERIC(bindingIndex));
diff --git a/mesalib/src/mesa/main/version.c b/mesalib/src/mesa/main/version.c
index 4951891cb..8e0c3ef5b 100644
--- a/mesalib/src/mesa/main/version.c
+++ b/mesalib/src/mesa/main/version.c
@@ -23,6 +23,7 @@
  */
 
 
+#include <stdio.h>
 #include "imports.h"
 #include "mtypes.h"
 #include "version.h"
diff --git a/mesalib/src/mesa/main/vtxfmt.h b/mesalib/src/mesa/main/vtxfmt.h
index 34ade8905..4f8bc9cdf 100644
--- a/mesalib/src/mesa/main/vtxfmt.h
+++ b/mesalib/src/mesa/main/vtxfmt.h
@@ -33,11 +33,18 @@
 #ifndef _VTXFMT_H_
 #define _VTXFMT_H_
 
-#include "compiler.h"
 #include "mtypes.h"
 
+#ifdef __cplusplus
+extern "C" {
+#endif
+
 extern void _mesa_install_exec_vtxfmt( struct gl_context *ctx, const GLvertexformat *vfmt );
 extern void _mesa_install_save_vtxfmt( struct gl_context *ctx, const GLvertexformat *vfmt );
 extern void _mesa_initialize_vbo_vtxfmt(struct gl_context *ctx);
 
+#ifdef __cplusplus
+} // extern "C"
+#endif
+
 #endif /* _VTXFMT_H_ */
diff --git a/mesalib/src/mesa/math/m_debug_norm.c b/mesalib/src/mesa/math/m_debug_norm.c
index 00e72be54..197b43cf2 100644
--- a/mesalib/src/mesa/math/m_debug_norm.c
+++ b/mesalib/src/mesa/math/m_debug_norm.c
@@ -26,6 +26,7 @@
  *    Gareth Hughes
  */
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/macros.h"
@@ -165,7 +166,7 @@ static void ref_norm_transform_normalize( const GLmatrix *mat,
 	    /* Hmmm, don't know how we could test the precalculated
 	     * length case...
 	     */
-            scale = INV_SQRTF( len );
+            scale = 1.0f / sqrtf(len);
 	    SCALE_SCALAR_3V( out[i], scale, t );
          } else {
             out[i][0] = out[i][1] = out[i][2] = 0;
@@ -241,7 +242,7 @@ static int test_norm_function( normal_func func, int mtype, long *cycles )
       ASSIGN_3V( d2[i], 0.0, 0.0, 0.0 );
       for ( j = 0 ; j < 3 ; j++ )
          s[i][j] = rnd();
-      length[i] = INV_SQRTF( LEN_SQUARED_3FV( s[i] ) );
+      length[i] = 1.0f / sqrtf( LEN_SQUARED_3FV( s[i] ) );
    }
 
    source->data = (GLfloat(*)[4]) s;
diff --git a/mesalib/src/mesa/math/m_debug_util.h b/mesalib/src/mesa/math/m_debug_util.h
index d05da8994..25ee029ce 100644
--- a/mesalib/src/mesa/math/m_debug_util.h
+++ b/mesalib/src/mesa/math/m_debug_util.h
@@ -32,6 +32,9 @@
 #ifdef DEBUG_MATH  /* This code only used for debugging */
 
 
+#include "c99_math.h"
+
+
 /* Comment this out to deactivate the cycle counter.
  * NOTE: it works only on CPUs which know the 'rdtsc' command (586 or higher)
  * (hope, you don't try to debug Mesa on a 386 ;)
@@ -286,9 +289,9 @@ static int significand_match( GLfloat a, GLfloat b )
       return 0;
    }
 
-   FREXPF( a, &a_ex );
-   FREXPF( b, &b_ex );
-   FREXPF( d, &d_ex );
+   frexpf( a, &a_ex );
+   frexpf( b, &b_ex );
+   frexpf( d, &d_ex );
 
    if ( a_ex < b_ex ) {
       return a_ex - d_ex;
diff --git a/mesalib/src/mesa/math/m_debug_xform.c b/mesalib/src/mesa/math/m_debug_xform.c
index f56b4cfc7..632c82ea2 100644
--- a/mesalib/src/mesa/math/m_debug_xform.c
+++ b/mesalib/src/mesa/math/m_debug_xform.c
@@ -187,7 +187,7 @@ static int test_transform_function( transform_func func, int psize,
    mat->type = mtypes[mtype];
 
    m = mat->m;
-   ASSERT( ((long)m & 15) == 0 );
+   assert( ((long)m & 15) == 0 );
 
    init_matrix( m );
 
@@ -206,7 +206,7 @@ static int test_transform_function( transform_func func, int psize,
          case VAR:
             break;
          default:
-            ASSERT(0);
+            assert(0);
             return 0;
          }
       }
diff --git a/mesalib/src/mesa/math/m_matrix.c b/mesalib/src/mesa/math/m_matrix.c
index 9c9310d93..0475a7ab2 100644
--- a/mesalib/src/mesa/math/m_matrix.c
+++ b/mesalib/src/mesa/math/m_matrix.c
@@ -34,6 +34,7 @@
  */
 
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/imports.h"
 #include "main/macros.h"
@@ -376,9 +377,9 @@ static GLboolean invert_matrix_general( GLmatrix *mat )
    r3[7] = 1.0, r3[4] = r3[5] = r3[6] = 0.0;
 
    /* choose pivot - or die */
-   if (FABSF(r3[0])>FABSF(r2[0])) SWAP_ROWS(r3, r2);
-   if (FABSF(r2[0])>FABSF(r1[0])) SWAP_ROWS(r2, r1);
-   if (FABSF(r1[0])>FABSF(r0[0])) SWAP_ROWS(r1, r0);
+   if (fabsf(r3[0])>fabsf(r2[0])) SWAP_ROWS(r3, r2);
+   if (fabsf(r2[0])>fabsf(r1[0])) SWAP_ROWS(r2, r1);
+   if (fabsf(r1[0])>fabsf(r0[0])) SWAP_ROWS(r1, r0);
    if (0.0 == r0[0])  return GL_FALSE;
 
    /* eliminate first variable     */
@@ -396,8 +397,8 @@ static GLboolean invert_matrix_general( GLmatrix *mat )
    if (s != 0.0) { r1[7] -= m1 * s; r2[7] -= m2 * s; r3[7] -= m3 * s; }
 
    /* choose pivot - or die */
-   if (FABSF(r3[1])>FABSF(r2[1])) SWAP_ROWS(r3, r2);
-   if (FABSF(r2[1])>FABSF(r1[1])) SWAP_ROWS(r2, r1);
+   if (fabsf(r3[1])>fabsf(r2[1])) SWAP_ROWS(r3, r2);
+   if (fabsf(r2[1])>fabsf(r1[1])) SWAP_ROWS(r2, r1);
    if (0.0 == r1[1])  return GL_FALSE;
 
    /* eliminate second variable */
@@ -410,7 +411,7 @@ static GLboolean invert_matrix_general( GLmatrix *mat )
    s = r1[7]; if (0.0 != s) { r2[7] -= m2 * s; r3[7] -= m3 * s; }
 
    /* choose pivot - or die */
-   if (FABSF(r3[2])>FABSF(r2[2])) SWAP_ROWS(r3, r2);
+   if (fabsf(r3[2])>fabsf(r2[2])) SWAP_ROWS(r3, r2);
    if (0.0 == r2[2])  return GL_FALSE;
 
    /* eliminate third variable */
@@ -508,7 +509,7 @@ static GLboolean invert_matrix_3d_general( GLmatrix *mat )
 
    det = pos + neg;
 
-   if (FABSF(det) < 1e-25)
+   if (fabsf(det) < 1e-25)
       return GL_FALSE;
 
    det = 1.0F / det;
@@ -798,8 +799,8 @@ _math_matrix_rotate( GLmatrix *mat,
    GLfloat m[16];
    GLboolean optimized;
 
-   s = (GLfloat) sin( angle * DEG2RAD );
-   c = (GLfloat) cos( angle * DEG2RAD );
+   s = (GLfloat) sin( angle * M_PI / 180.0 );
+   c = (GLfloat) cos( angle * M_PI / 180.0 );
 
    memcpy(m, Identity, sizeof(GLfloat)*16);
    optimized = GL_FALSE;
@@ -1069,7 +1070,7 @@ _math_matrix_scale( GLmatrix *mat, GLfloat x, GLfloat y, GLfloat z )
    m[2] *= x;   m[6] *= y;   m[10] *= z;
    m[3] *= x;   m[7] *= y;   m[11] *= z;
 
-   if (FABSF(x - y) < 1e-8 && FABSF(x - z) < 1e-8)
+   if (fabsf(x - y) < 1e-8 && fabsf(x - z) < 1e-8)
       mat->flags |= MAT_FLAG_UNIFORM_SCALE;
    else
       mat->flags |= MAT_FLAG_GENERAL_SCALE;
diff --git a/mesalib/src/mesa/math/m_norm_tmp.h b/mesalib/src/mesa/math/m_norm_tmp.h
index 339c03ff8..c8fab0ed3 100644
--- a/mesalib/src/mesa/math/m_norm_tmp.h
+++ b/mesalib/src/mesa/math/m_norm_tmp.h
@@ -68,7 +68,7 @@ TAG(transform_normalize_normals)( const GLmatrix *mat,
 	 {
 	    GLdouble len = tx*tx + ty*ty + tz*tz;
 	    if (len > 1e-20) {
-	       GLfloat scale = INV_SQRTF(len);
+	       GLfloat scale = 1.0f / sqrtf(len);
 	       out[i][0] = tx * scale;
 	       out[i][1] = ty * scale;
 	       out[i][2] = tz * scale;
@@ -135,7 +135,7 @@ TAG(transform_normalize_normals_no_rot)( const GLmatrix *mat,
 	 {
 	    GLdouble len = tx*tx + ty*ty + tz*tz;
 	    if (len > 1e-20) {
-	       GLfloat scale = INV_SQRTF(len);
+	       GLfloat scale = 1.0f / sqrtf(len);
 	       out[i][0] = tx * scale;
 	       out[i][1] = ty * scale;
 	       out[i][2] = tz * scale;
@@ -322,7 +322,7 @@ TAG(normalize_normals)( const GLmatrix *mat,
 	 const GLfloat x = from[0], y = from[1], z = from[2];
 	 GLdouble len = x * x + y * y + z * z;
 	 if (len > 1e-50) {
-	    len = INV_SQRTF(len);
+	    len = 1.0f / sqrtf(len);
 	    out[i][0] = (GLfloat)(x * len);
 	    out[i][1] = (GLfloat)(y * len);
 	    out[i][2] = (GLfloat)(z * len);
diff --git a/mesalib/src/mesa/math/m_trans_tmp.h b/mesalib/src/mesa/math/m_trans_tmp.h
index f39a3c83a..0a8bfc538 100644
--- a/mesalib/src/mesa/math/m_trans_tmp.h
+++ b/mesalib/src/mesa/math/m_trans_tmp.h
@@ -40,13 +40,11 @@ static void DEST_4F( GLfloat (*t)[4],
    (void) first;
    (void) start;
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	 if (SZ >= 1) t[i][0] = TRX_4F(f, 0);
-	 if (SZ >= 2) t[i][1] = TRX_4F(f, 1);
-	 if (SZ >= 3) t[i][2] = TRX_4F(f, 2);
-	 if (SZ == 4) t[i][3] = TRX_4F(f, 3); else t[i][3] = 1.0;
-      }
+      NEXT_F2;
+      if (SZ >= 1) t[i][0] = TRX_4F(f, 0);
+      if (SZ >= 2) t[i][1] = TRX_4F(f, 1);
+      if (SZ >= 3) t[i][2] = TRX_4F(f, 2);
+      if (SZ == 4) t[i][3] = TRX_4F(f, 3); else t[i][3] = 1.0;
    }
 }
 #endif
@@ -66,13 +64,11 @@ static void DEST_4FN( GLfloat (*t)[4],
    (void) first;
    (void) start;
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	 if (SZ >= 1) t[i][0] = TRX_4FN(f, 0);
-	 if (SZ >= 2) t[i][1] = TRX_4FN(f, 1);
-	 if (SZ >= 3) t[i][2] = TRX_4FN(f, 2);
-	 if (SZ == 4) t[i][3] = TRX_4FN(f, 3); else t[i][3] = 1.0;
-      }
+      NEXT_F2;
+      if (SZ >= 1) t[i][0] = TRX_4FN(f, 0);
+      if (SZ >= 2) t[i][1] = TRX_4FN(f, 1);
+      if (SZ >= 3) t[i][2] = TRX_4FN(f, 2);
+      if (SZ == 4) t[i][3] = TRX_4FN(f, 3); else t[i][3] = 1.0;
    }
 }
 #endif
@@ -90,12 +86,10 @@ static void DEST_3FN( GLfloat (*t)[3],
    (void) first;
    (void) start;
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	 t[i][0] = TRX_3FN(f, 0);
-	 t[i][1] = TRX_3FN(f, 1);
-	 t[i][2] = TRX_3FN(f, 2);
-      }
+      NEXT_F2;
+      t[i][0] = TRX_3FN(f, 0);
+      t[i][1] = TRX_3FN(f, 1);
+      t[i][2] = TRX_3FN(f, 2);
    }
 }
 #endif
@@ -112,10 +106,8 @@ static void DEST_1F( GLfloat *t,
    (void) first;
    (void) start;
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	 t[i] = TRX_1F(f, 0);
-      }
+      NEXT_F2;
+      t[i] = TRX_1F(f, 0);
    }
 }
 #endif
@@ -132,13 +124,11 @@ static void DEST_4UB( GLubyte (*t)[4],
    (void) start;
    (void) first;
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	 if (SZ >= 1) TRX_UB(t[i][0], f, 0);
-	 if (SZ >= 2) TRX_UB(t[i][1], f, 1);
-	 if (SZ >= 3) TRX_UB(t[i][2], f, 2);
-	 if (SZ == 4) TRX_UB(t[i][3], f, 3); else t[i][3] = 255;
-      }
+      NEXT_F2;
+      if (SZ >= 1) TRX_UB(t[i][0], f, 0);
+      if (SZ >= 2) TRX_UB(t[i][1], f, 1);
+      if (SZ >= 3) TRX_UB(t[i][2], f, 2);
+      if (SZ == 4) TRX_UB(t[i][3], f, 3); else t[i][3] = 255;
    }
 }
 #endif
@@ -156,13 +146,11 @@ static void DEST_4US( GLushort (*t)[4],
    (void) start;
    (void) first;
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	 if (SZ >= 1) TRX_US(t[i][0], f, 0);
-	 if (SZ >= 2) TRX_US(t[i][1], f, 1);
-	 if (SZ >= 3) TRX_US(t[i][2], f, 2);
-	 if (SZ == 4) TRX_US(t[i][3], f, 3); else t[i][3] = 65535;
-      }
+      NEXT_F2;
+      if (SZ >= 1) TRX_US(t[i][0], f, 0);
+      if (SZ >= 2) TRX_US(t[i][1], f, 1);
+      if (SZ >= 3) TRX_US(t[i][2], f, 2);
+      if (SZ == 4) TRX_US(t[i][3], f, 3); else t[i][3] = 65535;
    }
 }
 #endif
@@ -180,10 +168,8 @@ static void DEST_1UB( GLubyte *t,
    (void) start;
    (void) first;
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	  TRX_UB(t[i], f, 0);
-      }
+      NEXT_F2;
+      TRX_UB(t[i], f, 0);
    }
 }
 #endif
@@ -202,10 +188,8 @@ static void DEST_1UI( GLuint *t,
    (void) first;
 
    for (i = DST_START ; i < n ; i++, NEXT_F) {
-      CHECK {
-         NEXT_F2;
-	 t[i] = TRX_UI(f, 0);
-      }
+      NEXT_F2;
+      t[i] = TRX_UI(f, 0);
    }
 }
 #endif
@@ -214,19 +198,19 @@ static void DEST_1UI( GLuint *t,
 static void INIT(void)
 {
 #ifdef DEST_1UI
-   ASSERT(SZ == 1);
+   assert(SZ == 1);
    TAB(_1ui)[SRC_IDX] = DEST_1UI;
 #endif
 #ifdef DEST_1UB
-   ASSERT(SZ == 1);
+   assert(SZ == 1);
    TAB(_1ub)[SRC_IDX] = DEST_1UB;
 #endif
 #ifdef DEST_1F
-   ASSERT(SZ == 1);
+   assert(SZ == 1);
    TAB(_1f)[SRC_IDX] = DEST_1F;
 #endif
 #ifdef DEST_3FN
-   ASSERT(SZ == 3);
+   assert(SZ == 3);
    TAB(_3fn)[SRC_IDX] = DEST_3FN;
 #endif
 #ifdef DEST_4UB
diff --git a/mesalib/src/mesa/math/m_translate.c b/mesalib/src/mesa/math/m_translate.c
index 3a8ca74f6..a5bf5d76f 100644
--- a/mesalib/src/mesa/math/m_translate.c
+++ b/mesalib/src/mesa/math/m_translate.c
@@ -84,10 +84,6 @@ typedef void (*trans_3fn_func)(GLfloat (*to)[3],
 #define MAX_TYPES TYPE_IDX(GL_DOUBLE)+1      /* 0xa + 1 */
 
 
-/* This macro is used on other systems, so undefine it for this module */
-
-#undef	CHECK
-
 static trans_1f_func  _math_trans_1f_tab[MAX_TYPES];
 static trans_1ui_func _math_trans_1ui_tab[MAX_TYPES];
 static trans_1ub_func _math_trans_1ub_tab[MAX_TYPES];
@@ -108,7 +104,6 @@ static trans_4f_func  _math_trans_4fn_tab[5][MAX_TYPES];
 #define STRIDE stride
 #define NEXT_F f += stride
 #define NEXT_F2
-#define CHECK
 
 
 
@@ -604,7 +599,6 @@ static void init_translate_raw(void)
 #undef CLASS
 #endif
 #undef ARGS
-#undef CHECK
 #undef SRC_START
 #undef DST_START
 #undef NEXT_F
diff --git a/mesalib/src/mesa/math/m_translate.h b/mesalib/src/mesa/math/m_translate.h
index bdfa4c770..62f676ca6 100644
--- a/mesalib/src/mesa/math/m_translate.h
+++ b/mesalib/src/mesa/math/m_translate.h
@@ -26,7 +26,6 @@
 #ifndef _M_TRANSLATE_H_
 #define _M_TRANSLATE_H_
 
-#include "main/compiler.h"
 #include "main/glheader.h"
 
 /**
diff --git a/mesalib/src/mesa/math/m_vector.c b/mesalib/src/mesa/math/m_vector.c
index ff42396b7..831f953d6 100644
--- a/mesalib/src/mesa/math/m_vector.c
+++ b/mesalib/src/mesa/math/m_vector.c
@@ -26,6 +26,7 @@
  * New (3.1) transformation code written by Keith Whitwell.
  */
 
+#include <stdio.h>
 
 #include "main/glheader.h"
 #include "main/imports.h"
diff --git a/mesalib/src/mesa/math/m_xform.c b/mesalib/src/mesa/math/m_xform.c
index 14d1c6453..718ad4993 100644
--- a/mesalib/src/mesa/math/m_xform.c
+++ b/mesalib/src/mesa/math/m_xform.c
@@ -33,6 +33,7 @@
  * 3. Transformation of a point p by a matrix M is: p' = M * p
  */
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/macros.h"
 
diff --git a/mesalib/src/mesa/math/m_xform.h b/mesalib/src/mesa/math/m_xform.h
index e45880322..2ed62e711 100644
--- a/mesalib/src/mesa/math/m_xform.h
+++ b/mesalib/src/mesa/math/m_xform.h
@@ -119,7 +119,7 @@ typedef void (*vec_copy_func)( GLvector4f *to,
 /*
  * Functions for transformation of normals in the VB.
  */
-typedef void (_NORMAPIP normal_func)( const GLmatrix *mat,
+typedef void (_XFORMAPIP normal_func)( const GLmatrix *mat,
 				      GLfloat scale,
 				      const GLvector4f *in,
 				      const GLfloat lengths[],
diff --git a/mesalib/src/mesa/program/arbprogparse.c b/mesalib/src/mesa/program/arbprogparse.c
index 53a6f37cb..3ddaeca8a 100644
--- a/mesalib/src/mesa/program/arbprogparse.c
+++ b/mesalib/src/mesa/program/arbprogparse.c
@@ -73,7 +73,7 @@ _mesa_parse_arb_fragment_program(struct gl_context* ctx, GLenum target,
    struct asm_parser_state state;
    GLuint i;
 
-   ASSERT(target == GL_FRAGMENT_PROGRAM_ARB);
+   assert(target == GL_FRAGMENT_PROGRAM_ARB);
 
    memset(&prog, 0, sizeof(prog));
    memset(&state, 0, sizeof(state));
@@ -166,7 +166,7 @@ _mesa_parse_arb_vertex_program(struct gl_context *ctx, GLenum target,
    struct gl_program prog;
    struct asm_parser_state state;
 
-   ASSERT(target == GL_VERTEX_PROGRAM_ARB);
+   assert(target == GL_VERTEX_PROGRAM_ARB);
 
    memset(&prog, 0, sizeof(prog));
    memset(&state, 0, sizeof(state));
diff --git a/mesalib/src/mesa/program/ir_to_mesa.cpp b/mesalib/src/mesa/program/ir_to_mesa.cpp
index b2776da45..39790ec8e 100644
--- a/mesalib/src/mesa/program/ir_to_mesa.cpp
+++ b/mesalib/src/mesa/program/ir_to_mesa.cpp
@@ -1018,7 +1018,7 @@ void
 ir_to_mesa_visitor::visit(ir_expression *ir)
 {
    unsigned int operand;
-   src_reg op[Elements(ir->operands)];
+   src_reg op[ARRAY_SIZE(ir->operands)];
    src_reg result_src;
    dst_reg result_dst;
 
diff --git a/mesalib/src/mesa/program/prog_execute.c b/mesalib/src/mesa/program/prog_execute.c
index b2fbc808a..dc4919ae8 100644
--- a/mesalib/src/mesa/program/prog_execute.c
+++ b/mesalib/src/mesa/program/prog_execute.c
@@ -35,6 +35,7 @@
  */
 
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/colormac.h"
 #include "main/macros.h"
@@ -123,7 +124,7 @@ get_src_register_pointer(const struct prog_src_register *source,
       return (GLfloat *) prog->Parameters->ParameterValues[reg];
 
    case PROGRAM_SYSTEM_VALUE:
-      assert(reg < (GLint) Elements(machine->SystemValues));
+      assert(reg < (GLint) ARRAY_SIZE(machine->SystemValues));
       return machine->SystemValues[reg];
 
    default:
@@ -190,10 +191,10 @@ fetch_vector4(const struct prog_src_register *source,
       COPY_4V(result, src);
    }
    else {
-      ASSERT(GET_SWZ(source->Swizzle, 0) <= 3);
-      ASSERT(GET_SWZ(source->Swizzle, 1) <= 3);
-      ASSERT(GET_SWZ(source->Swizzle, 2) <= 3);
-      ASSERT(GET_SWZ(source->Swizzle, 3) <= 3);
+      assert(GET_SWZ(source->Swizzle, 0) <= 3);
+      assert(GET_SWZ(source->Swizzle, 1) <= 3);
+      assert(GET_SWZ(source->Swizzle, 2) <= 3);
+      assert(GET_SWZ(source->Swizzle, 3) <= 3);
       result[0] = src[GET_SWZ(source->Swizzle, 0)];
       result[1] = src[GET_SWZ(source->Swizzle, 1)];
       result[2] = src[GET_SWZ(source->Swizzle, 2)];
@@ -201,13 +202,13 @@ fetch_vector4(const struct prog_src_register *source,
    }
 
    if (source->Abs) {
-      result[0] = FABSF(result[0]);
-      result[1] = FABSF(result[1]);
-      result[2] = FABSF(result[2]);
-      result[3] = FABSF(result[3]);
+      result[0] = fabsf(result[0]);
+      result[1] = fabsf(result[1]);
+      result[2] = fabsf(result[2]);
+      result[3] = fabsf(result[3]);
    }
    if (source->Negate) {
-      ASSERT(source->Negate == NEGATE_XYZW);
+      assert(source->Negate == NEGATE_XYZW);
       result[0] = -result[0];
       result[1] = -result[1];
       result[2] = -result[2];
@@ -259,13 +260,13 @@ fetch_vector4_deriv(struct gl_context * ctx,
       result[3] = deriv[GET_SWZ(source->Swizzle, 3)];
       
       if (source->Abs) {
-         result[0] = FABSF(result[0]);
-         result[1] = FABSF(result[1]);
-         result[2] = FABSF(result[2]);
-         result[3] = FABSF(result[3]);
+         result[0] = fabsf(result[0]);
+         result[1] = fabsf(result[1]);
+         result[2] = fabsf(result[2]);
+         result[3] = fabsf(result[3]);
       }
       if (source->Negate) {
-         ASSERT(source->Negate == NEGATE_XYZW);
+         assert(source->Negate == NEGATE_XYZW);
          result[0] = -result[0];
          result[1] = -result[1];
          result[2] = -result[2];
@@ -290,7 +291,7 @@ fetch_vector1(const struct prog_src_register *source,
    result[0] = src[GET_SWZ(source->Swizzle, 0)];
 
    if (source->Abs) {
-      result[0] = FABSF(result[0]);
+      result[0] = fabsf(result[0]);
    }
    if (source->Negate) {
       result[0] = -result[0];
@@ -520,10 +521,10 @@ _mesa_execute_program(struct gl_context * ctx,
          {
             GLfloat a[4], result[4];
             fetch_vector4(&inst->SrcReg[0], machine, a);
-            result[0] = FABSF(a[0]);
-            result[1] = FABSF(a[1]);
-            result[2] = FABSF(a[2]);
-            result[3] = FABSF(a[3]);
+            result[0] = fabsf(a[0]);
+            result[1] = fabsf(a[1]);
+            result[2] = fabsf(a[2]);
+            result[3] = fabsf(a[3]);
             store_vector4(inst, machine, result);
          }
          break;
@@ -556,12 +557,12 @@ _mesa_execute_program(struct gl_context * ctx,
          break;
       case OPCODE_BGNLOOP:
          /* no-op */
-         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+         assert(program->Instructions[inst->BranchTarget].Opcode
                 == OPCODE_ENDLOOP);
          break;
       case OPCODE_ENDLOOP:
          /* subtract 1 here since pc is incremented by for(pc) loop */
-         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+         assert(program->Instructions[inst->BranchTarget].Opcode
                 == OPCODE_BGNLOOP);
          pc = inst->BranchTarget - 1;   /* go to matching BNGLOOP */
          break;
@@ -570,7 +571,7 @@ _mesa_execute_program(struct gl_context * ctx,
       case OPCODE_ENDSUB:      /* end subroutine */
          break;
       case OPCODE_BRK:         /* break out of loop (conditional) */
-         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+         assert(program->Instructions[inst->BranchTarget].Opcode
                 == OPCODE_ENDLOOP);
          if (eval_condition(machine, inst)) {
             /* break out of loop */
@@ -579,7 +580,7 @@ _mesa_execute_program(struct gl_context * ctx,
          }
          break;
       case OPCODE_CONT:        /* continue loop (conditional) */
-         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+         assert(program->Instructions[inst->BranchTarget].Opcode
                 == OPCODE_ENDLOOP);
          if (eval_condition(machine, inst)) {
             /* continue at ENDLOOP */
@@ -708,7 +709,7 @@ _mesa_execute_program(struct gl_context * ctx,
          {
             GLfloat t[4], q[4], floor_t0;
             fetch_vector1(&inst->SrcReg[0], machine, t);
-            floor_t0 = FLOORF(t[0]);
+            floor_t0 = floorf(t[0]);
             if (floor_t0 > FLT_MAX_EXP) {
                SET_POS_INFINITY(q[0]);
                SET_POS_INFINITY(q[2]);
@@ -718,7 +719,7 @@ _mesa_execute_program(struct gl_context * ctx,
                q[2] = 0.0F;
             }
             else {
-               q[0] = LDEXPF(1.0, (int) floor_t0);
+               q[0] = ldexpf(1.0, (int) floor_t0);
                /* Note: GL_NV_vertex_program expects 
                 * result.z = result.x * APPX(result.y)
                 * We do what the ARB extension says.
@@ -747,10 +748,10 @@ _mesa_execute_program(struct gl_context * ctx,
          {
             GLfloat a[4], result[4];
             fetch_vector4(&inst->SrcReg[0], machine, a);
-            result[0] = FLOORF(a[0]);
-            result[1] = FLOORF(a[1]);
-            result[2] = FLOORF(a[2]);
-            result[3] = FLOORF(a[3]);
+            result[0] = floorf(a[0]);
+            result[1] = floorf(a[1]);
+            result[2] = floorf(a[2]);
+            result[3] = floorf(a[3]);
             store_vector4(inst, machine, result);
          }
          break;
@@ -758,17 +759,17 @@ _mesa_execute_program(struct gl_context * ctx,
          {
             GLfloat a[4], result[4];
             fetch_vector4(&inst->SrcReg[0], machine, a);
-            result[0] = a[0] - FLOORF(a[0]);
-            result[1] = a[1] - FLOORF(a[1]);
-            result[2] = a[2] - FLOORF(a[2]);
-            result[3] = a[3] - FLOORF(a[3]);
+            result[0] = a[0] - floorf(a[0]);
+            result[1] = a[1] - floorf(a[1]);
+            result[2] = a[2] - floorf(a[2]);
+            result[3] = a[3] - floorf(a[3]);
             store_vector4(inst, machine, result);
          }
          break;
       case OPCODE_IF:
          {
             GLboolean cond;
-            ASSERT(program->Instructions[inst->BranchTarget].Opcode
+            assert(program->Instructions[inst->BranchTarget].Opcode
                    == OPCODE_ELSE ||
                    program->Instructions[inst->BranchTarget].Opcode
                    == OPCODE_ENDIF);
@@ -797,7 +798,7 @@ _mesa_execute_program(struct gl_context * ctx,
          break;
       case OPCODE_ELSE:
          /* goto ENDIF */
-         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+         assert(program->Instructions[inst->BranchTarget].Opcode
                 == OPCODE_ENDIF);
          assert(inst->BranchTarget >= 0);
          pc = inst->BranchTarget;
@@ -874,7 +875,7 @@ _mesa_execute_program(struct gl_context * ctx,
          {
             GLfloat t[4], q[4], abs_t0;
             fetch_vector1(&inst->SrcReg[0], machine, t);
-            abs_t0 = FABSF(t[0]);
+            abs_t0 = fabsf(t[0]);
             if (abs_t0 != 0.0F) {
                if (IS_INF_OR_NAN(abs_t0))
                {
@@ -884,7 +885,7 @@ _mesa_execute_program(struct gl_context * ctx,
                }
                else {
                   int exponent;
-                  GLfloat mantissa = FREXPF(t[0], &exponent);
+                  GLfloat mantissa = frexpf(t[0], &exponent);
                   q[0] = (GLfloat) (exponent - 1);
                   q[1] = (GLfloat) (2.0 * mantissa); /* map [.5, 1) -> [1, 2) */
 
@@ -1083,8 +1084,8 @@ _mesa_execute_program(struct gl_context * ctx,
          {
             GLfloat a[4], result[4];
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            a[0] = FABSF(a[0]);
-            result[0] = result[1] = result[2] = result[3] = INV_SQRTF(a[0]);
+            a[0] = fabsf(a[0]);
+            result[0] = result[1] = result[2] = result[3] = 1.0f / sqrtf(a[0]);
             store_vector4(inst, machine, result);
             if (DEBUG_PROG) {
                printf("RSQ %g = 1/sqrt(|%g|)\n", result[0], a[0]);
@@ -1260,7 +1261,7 @@ _mesa_execute_program(struct gl_context * ctx,
                else if (swz == SWIZZLE_ONE)
                   result[i] = 1.0;
                else {
-                  ASSERT(swz <= 3);
+                  assert(swz <= 3);
                   result[i] = src[swz];
                }
                if (source->Negate & (1 << i))
@@ -1357,7 +1358,7 @@ _mesa_execute_program(struct gl_context * ctx,
 
             fetch_vector4(&inst->SrcReg[0], machine, texcoord);
             /* Not so sure about this test - if texcoord[3] is
-             * zero, we'd probably be fine except for an ASSERT in
+             * zero, we'd probably be fine except for an assert in
              * IROUND_POS() which gets triggered by the inf values created.
              */
             if (texcoord[3] != 0.0) {
diff --git a/mesalib/src/mesa/program/prog_instruction.c b/mesalib/src/mesa/program/prog_instruction.c
index 254c0128f..6a9bcb7b5 100644
--- a/mesalib/src/mesa/program/prog_instruction.c
+++ b/mesalib/src/mesa/program/prog_instruction.c
@@ -200,9 +200,9 @@ static const struct instruction_info InstInfo[MAX_OPCODE] = {
 GLuint
 _mesa_num_inst_src_regs(gl_inst_opcode opcode)
 {
-   ASSERT(opcode < MAX_OPCODE);
-   ASSERT(opcode == InstInfo[opcode].Opcode);
-   ASSERT(OPCODE_XPD == InstInfo[OPCODE_XPD].Opcode);
+   assert(opcode < MAX_OPCODE);
+   assert(opcode == InstInfo[opcode].Opcode);
+   assert(OPCODE_XPD == InstInfo[OPCODE_XPD].Opcode);
    return InstInfo[opcode].NumSrcRegs;
 }
 
@@ -213,9 +213,9 @@ _mesa_num_inst_src_regs(gl_inst_opcode opcode)
 GLuint
 _mesa_num_inst_dst_regs(gl_inst_opcode opcode)
 {
-   ASSERT(opcode < MAX_OPCODE);
-   ASSERT(opcode == InstInfo[opcode].Opcode);
-   ASSERT(OPCODE_XPD == InstInfo[OPCODE_XPD].Opcode);
+   assert(opcode < MAX_OPCODE);
+   assert(opcode == InstInfo[opcode].Opcode);
+   assert(OPCODE_XPD == InstInfo[OPCODE_XPD].Opcode);
    return InstInfo[opcode].NumDstRegs;
 }
 
diff --git a/mesalib/src/mesa/program/prog_optimize.c b/mesalib/src/mesa/program/prog_optimize.c
index 65d427cb4..6d4485acb 100644
--- a/mesalib/src/mesa/program/prog_optimize.c
+++ b/mesalib/src/mesa/program/prog_optimize.c
@@ -57,7 +57,7 @@ get_src_arg_mask(const struct prog_instruction *inst,
    GLuint read_mask, channel_mask;
    GLuint comp;
 
-   ASSERT(arg < _mesa_num_inst_src_regs(inst->Opcode));
+   assert(arg < _mesa_num_inst_src_regs(inst->Opcode));
 
    /* Form the dst register, find the written channels */
    if (inst->CondUpdate) {
@@ -133,7 +133,7 @@ get_dst_mask_for_mov(const struct prog_instruction *mov, GLuint src_mask)
    GLuint comp;
    GLuint updated_mask = 0x0;
 
-   ASSERT(mov->Opcode == OPCODE_MOV);
+   assert(mov->Opcode == OPCODE_MOV);
 
    for (comp = 0; comp < 4; ++comp) {
       GLuint src_comp;
@@ -225,13 +225,13 @@ replace_regs(struct gl_program *prog, gl_register_file file, const GLint map[])
       for (j = 0; j < numSrc; j++) {
          if (inst->SrcReg[j].File == file) {
             GLuint index = inst->SrcReg[j].Index;
-            ASSERT(map[index] >= 0);
+            assert(map[index] >= 0);
             inst->SrcReg[j].Index = map[index];
          }
       }
       if (inst->DstReg.File == file) {
          const GLuint index = inst->DstReg.Index;
-         ASSERT(map[index] >= 0);
+         assert(map[index] >= 0);
          inst->DstReg.Index = map[index];
       }
    }
@@ -272,7 +272,7 @@ _mesa_remove_dead_code_global(struct gl_program *prog)
          if (inst->SrcReg[j].File == PROGRAM_TEMPORARY) {
             const GLuint index = inst->SrcReg[j].Index;
             GLuint read_mask;
-            ASSERT(index < REG_ALLOCATE_MAX_PROGRAM_TEMPS);
+            assert(index < REG_ALLOCATE_MAX_PROGRAM_TEMPS);
 	    read_mask = get_src_arg_mask(inst, j, NO_MASK);
 
             if (inst->SrcReg[j].RelAddr) {
@@ -295,7 +295,7 @@ _mesa_remove_dead_code_global(struct gl_program *prog)
       /* check dst reg */
       if (inst->DstReg.File == PROGRAM_TEMPORARY) {
          const GLuint index = inst->DstReg.Index;
-         ASSERT(index < REG_ALLOCATE_MAX_PROGRAM_TEMPS);
+         assert(index < REG_ALLOCATE_MAX_PROGRAM_TEMPS);
 
          if (inst->DstReg.RelAddr) {
             if (dbg)
@@ -676,7 +676,7 @@ _mesa_merge_mov_into_inst(struct prog_instruction *inst,
       for (dst_comp = 0; dst_comp < 4; ++dst_comp) {
          if (mov->DstReg.WriteMask & (1 << dst_comp)) {
             const GLuint src_comp = GET_SWZ(mov->SrcReg[0].Swizzle, dst_comp);
-            ASSERT(src_comp < 4);
+            assert(src_comp < 4);
             dst_to_src_comp[dst_comp] = src_comp;
          }
       }
@@ -692,9 +692,9 @@ _mesa_merge_mov_into_inst(struct prog_instruction *inst,
             if ((mov->DstReg.WriteMask & (1 << dst_comp)) == 0)
                continue;
             src_comp = dst_to_src_comp[dst_comp];
-            ASSERT(src_comp < 4);
+            assert(src_comp < 4);
             arg_comp = GET_SWZ(arg_swz, src_comp);
-            ASSERT(arg_comp < 4);
+            assert(arg_comp < 4);
             inst->SrcReg[arg].Swizzle |= arg_comp << (3*dst_comp);
          }
       }
@@ -871,7 +871,7 @@ insert_interval_by_end(struct interval_list *list, const struct interval *inv)
    {
       GLuint i;
       for (i = 0; i + 1 < list->Num; i++) {
-         ASSERT(list->Intervals[i].End <= list->Intervals[i + 1].End);
+         assert(list->Intervals[i].End <= list->Intervals[i + 1].End);
       }
    }
 #endif
@@ -887,8 +887,8 @@ remove_interval(struct interval_list *list, const struct interval *inv)
    for (k = 0; k < list->Num; k++) {
       if (list->Intervals[k].Reg == inv->Reg) {
          /* found, remove it */
-         ASSERT(list->Intervals[k].Start == inv->Start);
-         ASSERT(list->Intervals[k].End == inv->End);
+         assert(list->Intervals[k].Start == inv->Start);
+         assert(list->Intervals[k].End == inv->End);
          while (k < list->Num - 1) {
             list->Intervals[k] = list->Intervals[k + 1];
             k++;
@@ -924,7 +924,7 @@ sort_interval_list_by_start(struct interval_list *list)
    {
       GLuint i;
       for (i = 0; i + 1 < list->Num; i++) {
-         ASSERT(list->Intervals[i].Start <= list->Intervals[i + 1].Start);
+         assert(list->Intervals[i].Start <= list->Intervals[i + 1].Start);
       }
    }
 #endif
@@ -966,9 +966,9 @@ update_interval(GLint intBegin[], GLint intEnd[],
       begin = loopStack[0].Start;
    }
 
-   ASSERT(index < REG_ALLOCATE_MAX_PROGRAM_TEMPS);
+   assert(index < REG_ALLOCATE_MAX_PROGRAM_TEMPS);
    if (intBegin[index] == -1) {
-      ASSERT(intEnd[index] == -1);
+      assert(intEnd[index] == -1);
       intBegin[index] = begin;
       intEnd[index] = end;
    }
@@ -1176,7 +1176,7 @@ _mesa_reallocate_registers(struct gl_program *prog)
                else {
                   /* Interval 'inv' has expired */
                   const GLint regNew = registerMap[inv->Reg];
-                  ASSERT(regNew >= 0);
+                  assert(regNew >= 0);
 
                   if (dbg)
                      printf("  expire interval for reg %u\n", inv->Reg);
@@ -1188,7 +1188,7 @@ _mesa_reallocate_registers(struct gl_program *prog)
                   /* return register regNew to the free pool */
                   if (dbg)
                      printf("  free reg %d\n", regNew);
-                  ASSERT(usedRegs[regNew] == GL_TRUE);
+                  assert(usedRegs[regNew] == GL_TRUE);
                   usedRegs[regNew] = GL_FALSE;
                }
             }
diff --git a/mesalib/src/mesa/program/prog_parameter.c b/mesalib/src/mesa/program/prog_parameter.c
index 0ef46415d..5939f6f72 100644
--- a/mesalib/src/mesa/program/prog_parameter.c
+++ b/mesalib/src/mesa/program/prog_parameter.c
@@ -241,8 +241,8 @@ _mesa_add_typed_unnamed_constant(struct gl_program_parameter_list *paramList,
                            GLenum datatype, GLuint *swizzleOut)
 {
    GLint pos;
-   ASSERT(size >= 1);
-   ASSERT(size <= 4);
+   assert(size >= 1);
+   assert(size <= 4);
 
    if (swizzleOut &&
        _mesa_lookup_parameter_constant(paramList, values,
@@ -528,7 +528,7 @@ _mesa_clone_parameter_list(const struct gl_program_parameter_list *list)
       GLuint size = MIN2(p->Size, 4);
       GLint j = _mesa_add_parameter(clone, p->Type, p->Name, size, p->DataType,
                                     list->ParameterValues[i], NULL);
-      ASSERT(j >= 0);
+      assert(j >= 0);
       pCopy = clone->Parameters + j;
       /* copy state indexes */
       if (p->Type == PROGRAM_STATE_VAR) {
diff --git a/mesalib/src/mesa/program/prog_parameter_layout.c b/mesalib/src/mesa/program/prog_parameter_layout.c
index e83469059..282a367ad 100644
--- a/mesalib/src/mesa/program/prog_parameter_layout.c
+++ b/mesalib/src/mesa/program/prog_parameter_layout.c
@@ -28,7 +28,7 @@
  * \author Ian Romanick <ian.d.romanick@intel.com>
  */
 
-#include "main/compiler.h"
+#include "main/imports.h"
 #include "main/mtypes.h"
 #include "prog_parameter.h"
 #include "prog_parameter_layout.h"
diff --git a/mesalib/src/mesa/program/prog_print.c b/mesalib/src/mesa/program/prog_print.c
index 3f499749a..d588d07ff 100644
--- a/mesalib/src/mesa/program/prog_print.c
+++ b/mesalib/src/mesa/program/prog_print.c
@@ -182,20 +182,20 @@ arb_input_attrib_string(GLuint index, GLenum progType)
    };
 
    /* sanity checks */
-   STATIC_ASSERT(Elements(vertAttribs) == VERT_ATTRIB_MAX);
-   STATIC_ASSERT(Elements(fragAttribs) == VARYING_SLOT_MAX);
+   STATIC_ASSERT(ARRAY_SIZE(vertAttribs) == VERT_ATTRIB_MAX);
+   STATIC_ASSERT(ARRAY_SIZE(fragAttribs) == VARYING_SLOT_MAX);
    assert(strcmp(vertAttribs[VERT_ATTRIB_TEX0], "vertex.texcoord[0]") == 0);
    assert(strcmp(vertAttribs[VERT_ATTRIB_GENERIC15], "vertex.attrib[15]") == 0);
    assert(strcmp(fragAttribs[VARYING_SLOT_TEX0], "fragment.texcoord[0]") == 0);
    assert(strcmp(fragAttribs[VARYING_SLOT_VAR0+15], "fragment.varying[15]") == 0);
 
    if (progType == GL_VERTEX_PROGRAM_ARB) {
-      assert(index < Elements(vertAttribs));
+      assert(index < ARRAY_SIZE(vertAttribs));
       return vertAttribs[index];
    }
    else {
       assert(progType == GL_FRAGMENT_PROGRAM_ARB);
-      assert(index < Elements(fragAttribs));
+      assert(index < ARRAY_SIZE(fragAttribs));
       return fragAttribs[index];
    }
 }
@@ -321,19 +321,19 @@ arb_output_attrib_string(GLuint index, GLenum progType)
    };
 
    /* sanity checks */
-   STATIC_ASSERT(Elements(vertResults) == VARYING_SLOT_MAX);
-   STATIC_ASSERT(Elements(fragResults) == FRAG_RESULT_MAX);
+   STATIC_ASSERT(ARRAY_SIZE(vertResults) == VARYING_SLOT_MAX);
+   STATIC_ASSERT(ARRAY_SIZE(fragResults) == FRAG_RESULT_MAX);
    assert(strcmp(vertResults[VARYING_SLOT_POS], "result.position") == 0);
    assert(strcmp(vertResults[VARYING_SLOT_VAR0], "result.varying[0]") == 0);
    assert(strcmp(fragResults[FRAG_RESULT_DATA0], "result.color[0]") == 0);
 
    if (progType == GL_VERTEX_PROGRAM_ARB) {
-      assert(index < Elements(vertResults));
+      assert(index < ARRAY_SIZE(vertResults));
       return vertResults[index];
    }
    else {
       assert(progType == GL_FRAGMENT_PROGRAM_ARB);
-      assert(index < Elements(fragResults));
+      assert(index < ARRAY_SIZE(fragResults));
       return fragResults[index];
    }
 }
diff --git a/mesalib/src/mesa/program/prog_statevars.c b/mesalib/src/mesa/program/prog_statevars.c
index 7f5daf8c6..57b25a7e3 100644
--- a/mesalib/src/mesa/program/prog_statevars.c
+++ b/mesalib/src/mesa/program/prog_statevars.c
@@ -29,6 +29,7 @@
  */
 
 
+#include <stdio.h>
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/blend.h"
@@ -41,6 +42,9 @@
 #include "main/samplerobj.h"
 
 
+#define ONE_DIV_SQRT_LN2 (1.201122408786449815)
+
+
 /**
  * Use the list of tokens in the state[] array to find global GL state
  * and return it in <value>.  Usually, four values are returned in <value>
@@ -58,9 +62,9 @@ _mesa_fetch_state(struct gl_context *ctx, const gl_state_index state[],
          /* state[1] is either 0=front or 1=back side */
          const GLuint face = (GLuint) state[1];
          const struct gl_material *mat = &ctx->Light.Material;
-         ASSERT(face == 0 || face == 1);
+         assert(face == 0 || face == 1);
          /* we rely on tokens numbered so that _BACK_ == _FRONT_+ 1 */
-         ASSERT(MAT_ATTRIB_FRONT_AMBIENT + 1 == MAT_ATTRIB_BACK_AMBIENT);
+         assert(MAT_ATTRIB_FRONT_AMBIENT + 1 == MAT_ATTRIB_BACK_AMBIENT);
          /* XXX we could get rid of this switch entirely with a little
           * work in arbprogparse.c's parse_state_single_item().
           */
@@ -170,7 +174,7 @@ _mesa_fetch_state(struct gl_context *ctx, const gl_state_index state[],
          const GLuint ln = (GLuint) state[1];
          const GLuint face = (GLuint) state[2];
          GLint i;
-         ASSERT(face == 0 || face == 1);
+         assert(face == 0 || face == 1);
          switch (state[3]) {
             case STATE_AMBIENT:
                for (i = 0; i < 3; i++) {
@@ -295,8 +299,8 @@ _mesa_fetch_state(struct gl_context *ctx, const gl_state_index state[],
          const gl_state_index modifier = state[4];
          const GLfloat *m;
          GLuint row, i;
-         ASSERT(firstRow < 4);
-         ASSERT(lastRow < 4);
+         assert(firstRow < 4);
+         assert(lastRow < 4);
          if (mat == STATE_MODELVIEW_MATRIX) {
             matrix = ctx->ModelviewMatrixStack.Top;
          }
@@ -307,11 +311,11 @@ _mesa_fetch_state(struct gl_context *ctx, const gl_state_index state[],
             matrix = &ctx->_ModelProjectMatrix;
          }
          else if (mat == STATE_TEXTURE_MATRIX) {
-            ASSERT(index < Elements(ctx->TextureMatrixStack));
+            assert(index < ARRAY_SIZE(ctx->TextureMatrixStack));
             matrix = ctx->TextureMatrixStack[index].Top;
          }
          else if (mat == STATE_PROGRAM_MATRIX) {
-            ASSERT(index < Elements(ctx->ProgramMatrixStack));
+            assert(index < ARRAY_SIZE(ctx->ProgramMatrixStack));
             matrix = ctx->ProgramMatrixStack[index].Top;
          }
          else {
diff --git a/mesalib/src/mesa/program/program.c b/mesalib/src/mesa/program/program.c
index 6e17fe65f..61a9e97c9 100644
--- a/mesalib/src/mesa/program/program.c
+++ b/mesalib/src/mesa/program/program.c
@@ -56,21 +56,21 @@ _mesa_init_program(struct gl_context *ctx)
     * If this assertion fails, we need to increase the field
     * size for register indexes (see INST_INDEX_BITS).
     */
-   ASSERT(ctx->Const.Program[MESA_SHADER_VERTEX].MaxUniformComponents / 4
+   assert(ctx->Const.Program[MESA_SHADER_VERTEX].MaxUniformComponents / 4
           <= (1 << INST_INDEX_BITS));
-   ASSERT(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxUniformComponents / 4
+   assert(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxUniformComponents / 4
           <= (1 << INST_INDEX_BITS));
 
-   ASSERT(ctx->Const.Program[MESA_SHADER_VERTEX].MaxTemps <= (1 << INST_INDEX_BITS));
-   ASSERT(ctx->Const.Program[MESA_SHADER_VERTEX].MaxLocalParams <= (1 << INST_INDEX_BITS));
-   ASSERT(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxTemps <= (1 << INST_INDEX_BITS));
-   ASSERT(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxLocalParams <= (1 << INST_INDEX_BITS));
+   assert(ctx->Const.Program[MESA_SHADER_VERTEX].MaxTemps <= (1 << INST_INDEX_BITS));
+   assert(ctx->Const.Program[MESA_SHADER_VERTEX].MaxLocalParams <= (1 << INST_INDEX_BITS));
+   assert(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxTemps <= (1 << INST_INDEX_BITS));
+   assert(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxLocalParams <= (1 << INST_INDEX_BITS));
 
-   ASSERT(ctx->Const.Program[MESA_SHADER_VERTEX].MaxUniformComponents <= 4 * MAX_UNIFORMS);
-   ASSERT(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxUniformComponents <= 4 * MAX_UNIFORMS);
+   assert(ctx->Const.Program[MESA_SHADER_VERTEX].MaxUniformComponents <= 4 * MAX_UNIFORMS);
+   assert(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxUniformComponents <= 4 * MAX_UNIFORMS);
 
-   ASSERT(ctx->Const.Program[MESA_SHADER_VERTEX].MaxAddressOffset <= (1 << INST_INDEX_BITS));
-   ASSERT(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxAddressOffset <= (1 << INST_INDEX_BITS));
+   assert(ctx->Const.Program[MESA_SHADER_VERTEX].MaxAddressOffset <= (1 << INST_INDEX_BITS));
+   assert(ctx->Const.Program[MESA_SHADER_FRAGMENT].MaxAddressOffset <= (1 << INST_INDEX_BITS));
 
    /* If this fails, increase prog_instruction::TexSrcUnit size */
    STATIC_ASSERT(MAX_TEXTURE_UNITS <= (1 << 5));
@@ -364,8 +364,8 @@ void
 _mesa_delete_program(struct gl_context *ctx, struct gl_program *prog)
 {
    (void) ctx;
-   ASSERT(prog);
-   ASSERT(prog->RefCount==0);
+   assert(prog);
+   assert(prog->RefCount==0);
 
    if (prog == &_mesa_DummyProgram)
       return;
@@ -414,12 +414,12 @@ _mesa_reference_program_(struct gl_context *ctx,
    if (*ptr && prog) {
       /* sanity check */
       if ((*ptr)->Target == GL_VERTEX_PROGRAM_ARB)
-         ASSERT(prog->Target == GL_VERTEX_PROGRAM_ARB);
+         assert(prog->Target == GL_VERTEX_PROGRAM_ARB);
       else if ((*ptr)->Target == GL_FRAGMENT_PROGRAM_ARB)
-         ASSERT(prog->Target == GL_FRAGMENT_PROGRAM_ARB ||
+         assert(prog->Target == GL_FRAGMENT_PROGRAM_ARB ||
                 prog->Target == GL_FRAGMENT_PROGRAM_NV);
       else if ((*ptr)->Target == MESA_GEOMETRY_PROGRAM)
-         ASSERT(prog->Target == MESA_GEOMETRY_PROGRAM);
+         assert(prog->Target == MESA_GEOMETRY_PROGRAM);
    }
 #endif
 
@@ -434,14 +434,14 @@ _mesa_reference_program_(struct gl_context *ctx,
               ((*ptr)->Target == MESA_GEOMETRY_PROGRAM ? "GP" : "FP")),
              (*ptr)->RefCount - 1);
 #endif
-      ASSERT((*ptr)->RefCount > 0);
+      assert((*ptr)->RefCount > 0);
       (*ptr)->RefCount--;
 
       deleteFlag = ((*ptr)->RefCount == 0);
       /*mtx_lock(&(*ptr)->Mutex);*/
 
       if (deleteFlag) {
-         ASSERT(ctx);
+         assert(ctx);
          ctx->Driver.DeleteProgram(ctx, *ptr);
       }
 
@@ -733,7 +733,7 @@ _mesa_combine_programs(struct gl_context *ctx,
    GLbitfield64 inputsB;
    GLuint i;
 
-   ASSERT(progA->Target == progB->Target);
+   assert(progA->Target == progB->Target);
 
    newInst = _mesa_alloc_instructions(newLength);
    if (!newInst)
@@ -867,14 +867,14 @@ _mesa_find_used_registers(const struct gl_program *prog,
       const GLuint n = _mesa_num_inst_src_regs(inst->Opcode);
 
       if (inst->DstReg.File == file) {
-         ASSERT(inst->DstReg.Index < usedSize);
+         assert(inst->DstReg.Index < usedSize);
          if(inst->DstReg.Index < usedSize)
             used[inst->DstReg.Index] = GL_TRUE;
       }
 
       for (j = 0; j < n; j++) {
          if (inst->SrcReg[j].File == file) {
-            ASSERT(inst->SrcReg[j].Index < (GLint) usedSize);
+            assert(inst->SrcReg[j].Index < (GLint) usedSize);
             if (inst->SrcReg[j].Index < (GLint) usedSize)
                used[inst->SrcReg[j].Index] = GL_TRUE;
          }
diff --git a/mesalib/src/mesa/program/program.h b/mesalib/src/mesa/program/program.h
index ef698242f..0b0d1ac6f 100644
--- a/mesalib/src/mesa/program/program.h
+++ b/mesalib/src/mesa/program/program.h
@@ -210,7 +210,7 @@ _mesa_program_enum_to_shader_stage(GLenum v)
    case GL_COMPUTE_PROGRAM_NV:
       return MESA_SHADER_COMPUTE;
    default:
-      ASSERT(0);
+      assert(0);
       return ~0;
    }
 }
diff --git a/mesalib/src/mesa/program/program_parse.y b/mesalib/src/mesa/program/program_parse.y
index 1664740b4..716b83d2d 100644
--- a/mesalib/src/mesa/program/program_parse.y
+++ b/mesalib/src/mesa/program/program_parse.y
@@ -21,6 +21,8 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  */
+
+#include <stdarg.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -2333,11 +2335,11 @@ set_dst_reg(struct prog_dst_register *r, gl_register_file file, GLint index)
 {
    const GLint maxIndex = 1 << INST_INDEX_BITS;
    const GLint minIndex = 0;
-   ASSERT(index >= minIndex);
+   assert(index >= minIndex);
    (void) minIndex;
-   ASSERT(index <= maxIndex);
+   assert(index <= maxIndex);
    (void) maxIndex;
-   ASSERT(file == PROGRAM_TEMPORARY ||
+   assert(file == PROGRAM_TEMPORARY ||
 	  file == PROGRAM_ADDRESS ||
 	  file == PROGRAM_OUTPUT);
    memset(r, 0, sizeof(*r));
@@ -2375,10 +2377,10 @@ set_src_reg_swz(struct asm_src_register *r, gl_register_file file, GLint index,
 {
    const GLint maxIndex = (1 << INST_INDEX_BITS) - 1;
    const GLint minIndex = -(1 << INST_INDEX_BITS);
-   ASSERT(file < PROGRAM_FILE_MAX);
-   ASSERT(index >= minIndex);
+   assert(file < PROGRAM_FILE_MAX);
+   assert(index >= minIndex);
    (void) minIndex;
-   ASSERT(index <= maxIndex);
+   assert(index <= maxIndex);
    (void) maxIndex;
    memset(r, 0, sizeof(*r));
    r->Base.File = file;
diff --git a/mesalib/src/mesa/program/programopt.c b/mesalib/src/mesa/program/programopt.c
index fdaa4a465..e82c68a53 100644
--- a/mesalib/src/mesa/program/programopt.c
+++ b/mesalib/src/mesa/program/programopt.c
@@ -335,7 +335,7 @@ _mesa_append_fog_code(struct gl_context *ctx,
       inst++;
    }
    else {
-      ASSERT(fog_mode == GL_EXP || fog_mode == GL_EXP2);
+      assert(fog_mode == GL_EXP || fog_mode == GL_EXP2);
       /* fogPRefOpt.z = d/ln(2), fogPRefOpt.w = d/sqrt(ln(2) */
       /* EXP: MUL fogFactorTemp.x, fogPRefOpt.z, fragment.fogcoord.x; */
       /* EXP2: MUL fogFactorTemp.x, fogPRefOpt.w, fragment.fogcoord.x; */
diff --git a/mesalib/src/mesa/program/sampler.cpp b/mesalib/src/mesa/program/sampler.cpp
index f8584c968..ea3024d51 100644
--- a/mesalib/src/mesa/program/sampler.cpp
+++ b/mesalib/src/mesa/program/sampler.cpp
@@ -29,7 +29,6 @@
 #include "../glsl/program.h"
 #include "ir_uniform.h"
 
-#include "main/compiler.h"
 #include "main/mtypes.h"
 #include "program/hash_table.h"
 #include "program/prog_parameter.h"
diff --git a/mesalib/src/mesa/state_tracker/st_atom.c b/mesalib/src/mesa/state_tracker/st_atom.c
index 625ea2965..f0fe11ffa 100644
--- a/mesalib/src/mesa/state_tracker/st_atom.c
+++ b/mesalib/src/mesa/state_tracker/st_atom.c
@@ -26,6 +26,7 @@
  **************************************************************************/
 
 
+#include <stdio.h>
 #include "main/glheader.h"
 #include "main/context.h"
 
@@ -197,7 +198,7 @@ void st_validate_state( struct st_context *st )
       memset(&examined, 0, sizeof(examined));
       prev = *state;
 
-      for (i = 0; i < Elements(atoms); i++) {	 
+      for (i = 0; i < ARRAY_SIZE(atoms); i++) {	 
 	 const struct st_tracked_state *atom = atoms[i];
 	 struct st_state_flags generated;
 	 
@@ -228,7 +229,7 @@ void st_validate_state( struct st_context *st )
 
    }
    else {
-      for (i = 0; i < Elements(atoms); i++) {	 
+      for (i = 0; i < ARRAY_SIZE(atoms); i++) {	 
 	 if (check_state(state, &atoms[i]->dirty))
 	    atoms[i]->update( st );
       }
diff --git a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
index f24805cf5..1dbc4b905 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.c
@@ -104,9 +104,9 @@ st_bufferobj_subdata(struct gl_context *ctx,
    struct st_buffer_object *st_obj = st_buffer_object(obj);
 
    /* we may be called from VBO code, so double-check params here */
-   ASSERT(offset >= 0);
-   ASSERT(size >= 0);
-   ASSERT(offset + size <= obj->Size);
+   assert(offset >= 0);
+   assert(size >= 0);
+   assert(offset + size <= obj->Size);
 
    if (!size)
       return;
@@ -148,9 +148,9 @@ st_bufferobj_get_subdata(struct gl_context *ctx,
    struct st_buffer_object *st_obj = st_buffer_object(obj);
 
    /* we may be called from VBO code, so double-check params here */
-   ASSERT(offset >= 0);
-   ASSERT(size >= 0);
-   ASSERT(offset + size <= obj->Size);
+   assert(offset >= 0);
+   assert(size >= 0);
+   assert(offset + size <= obj->Size);
 
    if (!size)
       return;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h
index a5ab12219..647efe4cc 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h
+++ b/mesalib/src/mesa/state_tracker/st_cb_bufferobjects.h
@@ -48,7 +48,7 @@ struct st_buffer_object
 
 
 /** cast wrapper */
-static INLINE struct st_buffer_object *
+static inline struct st_buffer_object *
 st_buffer_object(struct gl_buffer_object *obj)
 {
    return (struct st_buffer_object *) obj;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_clear.c b/mesalib/src/mesa/state_tracker/st_cb_clear.c
index 45dea594e..dd81a6273 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_clear.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_clear.c
@@ -102,7 +102,7 @@ st_destroy_clear(struct st_context *st)
 /**
  * Helper function to set the fragment shaders.
  */
-static INLINE void
+static inline void
 set_fragment_shader(struct st_context *st)
 {
    if (!st->clear.fs)
@@ -118,7 +118,7 @@ set_fragment_shader(struct st_context *st)
 /**
  * Helper function to set the vertex shader.
  */
-static INLINE void
+static inline void
 set_vertex_shader(struct st_context *st)
 {
    /* vertex shader - still required to provide the linkage between
@@ -381,7 +381,7 @@ clear_with_quad(struct gl_context *ctx, unsigned clear_buffers)
 /**
  * Return if the scissor must be enabled during the clear.
  */
-static INLINE GLboolean
+static inline GLboolean
 is_scissor_enabled(struct gl_context *ctx, struct gl_renderbuffer *rb)
 {
    return (ctx->Scissor.EnableFlags & 1) &&
@@ -395,7 +395,7 @@ is_scissor_enabled(struct gl_context *ctx, struct gl_renderbuffer *rb)
 /**
  * Return if all of the color channels are masked.
  */
-static INLINE GLboolean
+static inline GLboolean
 is_color_disabled(struct gl_context *ctx, int i)
 {
    return !ctx->Color.ColorMask[i][0] &&
@@ -408,7 +408,7 @@ is_color_disabled(struct gl_context *ctx, int i)
 /**
  * Return if any of the color channels are masked.
  */
-static INLINE GLboolean
+static inline GLboolean
 is_color_masked(struct gl_context *ctx, int i)
 {
    return !ctx->Color.ColorMask[i][0] ||
@@ -421,7 +421,7 @@ is_color_masked(struct gl_context *ctx, int i)
 /**
  * Return if all of the stencil bits are masked.
  */
-static INLINE GLboolean
+static inline GLboolean
 is_stencil_disabled(struct gl_context *ctx, struct gl_renderbuffer *rb)
 {
    const GLuint stencilMax = 0xff;
@@ -434,7 +434,7 @@ is_stencil_disabled(struct gl_context *ctx, struct gl_renderbuffer *rb)
 /**
  * Return if any of the stencil bits are masked.
  */
-static INLINE GLboolean
+static inline GLboolean
 is_stencil_masked(struct gl_context *ctx, struct gl_renderbuffer *rb)
 {
    const GLuint stencilMax = 0xff;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
index 14fc13952..3edf31bad 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_drawpixels.c
@@ -201,7 +201,7 @@ st_make_drawpix_z_stencil_program(struct st_context *st,
    GLuint ic = 0;
    const GLuint shaderIndex = write_depth * 2 + write_stencil;
 
-   assert(shaderIndex < Elements(st->drawpix.shaders));
+   assert(shaderIndex < ARRAY_SIZE(st->drawpix.shaders));
 
    if (st->drawpix.shaders[shaderIndex]) {
       /* already have the proper shader */
@@ -1684,7 +1684,7 @@ st_destroy_drawpix(struct st_context *st)
 {
    GLuint i;
 
-   for (i = 0; i < Elements(st->drawpix.shaders); i++) {
+   for (i = 0; i < ARRAY_SIZE(st->drawpix.shaders); i++) {
       if (st->drawpix.shaders[i])
          _mesa_reference_fragprog(st->ctx, &st->drawpix.shaders[i], NULL);
    }
diff --git a/mesalib/src/mesa/state_tracker/st_cb_fbo.h b/mesalib/src/mesa/state_tracker/st_cb_fbo.h
index ae13aa677..f3b310b07 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_fbo.h
+++ b/mesalib/src/mesa/state_tracker/st_cb_fbo.h
@@ -65,7 +65,7 @@ struct st_renderbuffer
 };
 
 
-static INLINE struct st_renderbuffer *
+static inline struct st_renderbuffer *
 st_renderbuffer(struct gl_renderbuffer *rb)
 {
    return (struct st_renderbuffer *) rb;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_feedback.c b/mesalib/src/mesa/state_tracker/st_cb_feedback.c
index 6b504b8ab..c5410886a 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_feedback.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_feedback.c
@@ -69,7 +69,7 @@ struct feedback_stage
  * GL Feedback functions
  **********************************************************************/
 
-static INLINE struct feedback_stage *
+static inline struct feedback_stage *
 feedback_stage( struct draw_stage *stage )
 {
    return (struct feedback_stage *)stage;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_flush.c b/mesalib/src/mesa/state_tracker/st_cb_flush.c
index 5544c6644..6df430049 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_flush.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_flush.c
@@ -47,7 +47,7 @@
 
 
 /** Check if we have a front color buffer and if it's been drawn to. */
-static INLINE GLboolean
+static inline GLboolean
 is_front_buffer_dirty(struct st_context *st)
 {
    struct gl_framebuffer *fb = st->ctx->DrawBuffer;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_program.c b/mesalib/src/mesa/state_tracker/st_cb_program.c
index faff9c4bd..aa301d830 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_program.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_program.c
@@ -143,11 +143,6 @@ st_delete_program(struct gl_context *ctx, struct gl_program *prog)
          
          if (stgp->glsl_to_tgsi)
             free_glsl_to_tgsi_visitor(stgp->glsl_to_tgsi);
-
-         if (stgp->tgsi.tokens) {
-            st_free_tokens((void *) stgp->tgsi.tokens);
-            stgp->tgsi.tokens = NULL;
-         }
       }
       break;
    case GL_FRAGMENT_PROGRAM_ARB:
@@ -207,11 +202,6 @@ st_program_string_notify( struct gl_context *ctx,
 
       st_release_gp_variants(st, stgp);
 
-      if (stgp->tgsi.tokens) {
-         st_free_tokens((void *) stgp->tgsi.tokens);
-         stgp->tgsi.tokens = NULL;
-      }
-
       if (st->gp == stgp)
 	 st->dirty.st |= ST_NEW_GEOMETRY_PROGRAM;
    }
diff --git a/mesalib/src/mesa/state_tracker/st_cb_queryobj.h b/mesalib/src/mesa/state_tracker/st_cb_queryobj.h
index e4c6c8643..24063211b 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_queryobj.h
+++ b/mesalib/src/mesa/state_tracker/st_cb_queryobj.h
@@ -49,7 +49,7 @@ struct st_query_object
 /**
  * Cast wrapper
  */
-static INLINE struct st_query_object *
+static inline struct st_query_object *
 st_query_object(struct gl_query_object *q)
 {
    return (struct st_query_object *) q;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_rasterpos.c b/mesalib/src/mesa/state_tracker/st_cb_rasterpos.c
index 370746536..272cbb91d 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_rasterpos.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_rasterpos.c
@@ -64,7 +64,7 @@ struct rastpos_stage
 };
 
 
-static INLINE struct rastpos_stage *
+static inline struct rastpos_stage *
 rastpos_stage( struct draw_stage *stage )
 {
    return (struct rastpos_stage *) stage;
@@ -191,7 +191,7 @@ new_draw_rastpos_stage(struct gl_context *ctx, struct draw_context *draw)
    rs->stage.destroy = rastpos_destroy;
    rs->ctx = ctx;
 
-   for (i = 0; i < Elements(rs->array); i++) {
+   for (i = 0; i < ARRAY_SIZE(rs->array); i++) {
       rs->array[i].Size = 4;
       rs->array[i].Type = GL_FLOAT;
       rs->array[i].Format = GL_RGBA;
diff --git a/mesalib/src/mesa/state_tracker/st_cb_texture.c b/mesalib/src/mesa/state_tracker/st_cb_texture.c
index 0525e879f..a8b19a1f3 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_texture.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_texture.c
@@ -25,6 +25,7 @@
  * 
  **************************************************************************/
 
+#include <stdio.h>
 #include "main/bufferobj.h"
 #include "main/enums.h"
 #include "main/fbobject.h"
diff --git a/mesalib/src/mesa/state_tracker/st_cb_viewport.c b/mesalib/src/mesa/state_tracker/st_cb_viewport.c
index 2bd55f6d2..d7a34125e 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_viewport.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_viewport.c
@@ -39,7 +39,7 @@
  * We'll only return non-null for window system framebuffers.
  * Note that this function may fail.
  */
-static INLINE struct st_framebuffer *
+static inline struct st_framebuffer *
 st_ws_framebuffer(struct gl_framebuffer *fb)
 {
    /* FBO cannot be casted.  See st_new_framebuffer */
diff --git a/mesalib/src/mesa/state_tracker/st_cb_xformfb.c b/mesalib/src/mesa/state_tracker/st_cb_xformfb.c
index a2bd86aff..07c118e22 100644
--- a/mesalib/src/mesa/state_tracker/st_cb_xformfb.c
+++ b/mesalib/src/mesa/state_tracker/st_cb_xformfb.c
@@ -59,7 +59,7 @@ struct st_transform_feedback_object {
    struct pipe_stream_output_target *draw_count;
 };
 
-static INLINE struct st_transform_feedback_object *
+static inline struct st_transform_feedback_object *
 st_transform_feedback_object(struct gl_transform_feedback_object *obj)
 {
    return (struct st_transform_feedback_object *) obj;
@@ -95,7 +95,7 @@ st_delete_transform_feedback(struct gl_context *ctx,
       pipe_so_target_reference(&sobj->targets[i], NULL);
    }
 
-   for (i = 0; i < Elements(sobj->base.Buffers); i++) {
+   for (i = 0; i < ARRAY_SIZE(sobj->base.Buffers); i++) {
       _mesa_reference_buffer_object(ctx, &sobj->base.Buffers[i], NULL);
    }
 
@@ -115,8 +115,8 @@ st_begin_transform_feedback(struct gl_context *ctx, GLenum mode,
    unsigned i, max_num_targets;
    unsigned offsets[PIPE_MAX_SO_BUFFERS] = {0};
 
-   max_num_targets = MIN2(Elements(sobj->base.Buffers),
-                          Elements(sobj->targets));
+   max_num_targets = MIN2(ARRAY_SIZE(sobj->base.Buffers),
+                          ARRAY_SIZE(sobj->targets));
 
    /* Convert the transform feedback state into the gallium representation. */
    for (i = 0; i < max_num_targets; i++) {
@@ -185,7 +185,7 @@ st_transform_feedback_get_draw_target(struct gl_transform_feedback_object *obj)
          st_transform_feedback_object(obj);
    unsigned i;
 
-   for (i = 0; i < Elements(sobj->targets); i++) {
+   for (i = 0; i < ARRAY_SIZE(sobj->targets); i++) {
       if (sobj->targets[i]) {
          return sobj->targets[i];
       }
diff --git a/mesalib/src/mesa/state_tracker/st_context.c b/mesalib/src/mesa/state_tracker/st_context.c
index 5834ebad3..5fe132ac2 100644
--- a/mesalib/src/mesa/state_tracker/st_context.c
+++ b/mesalib/src/mesa/state_tracker/st_context.c
@@ -117,8 +117,8 @@ st_destroy_context_priv(struct st_context *st)
    st_destroy_drawpix(st);
    st_destroy_drawtex(st);
 
-   for (shader = 0; shader < Elements(st->state.sampler_views); shader++) {
-      for (i = 0; i < Elements(st->state.sampler_views[0]); i++) {
+   for (shader = 0; shader < ARRAY_SIZE(st->state.sampler_views); shader++) {
+      for (i = 0; i < ARRAY_SIZE(st->state.sampler_views[0]); i++) {
          pipe_sampler_view_release(st->pipe,
                                    &st->state.sampler_views[shader][i]);
       }
@@ -200,7 +200,7 @@ st_create_context_priv( struct gl_context *ctx, struct pipe_context *pipe,
    /* Vertex element objects used for drawing rectangles for glBitmap,
     * glDrawPixels, glClear, etc.
     */
-   for (i = 0; i < Elements(st->velems_util_draw); i++) {
+   for (i = 0; i < ARRAY_SIZE(st->velems_util_draw); i++) {
       memset(&st->velems_util_draw[i], 0, sizeof(struct pipe_vertex_element));
       st->velems_util_draw[i].src_offset = i * 4 * sizeof(float);
       st->velems_util_draw[i].instance_divisor = 0;
diff --git a/mesalib/src/mesa/state_tracker/st_context.h b/mesalib/src/mesa/state_tracker/st_context.h
index b091a8856..8a9504bb7 100644
--- a/mesalib/src/mesa/state_tracker/st_context.h
+++ b/mesalib/src/mesa/state_tracker/st_context.h
@@ -214,7 +214,7 @@ struct st_context
 
 /* Need this so that we can implement Mesa callbacks in this module.
  */
-static INLINE struct st_context *st_context(struct gl_context *ctx)
+static inline struct st_context *st_context(struct gl_context *ctx)
 {
    return ctx->st;
 }
@@ -246,7 +246,7 @@ void st_invalidate_state(struct gl_context * ctx, GLuint new_state);
 #define Y_0_TOP 1
 #define Y_0_BOTTOM 2
 
-static INLINE GLuint
+static inline GLuint
 st_fb_orientation(const struct gl_framebuffer *fb)
 {
    if (fb && _mesa_is_winsys_fbo(fb)) {
diff --git a/mesalib/src/mesa/state_tracker/st_debug.h b/mesalib/src/mesa/state_tracker/st_debug.h
index 49b916fb3..cc8197836 100644
--- a/mesalib/src/mesa/state_tracker/st_debug.h
+++ b/mesalib/src/mesa/state_tracker/st_debug.h
@@ -58,7 +58,7 @@ extern int ST_DEBUG;
 
 void st_debug_init( void );
 
-static INLINE void
+static inline void
 ST_DBG( unsigned flag, const char *fmt, ... )
 {
     if (ST_DEBUG & flag)
diff --git a/mesalib/src/mesa/state_tracker/st_draw.h b/mesalib/src/mesa/state_tracker/st_draw.h
index 361fe7f91..780d4bde7 100644
--- a/mesalib/src/mesa/state_tracker/st_draw.h
+++ b/mesalib/src/mesa/state_tracker/st_draw.h
@@ -76,7 +76,7 @@ st_feedback_draw_vbo(struct gl_context *ctx,
  * This function is basically a cast wrapper to avoid warnings when building
  * in 64-bit mode.
  */
-static INLINE unsigned
+static inline unsigned
 pointer_to_offset(const void *ptr)
 {
    return (unsigned) (((GLsizeiptr) ptr) & 0xffffffffUL);
diff --git a/mesalib/src/mesa/state_tracker/st_extensions.c b/mesalib/src/mesa/state_tracker/st_extensions.c
index ce29d076c..bc20f7323 100644
--- a/mesalib/src/mesa/state_tracker/st_extensions.c
+++ b/mesalib/src/mesa/state_tracker/st_extensions.c
@@ -347,8 +347,8 @@ init_format_extensions(struct pipe_screen *screen,
    GLboolean *extension_table = (GLboolean *) extensions;
    unsigned i;
    int j;
-   int num_formats = Elements(mapping->format);
-   int num_ext = Elements(mapping->extension_offset);
+   int num_formats = ARRAY_SIZE(mapping->format);
+   int num_ext = ARRAY_SIZE(mapping->extension_offset);
 
    for (i = 0; i < num_mappings; i++) {
       int num_supported = 0;
@@ -630,7 +630,7 @@ void st_init_extensions(struct pipe_screen *screen,
    extensions->OES_draw_texture = GL_TRUE;
 
    /* Expose the extensions which directly correspond to gallium caps. */
-   for (i = 0; i < Elements(cap_mapping); i++) {
+   for (i = 0; i < ARRAY_SIZE(cap_mapping); i++) {
       if (screen->get_param(screen, cap_mapping[i].cap)) {
          extension_table[cap_mapping[i].extension_offset] = GL_TRUE;
       }
@@ -638,16 +638,16 @@ void st_init_extensions(struct pipe_screen *screen,
 
    /* Expose the extensions which directly correspond to gallium formats. */
    init_format_extensions(screen, extensions, rendertarget_mapping,
-                          Elements(rendertarget_mapping), PIPE_TEXTURE_2D,
+                          ARRAY_SIZE(rendertarget_mapping), PIPE_TEXTURE_2D,
                           PIPE_BIND_RENDER_TARGET | PIPE_BIND_SAMPLER_VIEW);
    init_format_extensions(screen, extensions, depthstencil_mapping,
-                          Elements(depthstencil_mapping), PIPE_TEXTURE_2D,
+                          ARRAY_SIZE(depthstencil_mapping), PIPE_TEXTURE_2D,
                           PIPE_BIND_DEPTH_STENCIL | PIPE_BIND_SAMPLER_VIEW);
    init_format_extensions(screen, extensions, texture_mapping,
-                          Elements(texture_mapping), PIPE_TEXTURE_2D,
+                          ARRAY_SIZE(texture_mapping), PIPE_TEXTURE_2D,
                           PIPE_BIND_SAMPLER_VIEW);
    init_format_extensions(screen, extensions, vertex_mapping,
-                          Elements(vertex_mapping), PIPE_BUFFER,
+                          ARRAY_SIZE(vertex_mapping), PIPE_BUFFER,
                           PIPE_BIND_VERTEX_BUFFER);
 
    /* Figure out GLSL support. */
@@ -758,22 +758,22 @@ void st_init_extensions(struct pipe_screen *screen,
       };
 
       consts->MaxSamples =
-         get_max_samples_for_formats(screen, Elements(color_formats),
+         get_max_samples_for_formats(screen, ARRAY_SIZE(color_formats),
                                      color_formats, 16,
                                      PIPE_BIND_RENDER_TARGET);
 
       consts->MaxColorTextureSamples =
-         get_max_samples_for_formats(screen, Elements(color_formats),
+         get_max_samples_for_formats(screen, ARRAY_SIZE(color_formats),
                                      color_formats, consts->MaxSamples,
                                      PIPE_BIND_SAMPLER_VIEW);
 
       consts->MaxDepthTextureSamples =
-         get_max_samples_for_formats(screen, Elements(depth_formats),
+         get_max_samples_for_formats(screen, ARRAY_SIZE(depth_formats),
                                      depth_formats, consts->MaxSamples,
                                      PIPE_BIND_SAMPLER_VIEW);
 
       consts->MaxIntegerSamples =
-         get_max_samples_for_formats(screen, Elements(int_formats),
+         get_max_samples_for_formats(screen, ARRAY_SIZE(int_formats),
                                      int_formats, consts->MaxSamples,
                                      PIPE_BIND_SAMPLER_VIEW);
    }
@@ -831,7 +831,7 @@ void st_init_extensions(struct pipe_screen *screen,
          extensions->ARB_texture_buffer_range = GL_TRUE;
 
       init_format_extensions(screen, extensions, tbo_rgb32,
-                             Elements(tbo_rgb32), PIPE_BUFFER,
+                             ARRAY_SIZE(tbo_rgb32), PIPE_BUFFER,
                              PIPE_BIND_SAMPLER_VIEW);
    }
 
diff --git a/mesalib/src/mesa/state_tracker/st_format.c b/mesalib/src/mesa/state_tracker/st_format.c
index 7868bb501..72dbf3bd4 100644
--- a/mesalib/src/mesa/state_tracker/st_format.c
+++ b/mesalib/src/mesa/state_tracker/st_format.c
@@ -1855,7 +1855,7 @@ st_choose_format(struct st_context *st, GLenum internalFormat,
       return pf;
 
    /* search table for internalFormat */
-   for (i = 0; i < Elements(format_map); i++) {
+   for (i = 0; i < ARRAY_SIZE(format_map); i++) {
       const struct format_mapping *mapping = &format_map[i];
       for (j = 0; mapping->glFormats[j]; j++) {
          if (mapping->glFormats[j] == internalFormat) {
diff --git a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
index 0b3477161..bd191d864 100644
--- a/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
+++ b/mesalib/src/mesa/state_tracker/st_glsl_to_tgsi.cpp
@@ -1497,7 +1497,7 @@ void
 glsl_to_tgsi_visitor::visit(ir_expression *ir)
 {
    unsigned int operand;
-   st_src_reg op[Elements(ir->operands)];
+   st_src_reg op[ARRAY_SIZE(ir->operands)];
    st_src_reg result_src;
    st_dst_reg result_dst;
 
@@ -3515,7 +3515,7 @@ get_src_arg_mask(st_dst_reg dst, st_src_reg src)
     */
    for (comp = 0; comp < 4; ++comp) {
       const unsigned coord = GET_SWZ(src.swizzle, comp);
-      ASSERT(coord < 4);
+      assert(coord < 4);
       if (dst.writemask & (1 << comp) && coord <= SWIZZLE_W)
          read_mask |= 1 << coord;
    }
@@ -4022,7 +4022,7 @@ glsl_to_tgsi_visitor::eliminate_dead_code(void)
          /* Continuing the block, clear any channels from the write array that
           * are read by this instruction.
           */
-         for (unsigned i = 0; i < Elements(inst->src); i++) {
+         for (unsigned i = 0; i < ARRAY_SIZE(inst->src); i++) {
             if (inst->src[i].file == PROGRAM_TEMPORARY && inst->src[i].reladdr){
                /* Any temporary might be read, so no dead code elimination
                 * across this instruction.
@@ -4067,7 +4067,7 @@ glsl_to_tgsi_visitor::eliminate_dead_code(void)
        * If there is already an instruction in the write array for one or more
        * of the channels, flag that channel write as dead.
        */
-      for (unsigned i = 0; i < Elements(inst->dst); i++) {
+      for (unsigned i = 0; i < ARRAY_SIZE(inst->dst); i++) {
          if (inst->dst[i].file == PROGRAM_TEMPORARY &&
              !inst->dst[i].reladdr &&
              !inst->saturate) {
@@ -4625,7 +4625,7 @@ dst_register(struct st_translate *t,
    case PROGRAM_ARRAY:
       array = index >> 16;
 
-      assert(array < Elements(t->arrays));
+      assert(array < ARRAY_SIZE(t->arrays));
 
       if (ureg_dst_is_undef(t->arrays[array]))
          t->arrays[array] = ureg_DECL_array_temporary(
@@ -4642,7 +4642,7 @@ dst_register(struct st_translate *t,
       else
          assert(index < VARYING_SLOT_MAX);
 
-      assert(t->outputMapping[index] < Elements(t->outputs));
+      assert(t->outputMapping[index] < ARRAY_SIZE(t->outputs));
 
       return t->outputs[t->outputMapping[index]];
 
@@ -4659,7 +4659,7 @@ dst_register(struct st_translate *t,
  * Map a glsl_to_tgsi src register to a TGSI ureg_src register.
  */
 static struct ureg_src
-src_register(struct st_translate *t, const struct st_src_reg *reg)
+src_register(struct st_translate *t, const st_src_reg *reg)
 {
    switch(reg->file) {
    case PROGRAM_UNDEFINED:
@@ -4685,18 +4685,18 @@ src_register(struct st_translate *t, const struct st_src_reg *reg)
       return t->immediates[reg->index];
 
    case PROGRAM_INPUT:
-      assert(t->inputMapping[reg->index] < Elements(t->inputs));
+      assert(t->inputMapping[reg->index] < ARRAY_SIZE(t->inputs));
       return t->inputs[t->inputMapping[reg->index]];
 
    case PROGRAM_OUTPUT:
-      assert(t->outputMapping[reg->index] < Elements(t->outputs));
+      assert(t->outputMapping[reg->index] < ARRAY_SIZE(t->outputs));
       return ureg_src(t->outputs[t->outputMapping[reg->index]]); /* not needed? */
 
    case PROGRAM_ADDRESS:
       return ureg_src(t->address[reg->index]);
 
    case PROGRAM_SYSTEM_VALUE:
-      assert(reg->index < (int) Elements(t->systemValues));
+      assert(reg->index < (int) ARRAY_SIZE(t->systemValues));
       return t->systemValues[reg->index];
 
    default:
@@ -4824,7 +4824,7 @@ translate_tex_offset(struct st_translate *t,
       array = in_offset->index >> 16;
 
       assert(array >= 0);
-      assert(array < (int) Elements(t->arrays));
+      assert(array < (int) ARRAY_SIZE(t->arrays));
 
       dst = t->arrays[array];
       offset.File = dst.File;
@@ -5200,8 +5200,8 @@ st_translate_program(
    unsigned i;
    enum pipe_error ret = PIPE_OK;
 
-   assert(numInputs <= Elements(t->inputs));
-   assert(numOutputs <= Elements(t->outputs));
+   assert(numInputs <= ARRAY_SIZE(t->inputs));
+   assert(numOutputs <= ARRAY_SIZE(t->outputs));
 
    assert(_mesa_sysval_to_semantic[SYSTEM_VALUE_FRONT_FACE] ==
           TGSI_SEMANTIC_FACE);
diff --git a/mesalib/src/mesa/state_tracker/st_manager.c b/mesalib/src/mesa/state_tracker/st_manager.c
index 5411d84b0..840f76a13 100644
--- a/mesalib/src/mesa/state_tracker/st_manager.c
+++ b/mesalib/src/mesa/state_tracker/st_manager.c
@@ -61,7 +61,7 @@
  * We'll only return non-null for window system framebuffers.
  * Note that this function may fail.
  */
-static INLINE struct st_framebuffer *
+static inline struct st_framebuffer *
 st_ws_framebuffer(struct gl_framebuffer *fb)
 {
    /* FBO cannot be casted.  See st_new_framebuffer */
@@ -73,7 +73,7 @@ st_ws_framebuffer(struct gl_framebuffer *fb)
 /**
  * Map an attachment to a buffer index.
  */
-static INLINE gl_buffer_index
+static inline gl_buffer_index
 attachment_to_buffer_index(enum st_attachment_type statt)
 {
    gl_buffer_index index;
@@ -109,7 +109,7 @@ attachment_to_buffer_index(enum st_attachment_type statt)
 /**
  * Map a buffer index to an attachment.
  */
-static INLINE enum st_attachment_type
+static inline enum st_attachment_type
 buffer_index_to_attachment(gl_buffer_index index)
 {
    enum st_attachment_type statt;
diff --git a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
index 3dd8a14b6..2f1016110 100644
--- a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
+++ b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.c
@@ -172,7 +172,7 @@ dst_register( struct st_translate *t,
       else
          assert(index < VARYING_SLOT_MAX);
 
-      assert(t->outputMapping[index] < Elements(t->outputs));
+      assert(t->outputMapping[index] < ARRAY_SIZE(t->outputs));
 
       return t->outputs[t->outputMapping[index]];
 
@@ -200,7 +200,7 @@ src_register( struct st_translate *t,
 
    case PROGRAM_TEMPORARY:
       assert(index >= 0);
-      assert(index < Elements(t->temps));
+      assert(index < ARRAY_SIZE(t->temps));
       if (ureg_dst_is_undef(t->temps[index]))
          t->temps[index] = ureg_DECL_temporary( t->ureg );
       return ureg_src(t->temps[index]);
@@ -216,18 +216,18 @@ src_register( struct st_translate *t,
          return t->constants[index];
 
    case PROGRAM_INPUT:
-      assert(t->inputMapping[index] < Elements(t->inputs));
+      assert(t->inputMapping[index] < ARRAY_SIZE(t->inputs));
       return t->inputs[t->inputMapping[index]];
 
    case PROGRAM_OUTPUT:
-      assert(t->outputMapping[index] < Elements(t->outputs));
+      assert(t->outputMapping[index] < ARRAY_SIZE(t->outputs));
       return ureg_src(t->outputs[t->outputMapping[index]]); /* not needed? */
 
    case PROGRAM_ADDRESS:
       return ureg_src(t->address[index]);
 
    case PROGRAM_SYSTEM_VALUE:
-      assert(index < Elements(t->systemValues));
+      assert(index < ARRAY_SIZE(t->systemValues));
       return t->systemValues[index];
 
    default:
@@ -1027,8 +1027,8 @@ st_translate_mesa_program(
    unsigned i;
    enum pipe_error ret = PIPE_OK;
 
-   assert(numInputs <= Elements(t->inputs));
-   assert(numOutputs <= Elements(t->outputs));
+   assert(numInputs <= ARRAY_SIZE(t->inputs));
+   assert(numOutputs <= ARRAY_SIZE(t->outputs));
 
    t = &translate;
    memset(t, 0, sizeof *t);
@@ -1254,14 +1254,3 @@ out:
 
    return ret;
 }
-
-
-/**
- * Tokens cannot be free with free otherwise the builtin gallium
- * malloc debugging will get confused.
- */
-void
-st_free_tokens(const struct tgsi_token *tokens)
-{
-   ureg_free_tokens(tokens);
-}
diff --git a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.h b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.h
index 1aa797a32..62bb654e9 100644
--- a/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.h
+++ b/mesalib/src/mesa/state_tracker/st_mesa_to_tgsi.h
@@ -62,9 +62,6 @@ st_translate_mesa_program(
    boolean passthrough_edgeflags,
    boolean clamp_color);
 
-void
-st_free_tokens(const struct tgsi_token *tokens);
-
 unsigned
 st_translate_texture_target(GLuint textarget, GLboolean shadow);
 
diff --git a/mesalib/src/mesa/state_tracker/st_program.c b/mesalib/src/mesa/state_tracker/st_program.c
index 10a5f2900..4cfd817ce 100644
--- a/mesalib/src/mesa/state_tracker/st_program.c
+++ b/mesalib/src/mesa/state_tracker/st_program.c
@@ -69,7 +69,7 @@ delete_vp_variant(struct st_context *st, struct st_vp_variant *vpv)
       draw_delete_vertex_shader( st->draw, vpv->draw_shader );
       
    if (vpv->tgsi.tokens)
-      st_free_tokens(vpv->tgsi.tokens);
+      ureg_free_tokens(vpv->tgsi.tokens);
       
    free( vpv );
 }
@@ -108,7 +108,7 @@ delete_fp_variant(struct st_context *st, struct st_fp_variant *fpv)
    if (fpv->parameters)
       _mesa_free_parameter_list(fpv->parameters);
    if (fpv->tgsi.tokens)
-      st_free_tokens(fpv->tgsi.tokens);
+      ureg_free_tokens(fpv->tgsi.tokens);
    free(fpv);
 }
 
@@ -869,106 +869,79 @@ st_translate_geometry_program(struct st_context *st,
    GLuint outputMapping[VARYING_SLOT_MAX];
    struct pipe_context *pipe = st->pipe;
    GLuint attr;
-   GLbitfield64 inputsRead;
-   GLuint vslot = 0;
 
    uint gs_num_inputs = 0;
-   uint gs_builtin_inputs = 0;
-   uint gs_array_offset = 0;
+
+   ubyte input_semantic_name[PIPE_MAX_SHADER_INPUTS];
+   ubyte input_semantic_index[PIPE_MAX_SHADER_INPUTS];
 
    ubyte gs_output_semantic_name[PIPE_MAX_SHADER_OUTPUTS];
    ubyte gs_output_semantic_index[PIPE_MAX_SHADER_OUTPUTS];
    uint gs_num_outputs = 0;
 
    GLint i;
-   GLuint maxSlot = 0;
    struct ureg_program *ureg;
-
+   struct pipe_shader_state state = {0};
    struct st_gp_variant *gpv;
 
    gpv = CALLOC_STRUCT(st_gp_variant);
    if (!gpv)
       return NULL;
 
-   if (!stgp->glsl_to_tgsi) {
-      _mesa_remove_output_reads(&stgp->Base.Base, PROGRAM_OUTPUT);
-   }
-
-   ureg = ureg_create( TGSI_PROCESSOR_GEOMETRY );
+   ureg = ureg_create(TGSI_PROCESSOR_GEOMETRY);
    if (ureg == NULL) {
       free(gpv);
       return NULL;
    }
 
-   /* which vertex output goes to the first geometry input */
-   vslot = 0;
-
    memset(inputMapping, 0, sizeof(inputMapping));
    memset(outputMapping, 0, sizeof(outputMapping));
 
    /*
     * Convert Mesa program inputs to TGSI input register semantics.
     */
-   inputsRead = stgp->Base.Base.InputsRead;
    for (attr = 0; attr < VARYING_SLOT_MAX; attr++) {
-      if ((inputsRead & BITFIELD64_BIT(attr)) != 0) {
-         const GLuint slot = gs_num_inputs;
-
-         gs_num_inputs++;
+      if ((stgp->Base.Base.InputsRead & BITFIELD64_BIT(attr)) != 0) {
+         const GLuint slot = gs_num_inputs++;
 
          inputMapping[attr] = slot;
 
-         stgp->input_map[slot + gs_array_offset] = vslot - gs_builtin_inputs;
-         stgp->input_to_index[attr] = vslot;
-         stgp->index_to_input[vslot] = attr;
-         ++vslot;
-
-         if (attr != VARYING_SLOT_PRIMITIVE_ID) {
-            gs_array_offset += 2;
-         } else
-            ++gs_builtin_inputs;
-
-#if 0
-         debug_printf("input map at %d = %d\n",
-                      slot + gs_array_offset, stgp->input_map[slot + gs_array_offset]);
-#endif
-
          switch (attr) {
          case VARYING_SLOT_PRIMITIVE_ID:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_PRIMID;
-            stgp->input_semantic_index[slot] = 0;
+            input_semantic_name[slot] = TGSI_SEMANTIC_PRIMID;
+            input_semantic_index[slot] = 0;
             break;
          case VARYING_SLOT_POS:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_POSITION;
-            stgp->input_semantic_index[slot] = 0;
+            input_semantic_name[slot] = TGSI_SEMANTIC_POSITION;
+            input_semantic_index[slot] = 0;
             break;
          case VARYING_SLOT_COL0:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
-            stgp->input_semantic_index[slot] = 0;
+            input_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
+            input_semantic_index[slot] = 0;
             break;
          case VARYING_SLOT_COL1:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
-            stgp->input_semantic_index[slot] = 1;
+            input_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
+            input_semantic_index[slot] = 1;
             break;
          case VARYING_SLOT_FOGC:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_FOG;
-            stgp->input_semantic_index[slot] = 0;
+            input_semantic_name[slot] = TGSI_SEMANTIC_FOG;
+            input_semantic_index[slot] = 0;
             break;
          case VARYING_SLOT_CLIP_VERTEX:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_CLIPVERTEX;
-            stgp->input_semantic_index[slot] = 0;
+            input_semantic_name[slot] = TGSI_SEMANTIC_CLIPVERTEX;
+            input_semantic_index[slot] = 0;
             break;
          case VARYING_SLOT_CLIP_DIST0:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
-            stgp->input_semantic_index[slot] = 0;
+            input_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
+            input_semantic_index[slot] = 0;
             break;
          case VARYING_SLOT_CLIP_DIST1:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
-            stgp->input_semantic_index[slot] = 1;
+            input_semantic_name[slot] = TGSI_SEMANTIC_CLIPDIST;
+            input_semantic_index[slot] = 1;
             break;
          case VARYING_SLOT_PSIZ:
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_PSIZE;
-            stgp->input_semantic_index[slot] = 0;
+            input_semantic_name[slot] = TGSI_SEMANTIC_PSIZE;
+            input_semantic_index[slot] = 0;
             break;
          case VARYING_SLOT_TEX0:
          case VARYING_SLOT_TEX1:
@@ -979,16 +952,16 @@ st_translate_geometry_program(struct st_context *st,
          case VARYING_SLOT_TEX6:
          case VARYING_SLOT_TEX7:
             if (st->needs_texcoord_semantic) {
-               stgp->input_semantic_name[slot] = TGSI_SEMANTIC_TEXCOORD;
-               stgp->input_semantic_index[slot] = attr - VARYING_SLOT_TEX0;
+               input_semantic_name[slot] = TGSI_SEMANTIC_TEXCOORD;
+               input_semantic_index[slot] = attr - VARYING_SLOT_TEX0;
                break;
             }
             /* fall through */
          case VARYING_SLOT_VAR0:
          default:
             assert(attr >= VARYING_SLOT_VAR0 && attr < VARYING_SLOT_MAX);
-            stgp->input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            stgp->input_semantic_index[slot] =
+            input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
+            input_semantic_index[slot] =
                st_get_generic_varying_index(st, attr);
          break;
          }
@@ -1007,10 +980,8 @@ st_translate_geometry_program(struct st_context *st,
     */
    for (attr = 0; attr < VARYING_SLOT_MAX; attr++) {
       if (stgp->Base.Base.OutputsWritten & BITFIELD64_BIT(attr)) {
-         GLuint slot;
+         GLuint slot = gs_num_outputs++;
 
-         slot = gs_num_outputs;
-         gs_num_outputs++;
          outputMapping[attr] = slot;
 
          switch (attr) {
@@ -1083,7 +1054,7 @@ st_translate_geometry_program(struct st_context *st,
             /* fall through */
          case VARYING_SLOT_VAR0:
          default:
-            assert(slot < Elements(gs_output_semantic_name));
+            assert(slot < ARRAY_SIZE(gs_output_semantic_name));
             assert(attr >= VARYING_SLOT_VAR0);
             gs_output_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
             gs_output_semantic_index[slot] =
@@ -1093,93 +1064,38 @@ st_translate_geometry_program(struct st_context *st,
       }
    }
 
-   /* find max output slot referenced to compute gs_num_outputs */
-   for (attr = 0; attr < VARYING_SLOT_MAX; attr++) {
-      if (outputMapping[attr] != ~0U && outputMapping[attr] > maxSlot)
-         maxSlot = outputMapping[attr];
-   }
-   gs_num_outputs = maxSlot + 1;
-
-#if 0 /* debug */
-   {
-      GLuint i;
-      printf("outputMapping? %d\n", outputMapping ? 1 : 0);
-      if (outputMapping) {
-         printf("attr -> slot\n");
-         for (i = 0; i < 16;  i++) {
-            printf(" %2d       %3d\n", i, outputMapping[i]);
-         }
-      }
-      printf("slot    sem_name  sem_index\n");
-      for (i = 0; i < gs_num_outputs; i++) {
-         printf(" %2d         %d         %d\n",
-                i,
-                gs_output_semantic_name[i],
-                gs_output_semantic_index[i]);
-      }
-   }
-#endif
-
-   /* free old shader state, if any */
-   if (stgp->tgsi.tokens) {
-      st_free_tokens(stgp->tgsi.tokens);
-      stgp->tgsi.tokens = NULL;
-   }
-
    ureg_property(ureg, TGSI_PROPERTY_GS_INPUT_PRIM, stgp->Base.InputType);
    ureg_property(ureg, TGSI_PROPERTY_GS_OUTPUT_PRIM, stgp->Base.OutputType);
    ureg_property(ureg, TGSI_PROPERTY_GS_MAX_OUTPUT_VERTICES,
                  stgp->Base.VerticesOut);
    ureg_property(ureg, TGSI_PROPERTY_GS_INVOCATIONS, stgp->Base.Invocations);
 
-   if (stgp->glsl_to_tgsi)
-      st_translate_program(st->ctx,
-                           TGSI_PROCESSOR_GEOMETRY,
-                           ureg,
-                           stgp->glsl_to_tgsi,
-                           &stgp->Base.Base,
-                           /* inputs */
-                           gs_num_inputs,
-                           inputMapping,
-                           stgp->input_semantic_name,
-                           stgp->input_semantic_index,
-                           NULL,
-                           NULL,
-                           /* outputs */
-                           gs_num_outputs,
-                           outputMapping,
-                           gs_output_semantic_name,
-                           gs_output_semantic_index,
-                           FALSE,
-                           FALSE);
-   else
-      st_translate_mesa_program(st->ctx,
-                                TGSI_PROCESSOR_GEOMETRY,
-                                ureg,
-                                &stgp->Base.Base,
-                                /* inputs */
-                                gs_num_inputs,
-                                inputMapping,
-                                stgp->input_semantic_name,
-                                stgp->input_semantic_index,
-                                NULL,
-                                /* outputs */
-                                gs_num_outputs,
-                                outputMapping,
-                                gs_output_semantic_name,
-                                gs_output_semantic_index,
-                                FALSE,
-                                FALSE);
-
-   stgp->num_inputs = gs_num_inputs;
-   stgp->tgsi.tokens = ureg_get_tokens( ureg, NULL );
-   ureg_destroy( ureg );
-
-   if (stgp->glsl_to_tgsi) {
-      st_translate_stream_output_info(stgp->glsl_to_tgsi,
-                                      outputMapping,
-                                      &stgp->tgsi.stream_output);
-   }
+   st_translate_program(st->ctx,
+                        TGSI_PROCESSOR_GEOMETRY,
+                        ureg,
+                        stgp->glsl_to_tgsi,
+                        &stgp->Base.Base,
+                        /* inputs */
+                        gs_num_inputs,
+                        inputMapping,
+                        input_semantic_name,
+                        input_semantic_index,
+                        NULL,
+                        NULL,
+                        /* outputs */
+                        gs_num_outputs,
+                        outputMapping,
+                        gs_output_semantic_name,
+                        gs_output_semantic_index,
+                        FALSE,
+                        FALSE);
+
+   state.tokens = ureg_get_tokens(ureg, NULL);
+   ureg_destroy(ureg);
+
+   st_translate_stream_output_info(stgp->glsl_to_tgsi,
+                                   outputMapping,
+                                   &state.stream_output);
 
    if ((ST_DEBUG & DEBUG_TGSI) && (ST_DEBUG & DEBUG_MESA)) {
       _mesa_print_program(&stgp->Base.Base);
@@ -1187,13 +1103,15 @@ st_translate_geometry_program(struct st_context *st,
    }
 
    if (ST_DEBUG & DEBUG_TGSI) {
-      tgsi_dump(stgp->tgsi.tokens, 0);
+      tgsi_dump(state.tokens, 0);
       debug_printf("\n");
    }
 
    /* fill in new variant */
-   gpv->driver_shader = pipe->create_gs_state(pipe, &stgp->tgsi);
+   gpv->driver_shader = pipe->create_gs_state(pipe, &state);
    gpv->key = *key;
+
+   ureg_free_tokens(state.tokens);
    return gpv;
 }
 
@@ -1229,51 +1147,6 @@ st_get_gp_variant(struct st_context *st,
 }
 
 
-
-
-/**
- * Debug- print current shader text
- */
-void
-st_print_shaders(struct gl_context *ctx)
-{
-   struct gl_shader_program **shProg = ctx->_Shader->CurrentProgram;
-   unsigned j;
-
-   for (j = 0; j < 3; j++) {
-      unsigned i;
-
-      if (shProg[j] == NULL)
-	 continue;
-
-      for (i = 0; i < shProg[j]->NumShaders; i++) {
-	 struct gl_shader *sh;
-
-	 switch (shProg[j]->Shaders[i]->Type) {
-	 case GL_VERTEX_SHADER:
-	    sh = (i != 0) ? NULL : shProg[j]->Shaders[i];
-	    break;
-	 case GL_GEOMETRY_SHADER_ARB:
-	    sh = (i != 1) ? NULL : shProg[j]->Shaders[i];
-	    break;
-	 case GL_FRAGMENT_SHADER:
-	    sh = (i != 2) ? NULL : shProg[j]->Shaders[i];
-	    break;
-	 default:
-	    assert(0);
-	    sh = NULL;
-	    break;
-	 }
-
-	 if (sh != NULL) {
-	    printf("GLSL shader %u of %u:\n", i, shProg[j]->NumShaders);
-	    printf("%s\n", sh->Source);
-	 }
-      }
-   }
-}
-
-
 /**
  * Vert/Geom/Frag programs have per-context variants.  Free all the
  * variants attached to the given program which match the given context.
@@ -1374,7 +1247,7 @@ destroy_shader_program_variants_cb(GLuint key, void *data, void *userData)
             destroy_program_variants(st, shProg->Shaders[i]->Program);
          }
 
-	 for (i = 0; i < Elements(shProg->_LinkedShaders); i++) {
+	 for (i = 0; i < ARRAY_SIZE(shProg->_LinkedShaders); i++) {
 	    if (shProg->_LinkedShaders[i])
                destroy_program_variants(st, shProg->_LinkedShaders[i]->Program);
 	 }
diff --git a/mesalib/src/mesa/state_tracker/st_program.h b/mesalib/src/mesa/state_tracker/st_program.h
index 870d0d57a..451d7bb6a 100644
--- a/mesalib/src/mesa/state_tracker/st_program.h
+++ b/mesalib/src/mesa/state_tracker/st_program.h
@@ -203,49 +203,31 @@ struct st_geometry_program
    struct gl_geometry_program Base;  /**< The Mesa geometry program */
    struct glsl_to_tgsi_visitor* glsl_to_tgsi;
 
-   /** map GP input back to VP output */
-   GLuint input_map[PIPE_MAX_SHADER_INPUTS];
-
-   /** maps a Mesa VARYING_SLOT_x to a packed TGSI input index */
-   GLuint input_to_index[VARYING_SLOT_MAX];
-   /** maps a TGSI input index back to a Mesa VARYING_SLOT_x */
-   GLuint index_to_input[PIPE_MAX_SHADER_INPUTS];
-
-   GLuint num_inputs;
-
-   GLuint input_to_slot[VARYING_SLOT_MAX];  /**< Maps VARYING_SLOT_x to slot */
-   GLuint num_input_slots;
-
-   ubyte input_semantic_name[PIPE_MAX_SHADER_INPUTS];
-   ubyte input_semantic_index[PIPE_MAX_SHADER_INPUTS];
-
-   struct pipe_shader_state tgsi;
-
    struct st_gp_variant *variants;
 };
 
 
 
-static INLINE struct st_fragment_program *
+static inline struct st_fragment_program *
 st_fragment_program( struct gl_fragment_program *fp )
 {
    return (struct st_fragment_program *)fp;
 }
 
 
-static INLINE struct st_vertex_program *
+static inline struct st_vertex_program *
 st_vertex_program( struct gl_vertex_program *vp )
 {
    return (struct st_vertex_program *)vp;
 }
 
-static INLINE struct st_geometry_program *
+static inline struct st_geometry_program *
 st_geometry_program( struct gl_geometry_program *gp )
 {
    return (struct st_geometry_program *)gp;
 }
 
-static INLINE void
+static inline void
 st_reference_vertprog(struct st_context *st,
                       struct st_vertex_program **ptr,
                       struct st_vertex_program *prog)
@@ -255,7 +237,7 @@ st_reference_vertprog(struct st_context *st,
                            (struct gl_program *) prog);
 }
 
-static INLINE void
+static inline void
 st_reference_geomprog(struct st_context *st,
                       struct st_geometry_program **ptr,
                       struct st_geometry_program *prog)
@@ -265,7 +247,7 @@ st_reference_geomprog(struct st_context *st,
                            (struct gl_program *) prog);
 }
 
-static INLINE void
+static inline void
 st_reference_fragprog(struct st_context *st,
                       struct st_fragment_program **ptr,
                       struct st_fragment_program *prog)
@@ -278,7 +260,7 @@ st_reference_fragprog(struct st_context *st,
 /**
  * This defines mapping from Mesa VARYING_SLOTs to TGSI GENERIC slots.
  */
-static INLINE unsigned
+static inline unsigned
 st_get_generic_varying_index(struct st_context *st, GLuint attr)
 {
    if (attr >= VARYING_SLOT_VAR0) {
@@ -340,10 +322,6 @@ extern void
 st_release_gp_variants(struct st_context *st,
                        struct st_geometry_program *stgp);
 
-
-extern void
-st_print_shaders(struct gl_context *ctx);
-
 extern void
 st_destroy_program_variants(struct st_context *st);
 
diff --git a/mesalib/src/mesa/state_tracker/st_texture.h b/mesalib/src/mesa/state_tracker/st_texture.h
index 2f540295f..6b7f8c750 100644
--- a/mesalib/src/mesa/state_tracker/st_texture.h
+++ b/mesalib/src/mesa/state_tracker/st_texture.h
@@ -117,32 +117,32 @@ struct st_texture_object
 };
 
 
-static INLINE struct st_texture_image *
+static inline struct st_texture_image *
 st_texture_image(struct gl_texture_image *img)
 {
    return (struct st_texture_image *) img;
 }
 
-static INLINE const struct st_texture_image *
+static inline const struct st_texture_image *
 st_texture_image_const(const struct gl_texture_image *img)
 {
    return (const struct st_texture_image *) img;
 }
 
-static INLINE struct st_texture_object *
+static inline struct st_texture_object *
 st_texture_object(struct gl_texture_object *obj)
 {
    return (struct st_texture_object *) obj;
 }
 
-static INLINE const struct st_texture_object *
+static inline const struct st_texture_object *
 st_texture_object_const(const struct gl_texture_object *obj)
 {
    return (const struct st_texture_object *) obj;
 }
 
 
-static INLINE struct pipe_resource *
+static inline struct pipe_resource *
 st_get_texobj_resource(struct gl_texture_object *texObj)
 {
    struct st_texture_object *stObj = st_texture_object(texObj);
@@ -150,14 +150,14 @@ st_get_texobj_resource(struct gl_texture_object *texObj)
 }
 
 
-static INLINE struct pipe_resource *
+static inline struct pipe_resource *
 st_get_stobj_resource(struct st_texture_object *stObj)
 {
    return stObj ? stObj->pt : NULL;
 }
 
 
-static INLINE struct pipe_sampler_view *
+static inline struct pipe_sampler_view *
 st_create_texture_sampler_view_format(struct pipe_context *pipe,
                                       struct pipe_resource *texture,
                                       enum pipe_format format)
@@ -169,7 +169,7 @@ st_create_texture_sampler_view_format(struct pipe_context *pipe,
    return pipe->create_sampler_view(pipe, texture, &templ);
 }
 
-static INLINE struct pipe_sampler_view *
+static inline struct pipe_sampler_view *
 st_create_texture_sampler_view(struct pipe_context *pipe,
                                struct pipe_resource *texture)
 {
diff --git a/mesalib/src/mesa/swrast/s_aaline.c b/mesalib/src/mesa/swrast/s_aaline.c
index 6aea9d545..f3258e813 100644
--- a/mesalib/src/mesa/swrast/s_aaline.c
+++ b/mesalib/src/mesa/swrast/s_aaline.c
@@ -23,6 +23,7 @@
  */
 
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/imports.h"
 #include "main/macros.h"
@@ -203,7 +204,7 @@ compute_lambda(const GLfloat sPlane[4], const GLfloat tPlane[4],
    if (rho2 == 0.0F)
       return 0.0;
    else
-      return (GLfloat) (LOGF(rho2) * 1.442695 * 0.5);/* 1.442695 = 1/log(2) */
+      return logf(rho2) * 1.442695f * 0.5f;/* 1.442695 = 1/log(2) */
 }
 
 
@@ -477,7 +478,7 @@ _swrast_choose_aa_line_function(struct gl_context *ctx)
 {
    SWcontext *swrast = SWRAST_CONTEXT(ctx);
 
-   ASSERT(ctx->Line.SmoothFlag);
+   assert(ctx->Line.SmoothFlag);
 
    if (ctx->Texture._EnabledCoordUnits != 0
        || _swrast_use_fragment_program(ctx)
diff --git a/mesalib/src/mesa/swrast/s_aatriangle.c b/mesalib/src/mesa/swrast/s_aatriangle.c
index 219282748..1d076cc7d 100644
--- a/mesalib/src/mesa/swrast/s_aatriangle.c
+++ b/mesalib/src/mesa/swrast/s_aatriangle.c
@@ -102,7 +102,7 @@ do {					\
 static inline GLfloat
 solve_plane(GLfloat x, GLfloat y, const GLfloat plane[4])
 {
-   ASSERT(plane[2] != 0.0F);
+   assert(plane[2] != 0.0F);
    return (plane[3] + plane[0] * x + plane[1] * y) / -plane[2];
 }
 
@@ -201,7 +201,7 @@ compute_coveragef(const GLfloat v0[3], const GLfloat v1[3],
    GLint stop = 4, i;
    GLfloat insideCount = 16.0F;
 
-   ASSERT(dx0 * dy1 - dx1 * dy0 >= 0.0); /* area >= 0.0 */
+   assert(dx0 * dy1 - dx1 * dy0 >= 0.0); /* area >= 0.0 */
 
    for (i = 0; i < stop; i++) {
       const GLfloat sx = x + samples[i][0];
@@ -282,7 +282,7 @@ _swrast_set_aa_triangle_function(struct gl_context *ctx)
 {
    SWcontext *swrast = SWRAST_CONTEXT(ctx);
 
-   ASSERT(ctx->Polygon.SmoothFlag);
+   assert(ctx->Polygon.SmoothFlag);
 
    if (ctx->Texture._EnabledCoordUnits != 0
        || _swrast_use_fragment_program(ctx)
@@ -294,5 +294,5 @@ _swrast_set_aa_triangle_function(struct gl_context *ctx)
       SWRAST_CONTEXT(ctx)->Triangle = rgba_aa_tri;
    }
 
-   ASSERT(SWRAST_CONTEXT(ctx)->Triangle);
+   assert(SWRAST_CONTEXT(ctx)->Triangle);
 }
diff --git a/mesalib/src/mesa/swrast/s_aatritemp.h b/mesalib/src/mesa/swrast/s_aatritemp.h
index fd374a524..230dab816 100644
--- a/mesalib/src/mesa/swrast/s_aatritemp.h
+++ b/mesalib/src/mesa/swrast/s_aatritemp.h
@@ -284,7 +284,7 @@
             /* (cx,cy) = center of fragment */
             const GLfloat cx = ix + 0.5F, cy = iy + 0.5F;
             SWspanarrays *array = span.array;
-            ASSERT(ix >= 0);
+            assert(ix >= 0);
             array->coverage[ix] = coverage;
 #ifdef DO_Z
             array->z[ix] = (GLuint) solve_plane(cx, cy, zPlane);
diff --git a/mesalib/src/mesa/swrast/s_alpha.c b/mesalib/src/mesa/swrast/s_alpha.c
index fba143a52..b1a7ff132 100644
--- a/mesalib/src/mesa/swrast/s_alpha.c
+++ b/mesalib/src/mesa/swrast/s_alpha.c
@@ -130,7 +130,7 @@ _swrast_alpha_test(const struct gl_context *ctx, SWspan *span)
    }
    else {
       /* Interpolate alpha values */
-      ASSERT(span->interpMask & SPAN_RGBA);
+      assert(span->interpMask & SPAN_RGBA);
       if (span->array->ChanType == GL_UNSIGNED_BYTE) {
          const GLfixed alphaStep = span->alphaStep;
          GLfixed alpha = span->alpha;
diff --git a/mesalib/src/mesa/swrast/s_atifragshader.c b/mesalib/src/mesa/swrast/s_atifragshader.c
index 1e91e2bee..0bf03771f 100644
--- a/mesalib/src/mesa/swrast/s_atifragshader.c
+++ b/mesalib/src/mesa/swrast/s_atifragshader.c
@@ -571,7 +571,7 @@ _swrast_exec_fragment_shader(struct gl_context * ctx, SWspan *span)
    GLuint i;
 
    /* incoming colors should be floats */
-   ASSERT(span->array->ChanType == GL_FLOAT);
+   assert(span->array->ChanType == GL_FLOAT);
 
    for (i = 0; i < span->end; i++) {
       if (span->array->mask[i]) {
diff --git a/mesalib/src/mesa/swrast/s_bitmap.c b/mesalib/src/mesa/swrast/s_bitmap.c
index e364ec129..324daea36 100644
--- a/mesalib/src/mesa/swrast/s_bitmap.c
+++ b/mesalib/src/mesa/swrast/s_bitmap.c
@@ -55,7 +55,7 @@ _swrast_Bitmap( struct gl_context *ctx, GLint px, GLint py,
    GLuint count = 0;
    SWspan span;
 
-   ASSERT(ctx->RenderMode == GL_RENDER);
+   assert(ctx->RenderMode == GL_RENDER);
 
    if (!_mesa_check_conditional_render(ctx))
       return; /* don't draw */
@@ -154,8 +154,8 @@ _swrast_Bitmap( struct gl_context *ctx, GLint px, GLint py,
    GLint row, col;
    SWspan span;
 
-   ASSERT(ctx->RenderMode == GL_RENDER);
-   ASSERT(bitmap);
+   assert(ctx->RenderMode == GL_RENDER);
+   assert(bitmap);
 
    swrast_render_start(ctx);
 
diff --git a/mesalib/src/mesa/swrast/s_blend.c b/mesalib/src/mesa/swrast/s_blend.c
index 1037b6236..7cb119407 100644
--- a/mesalib/src/mesa/swrast/s_blend.c
+++ b/mesalib/src/mesa/swrast/s_blend.c
@@ -75,10 +75,10 @@ blend_noop(struct gl_context *ctx, GLuint n, const GLubyte mask[],
 {
    GLint bytes;
 
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].SrcRGB == GL_ZERO);
-   ASSERT(ctx->Color.Blend[0].DstRGB == GL_ONE);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].SrcRGB == GL_ZERO);
+   assert(ctx->Color.Blend[0].DstRGB == GL_ONE);
    (void) ctx;
 
    /* just memcpy */
@@ -101,10 +101,10 @@ static void _BLENDAPI
 blend_replace(struct gl_context *ctx, GLuint n, const GLubyte mask[],
               GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].SrcRGB == GL_ONE);
-   ASSERT(ctx->Color.Blend[0].DstRGB == GL_ZERO);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].SrcRGB == GL_ONE);
+   assert(ctx->Color.Blend[0].DstRGB == GL_ZERO);
    (void) ctx;
    (void) n;
    (void) mask;
@@ -125,13 +125,13 @@ blend_transparency_ubyte(struct gl_context *ctx, GLuint n, const GLubyte mask[],
    const GLubyte (*dest)[4] = (const GLubyte (*)[4]) dst;
    GLuint i;
 
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].SrcRGB == GL_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].SrcA == GL_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].DstRGB == GL_ONE_MINUS_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].DstA == GL_ONE_MINUS_SRC_ALPHA);
-   ASSERT(chanType == GL_UNSIGNED_BYTE);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].SrcRGB == GL_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].SrcA == GL_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].DstRGB == GL_ONE_MINUS_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].DstA == GL_ONE_MINUS_SRC_ALPHA);
+   assert(chanType == GL_UNSIGNED_BYTE);
 
    (void) ctx;
 
@@ -148,10 +148,10 @@ blend_transparency_ubyte(struct gl_context *ctx, GLuint n, const GLubyte mask[],
             const GLint g = DIV255((rgba[i][GCOMP] - dest[i][GCOMP]) * t) + dest[i][GCOMP];
             const GLint b = DIV255((rgba[i][BCOMP] - dest[i][BCOMP]) * t) + dest[i][BCOMP];
             const GLint a = DIV255((rgba[i][ACOMP] - dest[i][ACOMP]) * t) + dest[i][ACOMP]; 
-            ASSERT(r <= 255);
-            ASSERT(g <= 255);
-            ASSERT(b <= 255);
-            ASSERT(a <= 255);
+            assert(r <= 255);
+            assert(g <= 255);
+            assert(b <= 255);
+            assert(a <= 255);
             rgba[i][RCOMP] = (GLubyte) r;
             rgba[i][GCOMP] = (GLubyte) g;
             rgba[i][BCOMP] = (GLubyte) b;
@@ -170,13 +170,13 @@ blend_transparency_ushort(struct gl_context *ctx, GLuint n, const GLubyte mask[]
    const GLushort (*dest)[4] = (const GLushort (*)[4]) dst;
    GLuint i;
 
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].SrcRGB == GL_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].SrcA == GL_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].DstRGB == GL_ONE_MINUS_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].DstA == GL_ONE_MINUS_SRC_ALPHA);
-   ASSERT(chanType == GL_UNSIGNED_SHORT);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].SrcRGB == GL_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].SrcA == GL_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].DstRGB == GL_ONE_MINUS_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].DstA == GL_ONE_MINUS_SRC_ALPHA);
+   assert(chanType == GL_UNSIGNED_SHORT);
 
    (void) ctx;
 
@@ -208,13 +208,13 @@ blend_transparency_float(struct gl_context *ctx, GLuint n, const GLubyte mask[],
    const GLfloat (*dest)[4] = (const GLfloat (*)[4]) dst;
    GLuint i;
 
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].SrcRGB == GL_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].SrcA == GL_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].DstRGB == GL_ONE_MINUS_SRC_ALPHA);
-   ASSERT(ctx->Color.Blend[0].DstA == GL_ONE_MINUS_SRC_ALPHA);
-   ASSERT(chanType == GL_FLOAT);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].SrcRGB == GL_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].SrcA == GL_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].DstRGB == GL_ONE_MINUS_SRC_ALPHA);
+   assert(ctx->Color.Blend[0].DstA == GL_ONE_MINUS_SRC_ALPHA);
+   assert(chanType == GL_FLOAT);
 
    (void) ctx;
 
@@ -248,10 +248,10 @@ blend_add(struct gl_context *ctx, GLuint n, const GLubyte mask[],
 {
    GLuint i;
 
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
-   ASSERT(ctx->Color.Blend[0].SrcRGB == GL_ONE);
-   ASSERT(ctx->Color.Blend[0].DstRGB == GL_ONE);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].EquationA == GL_FUNC_ADD);
+   assert(ctx->Color.Blend[0].SrcRGB == GL_ONE);
+   assert(ctx->Color.Blend[0].DstRGB == GL_ONE);
    (void) ctx;
 
    if (chanType == GL_UNSIGNED_BYTE) {
@@ -289,7 +289,7 @@ blend_add(struct gl_context *ctx, GLuint n, const GLubyte mask[],
    else {
       GLfloat (*rgba)[4] = (GLfloat (*)[4]) src;
       const GLfloat (*dest)[4] = (const GLfloat (*)[4]) dst;
-      ASSERT(chanType == GL_FLOAT);
+      assert(chanType == GL_FLOAT);
       for (i=0;i<n;i++) {
          if (mask[i]) {
             /* don't RGB clamp to max */
@@ -313,8 +313,8 @@ blend_min(struct gl_context *ctx, GLuint n, const GLubyte mask[],
           GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
    GLuint i;
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_MIN);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_MIN);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_MIN);
+   assert(ctx->Color.Blend[0].EquationA == GL_MIN);
    (void) ctx;
 
    if (chanType == GL_UNSIGNED_BYTE) {
@@ -344,7 +344,7 @@ blend_min(struct gl_context *ctx, GLuint n, const GLubyte mask[],
    else {
       GLfloat (*rgba)[4] = (GLfloat (*)[4]) src;
       const GLfloat (*dest)[4] = (const GLfloat (*)[4]) dst;
-      ASSERT(chanType == GL_FLOAT);
+      assert(chanType == GL_FLOAT);
       for (i=0;i<n;i++) {
          if (mask[i]) {
             rgba[i][RCOMP] = MIN2( rgba[i][RCOMP], dest[i][RCOMP] );
@@ -366,8 +366,8 @@ blend_max(struct gl_context *ctx, GLuint n, const GLubyte mask[],
           GLvoid *src, const GLvoid *dst, GLenum chanType)
 {
    GLuint i;
-   ASSERT(ctx->Color.Blend[0].EquationRGB == GL_MAX);
-   ASSERT(ctx->Color.Blend[0].EquationA == GL_MAX);
+   assert(ctx->Color.Blend[0].EquationRGB == GL_MAX);
+   assert(ctx->Color.Blend[0].EquationA == GL_MAX);
    (void) ctx;
 
    if (chanType == GL_UNSIGNED_BYTE) {
@@ -397,7 +397,7 @@ blend_max(struct gl_context *ctx, GLuint n, const GLubyte mask[],
    else {
       GLfloat (*rgba)[4] = (GLfloat (*)[4]) src;
       const GLfloat (*dest)[4] = (const GLfloat (*)[4]) dst;
-      ASSERT(chanType == GL_FLOAT);
+      assert(chanType == GL_FLOAT);
       for (i=0;i<n;i++) {
          if (mask[i]) {
             rgba[i][RCOMP] = MAX2( rgba[i][RCOMP], dest[i][RCOMP] );
@@ -450,7 +450,7 @@ blend_modulate(struct gl_context *ctx, GLuint n, const GLubyte mask[],
    else {
       GLfloat (*rgba)[4] = (GLfloat (*)[4]) src;
       const GLfloat (*dest)[4] = (const GLfloat (*)[4]) dst;
-      ASSERT(chanType == GL_FLOAT);
+      assert(chanType == GL_FLOAT);
       for (i=0;i<n;i++) {
          if (mask[i]) {
             rgba[i][RCOMP] = rgba[i][RCOMP] * dest[i][RCOMP];
@@ -998,9 +998,9 @@ _swrast_blend_span(struct gl_context *ctx, struct gl_renderbuffer *rb, SWspan *s
    SWcontext *swrast = SWRAST_CONTEXT(ctx);
    void *rbPixels;
 
-   ASSERT(span->end <= SWRAST_MAX_WIDTH);
-   ASSERT(span->arrayMask & SPAN_RGBA);
-   ASSERT(!ctx->Color.ColorLogicOpEnabled);
+   assert(span->end <= SWRAST_MAX_WIDTH);
+   assert(span->arrayMask & SPAN_RGBA);
+   assert(!ctx->Color.ColorLogicOpEnabled);
 
    rbPixels = _swrast_get_dest_rgba(ctx, rb, span);
 
diff --git a/mesalib/src/mesa/swrast/s_blit.c b/mesalib/src/mesa/swrast/s_blit.c
index 16e5b8c1c..3e838a41d 100644
--- a/mesalib/src/mesa/swrast/s_blit.c
+++ b/mesalib/src/mesa/swrast/s_blit.c
@@ -52,8 +52,8 @@ NAME(GLint srcWidth, GLint dstWidth,			\
    if (flip) {						\
       for (dstCol = 0; dstCol < dstWidth; dstCol++) {	\
          GLint srcCol = (dstCol * srcWidth) / dstWidth;	\
-         ASSERT(srcCol >= 0);				\
-         ASSERT(srcCol < srcWidth);			\
+         assert(srcCol >= 0);				\
+         assert(srcCol < srcWidth);			\
          srcCol = srcWidth - 1 - srcCol; /* flip */	\
          if (SIZE == 1) {				\
             dst[dstCol] = src[srcCol];			\
@@ -73,8 +73,8 @@ NAME(GLint srcWidth, GLint dstWidth,			\
    else {						\
       for (dstCol = 0; dstCol < dstWidth; dstCol++) {	\
          GLint srcCol = (dstCol * srcWidth) / dstWidth;	\
-         ASSERT(srcCol >= 0);				\
-         ASSERT(srcCol < srcWidth);			\
+         assert(srcCol >= 0);				\
+         assert(srcCol < srcWidth);			\
          if (SIZE == 1) {				\
             dst[dstCol] = src[srcCol];			\
          }						\
@@ -299,8 +299,8 @@ blit_nearest(struct gl_context *ctx,
          GLint srcRow = IROUND(srcRowF);
          GLubyte *dstRowStart = dstMap + dstRowStride * dstRow;
 
-         ASSERT(srcRow >= 0);
-         ASSERT(srcRow < srcHeight);
+         assert(srcRow >= 0);
+         assert(srcRow < srcHeight);
 
          if (invertY) {
             srcRow = srcHeight - 1 - srcRow;
@@ -412,8 +412,8 @@ resample_linear_row_ub(GLint srcWidth, GLint dstWidth,
       GLfloat colWeight = srcCol - srcCol0; /* fractional part of srcCol */
       GLfloat red, green, blue, alpha;
 
-      ASSERT(srcCol0 < srcWidth);
-      ASSERT(srcCol1 <= srcWidth);
+      assert(srcCol0 < srcWidth);
+      assert(srcCol1 <= srcWidth);
 
       if (srcCol1 == srcWidth) {
          /* last column fudge */
@@ -467,8 +467,8 @@ resample_linear_row_float(GLint srcWidth, GLint dstWidth,
       GLfloat colWeight = srcCol - srcCol0; /* fractional part of srcCol */
       GLfloat red, green, blue, alpha;
 
-      ASSERT(srcCol0 < srcWidth);
-      ASSERT(srcCol1 <= srcWidth);
+      assert(srcCol0 < srcWidth);
+      assert(srcCol1 <= srcWidth);
 
       if (srcCol1 == srcWidth) {
          /* last column fudge */
@@ -801,7 +801,7 @@ _swrast_BlitFramebuffer(struct gl_context *ctx,
       }
    }
    else {
-      ASSERT(filter == GL_LINEAR);
+      assert(filter == GL_LINEAR);
       if (mask & GL_COLOR_BUFFER_BIT) {  /* depth/stencil not allowed */
          blit_linear(ctx, readFb, drawFb, srcX0, srcY0, srcX1, srcY1,
                      dstX0, dstY0, dstX1, dstY1);
diff --git a/mesalib/src/mesa/swrast/s_context.c b/mesalib/src/mesa/swrast/s_context.c
index ad11aa2c1..51cc22760 100644
--- a/mesalib/src/mesa/swrast/s_context.c
+++ b/mesalib/src/mesa/swrast/s_context.c
@@ -251,6 +251,7 @@ _swrast_update_fog_state( struct gl_context *ctx )
    const struct gl_fragment_program *fp = ctx->FragmentProgram._Current;
 
    assert(fp == NULL || fp->Base.Target == GL_FRAGMENT_PROGRAM_ARB);
+   (void) fp; /* silence unused var warning */
 
    /* determine if fog is needed, and if so, which fog mode */
    swrast->_FogEnabled = (!_swrast_use_fragment_program(ctx) &&
@@ -351,7 +352,7 @@ _swrast_validate_triangle( struct gl_context *ctx,
 
    _swrast_validate_derived( ctx );
    swrast->choose_triangle( ctx );
-   ASSERT(swrast->Triangle);
+   assert(swrast->Triangle);
 
    if (swrast->SpecularVertexAdd) {
       /* separate specular color, but no texture */
@@ -373,7 +374,7 @@ _swrast_validate_line( struct gl_context *ctx, const SWvertex *v0, const SWverte
 
    _swrast_validate_derived( ctx );
    swrast->choose_line( ctx );
-   ASSERT(swrast->Line);
+   assert(swrast->Line);
 
    if (swrast->SpecularVertexAdd) {
       swrast->SpecLine = swrast->Line;
diff --git a/mesalib/src/mesa/swrast/s_copypix.c b/mesalib/src/mesa/swrast/s_copypix.c
index e21c69dbc..17140ad2d 100644
--- a/mesalib/src/mesa/swrast/s_copypix.c
+++ b/mesalib/src/mesa/swrast/s_copypix.c
@@ -158,7 +158,7 @@ copy_rgba_pixels(struct gl_context *ctx, GLint srcx, GLint srcy,
       p = NULL;
    }
 
-   ASSERT(width < SWRAST_MAX_WIDTH);
+   assert(width < SWRAST_MAX_WIDTH);
 
    for (row = 0; row < height; row++, sy += stepy, dy += stepy) {
       GLvoid *rgba = span.array->attribs[VARYING_SLOT_COL0];
@@ -468,7 +468,7 @@ swrast_fast_copy_pixels(struct gl_context *ctx,
       dstRb = dstFb->Attachment[BUFFER_DEPTH].Renderbuffer;
    }
    else {
-      ASSERT(type == GL_DEPTH_STENCIL_EXT);
+      assert(type == GL_DEPTH_STENCIL_EXT);
       /* XXX correct? */
       srcRb = srcFb->Attachment[BUFFER_DEPTH].Renderbuffer;
       dstRb = dstFb->Attachment[BUFFER_DEPTH].Renderbuffer;
diff --git a/mesalib/src/mesa/swrast/s_drawpix.c b/mesalib/src/mesa/swrast/s_drawpix.c
index c99251904..bf427266c 100644
--- a/mesalib/src/mesa/swrast/s_drawpix.c
+++ b/mesalib/src/mesa/swrast/s_drawpix.c
@@ -373,7 +373,7 @@ draw_depth_pixels( struct gl_context *ctx, GLint x, GLint y,
       while (skipPixels < width) {
          const GLint spanWidth = MIN2(width - skipPixels, SWRAST_MAX_WIDTH);
          GLint row;
-         ASSERT(span.end <= SWRAST_MAX_WIDTH);
+         assert(span.end <= SWRAST_MAX_WIDTH);
          for (row = 0; row < height; row++) {
             const GLvoid *zSrc = _mesa_image_address2d(unpack,
                                                       pixels, width, height,
@@ -595,8 +595,8 @@ draw_depth_stencil_pixels(struct gl_context *ctx, GLint x, GLint y,
    
    depthRb = ctx->ReadBuffer->Attachment[BUFFER_DEPTH].Renderbuffer;
    stencilRb = ctx->ReadBuffer->Attachment[BUFFER_STENCIL].Renderbuffer;
-   ASSERT(depthRb);
-   ASSERT(stencilRb);
+   assert(depthRb);
+   assert(stencilRb);
 
    if (depthRb == stencilRb &&
        (depthRb->Format == MESA_FORMAT_S8_UINT_Z24_UNORM ||
diff --git a/mesalib/src/mesa/swrast/s_fog.c b/mesalib/src/mesa/swrast/s_fog.c
index 34b7c2d28..e270b7e24 100644
--- a/mesalib/src/mesa/swrast/s_fog.c
+++ b/mesalib/src/mesa/swrast/s_fog.c
@@ -23,6 +23,7 @@
  */
 
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/colormac.h"
 #include "main/macros.h"
@@ -49,12 +50,12 @@ _swrast_z_to_fogfactor(struct gl_context *ctx, GLfloat z)
       return CLAMP(f, 0.0F, 1.0F);
    case GL_EXP:
       d = ctx->Fog.Density;
-      f = EXPF(-d * z);
+      f = expf(-d * z);
       f = CLAMP(f, 0.0F, 1.0F);
       return f;
    case GL_EXP2:
       d = ctx->Fog.Density;
-      f = EXPF(-(d * d * z * z));
+      f = expf(-(d * d * z * z));
       f = CLAMP(f, 0.0F, 1.0F);
       return f;
    default:
@@ -66,14 +67,14 @@ _swrast_z_to_fogfactor(struct gl_context *ctx, GLfloat z)
 
 #define LINEAR_FOG(f, coord)  f = (fogEnd - coord) * fogScale
 
-#define EXP_FOG(f, coord)  f = EXPF(density * coord)
+#define EXP_FOG(f, coord)  f = expf(density * coord)
 
 #define EXP2_FOG(f, coord)				\
 do {							\
    GLfloat tmp = negDensitySquared * coord * coord;	\
    if (tmp < FLT_MIN_10_EXP)				\
       tmp = FLT_MIN_10_EXP;				\
-   f = EXPF(tmp);					\
+   f = expf(tmp);					\
  } while(0)
 
 
@@ -91,7 +92,7 @@ if (span->arrayAttribs & VARYING_BIT_FOGC) {					\
    GLuint i;									\
    for (i = 0; i < span->end; i++) {						\
       const GLfloat fogCoord = span->array->attribs[VARYING_SLOT_FOGC][i][0];	\
-      const GLfloat c = FABSF(fogCoord);					\
+      const GLfloat c = fabsf(fogCoord);					\
       GLfloat f, oneMinusF;							\
       FOG_FUNC(f, c);								\
       f = CLAMP(f, 0.0F, 1.0F);							\
@@ -108,7 +109,7 @@ else {										\
    GLfloat w = span->attrStart[VARYING_SLOT_POS][3];				\
    GLuint i;									\
    for (i = 0; i < span->end; i++) {						\
-      const GLfloat c = FABSF(fogCoord) / w;					\
+      const GLfloat c = fabsf(fogCoord) / w;					\
       GLfloat f, oneMinusF;							\
       FOG_FUNC(f, c);								\
       f = CLAMP(f, 0.0F, 1.0F);							\
@@ -134,8 +135,8 @@ _swrast_fog_rgba_span( const struct gl_context *ctx, SWspan *span )
    const SWcontext *swrast = CONST_SWRAST_CONTEXT(ctx);
    GLfloat rFog, gFog, bFog;
 
-   ASSERT(swrast->_FogEnabled);
-   ASSERT(span->arrayMask & SPAN_RGBA);
+   assert(swrast->_FogEnabled);
+   assert(span->arrayMask & SPAN_RGBA);
 
    /* compute (scaled) fog color */
    if (span->array->ChanType == GL_UNSIGNED_BYTE) {
@@ -174,7 +175,7 @@ _swrast_fog_rgba_span( const struct gl_context *ctx, SWspan *span )
             }
             else {
                GLfloat (*rgba)[4] = span->array->attribs[VARYING_SLOT_COL0];
-               ASSERT(span->array->ChanType == GL_FLOAT);
+               assert(span->array->ChanType == GL_FLOAT);
                FOG_LOOP(GLfloat, LINEAR_FOG);
             }
          }
@@ -193,7 +194,7 @@ _swrast_fog_rgba_span( const struct gl_context *ctx, SWspan *span )
             }
             else {
                GLfloat (*rgba)[4] = span->array->attribs[VARYING_SLOT_COL0];
-               ASSERT(span->array->ChanType == GL_FLOAT);
+               assert(span->array->ChanType == GL_FLOAT);
                FOG_LOOP(GLfloat, EXP_FOG);
             }
          }
@@ -212,7 +213,7 @@ _swrast_fog_rgba_span( const struct gl_context *ctx, SWspan *span )
             }
             else {
                GLfloat (*rgba)[4] = span->array->attribs[VARYING_SLOT_COL0];
-               ASSERT(span->array->ChanType == GL_FLOAT);
+               assert(span->array->ChanType == GL_FLOAT);
                FOG_LOOP(GLfloat, EXP2_FOG);
             }
          }
@@ -237,7 +238,7 @@ _swrast_fog_rgba_span( const struct gl_context *ctx, SWspan *span )
       }
       else {
          GLfloat (*rgba)[4] = span->array->attribs[VARYING_SLOT_COL0];
-         ASSERT(span->array->ChanType == GL_FLOAT);
+         assert(span->array->ChanType == GL_FLOAT);
          FOG_LOOP(GLfloat, BLEND_FOG);
       }
    }
diff --git a/mesalib/src/mesa/swrast/s_fragprog.c b/mesalib/src/mesa/swrast/s_fragprog.c
index 1d7c33619..12bcda311 100644
--- a/mesalib/src/mesa/swrast/s_fragprog.c
+++ b/mesalib/src/mesa/swrast/s_fragprog.c
@@ -273,7 +273,7 @@ _swrast_exec_fragment_program( struct gl_context *ctx, SWspan *span )
 
    /* incoming colors should be floats */
    if (program->Base.InputsRead & VARYING_BIT_COL0) {
-      ASSERT(span->array->ChanType == GL_FLOAT);
+      assert(span->array->ChanType == GL_FLOAT);
    }
 
    run_program(ctx, span, 0, span->end);
diff --git a/mesalib/src/mesa/swrast/s_lines.c b/mesalib/src/mesa/swrast/s_lines.c
index b63296aa4..3e626b9e0 100644
--- a/mesalib/src/mesa/swrast/s_lines.c
+++ b/mesalib/src/mesa/swrast/s_lines.c
@@ -67,7 +67,7 @@ draw_wide_line( struct gl_context *ctx, SWspan *span, GLboolean xMajor )
                                      ctx->Const.MaxLineWidth);
    GLint start;
 
-   ASSERT(span->end < SWRAST_MAX_WIDTH);
+   assert(span->end < SWRAST_MAX_WIDTH);
 
    if (width & 1)
       start = width / 2;
@@ -233,7 +233,7 @@ _swrast_choose_line( struct gl_context *ctx )
       if (ctx->Line.SmoothFlag) {
          /* antialiased lines */
          _swrast_choose_aa_line_function(ctx);
-         ASSERT(swrast->Line);
+         assert(swrast->Line);
       }
       else if (ctx->Texture._EnabledCoordUnits
                || _swrast_use_fragment_program(ctx)
@@ -252,8 +252,8 @@ _swrast_choose_line( struct gl_context *ctx )
 #endif
       }
       else {
-         ASSERT(!ctx->Depth.Test);
-         ASSERT(ctx->Line.Width == 1.0);
+         assert(!ctx->Depth.Test);
+         assert(ctx->Line.Width == 1.0);
          /* simple lines */
          USE(simple_no_z_rgba_line);
       }
@@ -262,7 +262,7 @@ _swrast_choose_line( struct gl_context *ctx )
       USE(_swrast_feedback_line);
    }
    else {
-      ASSERT(ctx->RenderMode == GL_SELECT);
+      assert(ctx->RenderMode == GL_SELECT);
       USE(_swrast_select_line);
    }
 }
diff --git a/mesalib/src/mesa/swrast/s_linetemp.h b/mesalib/src/mesa/swrast/s_linetemp.h
index 7b1eac36a..352c88428 100644
--- a/mesalib/src/mesa/swrast/s_linetemp.h
+++ b/mesalib/src/mesa/swrast/s_linetemp.h
@@ -212,8 +212,8 @@ NAME( struct gl_context *ctx, const SWvertex *vert0, const SWvertex *vert1 )
 #endif
    }
 
-   ASSERT(dx >= 0);
-   ASSERT(dy >= 0);
+   assert(dx >= 0);
+   assert(dy >= 0);
 
    numPixels = MAX2(dx, dy);
 
diff --git a/mesalib/src/mesa/swrast/s_logic.c b/mesalib/src/mesa/swrast/s_logic.c
index f68273fd2..c7cac2e40 100644
--- a/mesalib/src/mesa/swrast/s_logic.c
+++ b/mesalib/src/mesa/swrast/s_logic.c
@@ -193,8 +193,8 @@ _swrast_logicop_rgba_span(struct gl_context *ctx, struct gl_renderbuffer *rb,
 {
    void *rbPixels;
 
-   ASSERT(span->end < SWRAST_MAX_WIDTH);
-   ASSERT(span->arrayMask & SPAN_RGBA);
+   assert(span->end < SWRAST_MAX_WIDTH);
+   assert(span->arrayMask & SPAN_RGBA);
 
    rbPixels = _swrast_get_dest_rgba(ctx, rb, span);
 
diff --git a/mesalib/src/mesa/swrast/s_masking.c b/mesalib/src/mesa/swrast/s_masking.c
index c3aa41891..c95587b20 100644
--- a/mesalib/src/mesa/swrast/s_masking.c
+++ b/mesalib/src/mesa/swrast/s_masking.c
@@ -46,8 +46,8 @@ _swrast_mask_rgba_span(struct gl_context *ctx, struct gl_renderbuffer *rb,
    const GLuint n = span->end;
    void *rbPixels;
 
-   ASSERT(n < SWRAST_MAX_WIDTH);
-   ASSERT(span->arrayMask & SPAN_RGBA);
+   assert(n < SWRAST_MAX_WIDTH);
+   assert(span->arrayMask & SPAN_RGBA);
 
    rbPixels = _swrast_get_dest_rgba(ctx, rb, span);
 
diff --git a/mesalib/src/mesa/swrast/s_points.c b/mesalib/src/mesa/swrast/s_points.c
index b9bec237e..8180483ba 100644
--- a/mesalib/src/mesa/swrast/s_points.c
+++ b/mesalib/src/mesa/swrast/s_points.c
@@ -140,7 +140,7 @@ sprite_point(struct gl_context *ctx, const SWvertex *vert)
          if (attr >= VARYING_SLOT_TEX0 && attr <= VARYING_SLOT_TEX7) {
             /* a texcoord attribute */
             const GLuint u = attr - VARYING_SLOT_TEX0;
-            ASSERT(u < Elements(ctx->Point.CoordReplace));
+            assert(u < ARRAY_SIZE(ctx->Point.CoordReplace));
             if (ctx->Point.CoordReplace[u]) {
                tCoords[numTcoords++] = attr;
 
@@ -504,7 +504,7 @@ pixel_point(struct gl_context *ctx, const SWvertex *vert)
    span->array->z[count] = (GLint) (vert->attrib[VARYING_SLOT_POS][2] + 0.5F);
 
    span->end = count + 1;
-   ASSERT(span->end <= SWRAST_MAX_WIDTH);
+   assert(span->end <= SWRAST_MAX_WIDTH);
 }
 
 
diff --git a/mesalib/src/mesa/swrast/s_renderbuffer.c b/mesalib/src/mesa/swrast/s_renderbuffer.c
index dfd3a6057..d8c4ed386 100644
--- a/mesalib/src/mesa/swrast/s_renderbuffer.c
+++ b/mesalib/src/mesa/swrast/s_renderbuffer.c
@@ -153,7 +153,7 @@ soft_renderbuffer_storage(struct gl_context *ctx, struct gl_renderbuffer *rb,
    }
    else {
       /* the internalFormat should have been error checked long ago */
-      ASSERT(rb->_BaseFormat);
+      assert(rb->_BaseFormat);
    }
 
    return GL_TRUE;
diff --git a/mesalib/src/mesa/swrast/s_span.c b/mesalib/src/mesa/swrast/s_span.c
index 321959df9..5d618f048 100644
--- a/mesalib/src/mesa/swrast/s_span.c
+++ b/mesalib/src/mesa/swrast/s_span.c
@@ -31,6 +31,7 @@
  * \author Brian Paul
  */
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/colormac.h"
 #include "main/format_pack.h"
@@ -205,7 +206,7 @@ interpolate_active_attribs(struct gl_context *ctx, SWspan *span,
             v3 += dv3dx;
             w += dwdx;
          }
-         ASSERT((span->arrayAttribs & BITFIELD64_BIT(attr)) == 0);
+         assert((span->arrayAttribs & BITFIELD64_BIT(attr)) == 0);
          span->arrayAttribs |= BITFIELD64_BIT(attr);
       }
    ATTRIB_LOOP_END
@@ -223,7 +224,7 @@ interpolate_int_colors(struct gl_context *ctx, SWspan *span)
    const GLuint n = span->end;
    GLuint i;
 
-   ASSERT(!(span->arrayMask & SPAN_RGBA));
+   assert(!(span->arrayMask & SPAN_RGBA));
 #endif
 
    switch (span->array->ChanType) {
@@ -336,7 +337,7 @@ interpolate_float_colors(SWspan *span)
    }
    else {
       /* interpolate red/green/blue/alpha to get float colors */
-      ASSERT(span->interpMask & SPAN_RGBA);
+      assert(span->interpMask & SPAN_RGBA);
       if (span->interpMask & SPAN_FLAT) {
          GLfloat r = FixedToFloat(span->red);
          GLfloat g = FixedToFloat(span->green);
@@ -383,7 +384,7 @@ _swrast_span_interpolate_z( const struct gl_context *ctx, SWspan *span )
    const GLuint n = span->end;
    GLuint i;
 
-   ASSERT(!(span->arrayMask & SPAN_Z));
+   assert(!(span->arrayMask & SPAN_Z));
 
    if (ctx->DrawBuffer->Visual.depthBits <= 16) {
       GLfixed zval = span->z;
@@ -443,10 +444,10 @@ _swrast_compute_lambda(GLfloat dsdx, GLfloat dsdy, GLfloat dtdx, GLfloat dtdy,
    GLfloat dsdy2 = (s + dsdy) / (q + dqdy) - s * invQ;
    GLfloat dtdy2 = (t + dtdy) / (q + dqdy) - t * invQ;
    GLfloat maxU, maxV, rho, lambda;
-   dsdx2 = FABSF(dsdx2);
-   dsdy2 = FABSF(dsdy2);
-   dtdx2 = FABSF(dtdx2);
-   dtdy2 = FABSF(dtdy2);
+   dsdx2 = fabsf(dsdx2);
+   dsdy2 = fabsf(dsdy2);
+   dtdx2 = fabsf(dtdx2);
+   dtdy2 = fabsf(dtdy2);
    maxU = MAX2(dsdx2, dsdy2) * texW;
    maxV = MAX2(dtdx2, dtdy2) * texH;
    rho = MAX2(maxU, maxV);
@@ -658,7 +659,7 @@ stipple_polygon_span(struct gl_context *ctx, SWspan *span)
 {
    GLubyte *mask = span->array->mask;
 
-   ASSERT(ctx->Polygon.StippleFlag);
+   assert(ctx->Polygon.StippleFlag);
 
    if (span->arrayMask & SPAN_XY) {
       /* arrays of x/y pixel coords */
@@ -748,7 +749,7 @@ clip_span( struct gl_context *ctx, SWspan *span )
 
       /* Clip to right */
       if (x + n > xmax) {
-         ASSERT(x < xmax);
+         assert(x < xmax);
          n = span->end = xmax - x;
       }
 
@@ -757,8 +758,8 @@ clip_span( struct gl_context *ctx, SWspan *span )
          const GLint leftClip = xmin - x;
          GLuint i;
 
-         ASSERT(leftClip > 0);
-         ASSERT(x + n > xmin);
+         assert(leftClip > 0);
+         assert(x + n > xmin);
 
          /* Clip 'leftClip' pixels from the left side.
           * The span->leftClip field will be applied when we interpolate
@@ -813,10 +814,10 @@ clip_span( struct gl_context *ctx, SWspan *span )
          span->writeAll = GL_FALSE;
       }
 
-      ASSERT(span->x >= xmin);
-      ASSERT(span->x + span->end <= xmax);
-      ASSERT(span->y >= ymin);
-      ASSERT(span->y < ymax);
+      assert(span->x >= xmin);
+      assert(span->x + span->end <= xmax);
+      assert(span->y >= ymin);
+      assert(span->y < ymax);
 
       return GL_TRUE;  /* some pixels visible */
    }
@@ -837,9 +838,9 @@ add_specular(struct gl_context *ctx, SWspan *span)
    GLfloat (*col1)[4] = span->array->attribs[VARYING_SLOT_COL1];
    GLuint i;
 
-   ASSERT(!_swrast_use_fragment_program(ctx));
-   ASSERT(span->arrayMask & SPAN_RGBA);
-   ASSERT(swrast->_ActiveAttribMask & VARYING_BIT_COL1);
+   assert(!_swrast_use_fragment_program(ctx));
+   assert(span->arrayMask & SPAN_RGBA);
+   assert(swrast->_ActiveAttribMask & VARYING_BIT_COL1);
    (void) swrast; /* silence warning */
 
    if (span->array->ChanType == GL_FLOAT) {
@@ -859,8 +860,8 @@ add_specular(struct gl_context *ctx, SWspan *span)
       interpolate_active_attribs(ctx, span, VARYING_BIT_COL1);
    }
 
-   ASSERT(span->arrayAttribs & VARYING_BIT_COL0);
-   ASSERT(span->arrayAttribs & VARYING_BIT_COL1);
+   assert(span->arrayAttribs & VARYING_BIT_COL0);
+   assert(span->arrayAttribs & VARYING_BIT_COL1);
 
    for (i = 0; i < span->end; i++) {
       if (mask[i]) {
@@ -887,8 +888,8 @@ apply_aa_coverage(SWspan *span)
       for (i = 0; i < span->end; i++) {
          const GLfloat a = rgba[i][ACOMP] * coverage[i];
          rgba[i][ACOMP] = (GLubyte) CLAMP(a, 0.0, 255.0);
-         ASSERT(coverage[i] >= 0.0);
-         ASSERT(coverage[i] <= 1.0);
+         assert(coverage[i] >= 0.0);
+         assert(coverage[i] <= 1.0);
       }
    }
    else if (span->array->ChanType == GL_UNSIGNED_SHORT) {
@@ -916,7 +917,7 @@ clamp_colors(SWspan *span)
 {
    GLfloat (*rgba)[4] = span->array->attribs[VARYING_SLOT_COL0];
    GLuint i;
-   ASSERT(span->array->ChanType == GL_FLOAT);
+   assert(span->array->ChanType == GL_FLOAT);
    for (i = 0; i < span->end; i++) {
       rgba[i][RCOMP] = CLAMP(rgba[i][RCOMP], 0.0F, 1.0F);
       rgba[i][GCOMP] = CLAMP(rgba[i][GCOMP], 0.0F, 1.0F);
@@ -945,7 +946,7 @@ convert_color_type(SWspan *span, GLenum srcType, GLenum newType, GLuint output)
       src = span->array->rgba8;
    }
    else {
-      ASSERT(srcType == GL_UNSIGNED_SHORT);
+      assert(srcType == GL_UNSIGNED_SHORT);
       src = span->array->rgba16;
    }
 
@@ -1008,7 +1009,7 @@ shade_texture_span(struct gl_context *ctx, SWspan *span)
          _swrast_exec_fragment_program(ctx, span);
       }
       else {
-         ASSERT(ctx->ATIFragmentShader._Enabled);
+         assert(ctx->ATIFragmentShader._Enabled);
          _swrast_exec_fragment_shader(ctx, span);
       }
    }
@@ -1147,7 +1148,7 @@ _swrast_write_rgba_span( struct gl_context *ctx, SWspan *span)
           span->interpMask, span->arrayMask);
    */
 
-   ASSERT(span->primitive == GL_POINT ||
+   assert(span->primitive == GL_POINT ||
           span->primitive == GL_LINE ||
 	  span->primitive == GL_POLYGON ||
           span->primitive == GL_BITMAP);
@@ -1167,7 +1168,7 @@ _swrast_write_rgba_span( struct gl_context *ctx, SWspan *span)
       return;
    }
 
-   ASSERT(span->end <= SWRAST_MAX_WIDTH);
+   assert(span->end <= SWRAST_MAX_WIDTH);
 
    /* Depth bounds test */
    if (ctx->Depth.BoundsTest && fb->Visual.depthBits > 0) {
@@ -1229,8 +1230,8 @@ _swrast_write_rgba_span( struct gl_context *ctx, SWspan *span)
       }
       else if (fb->Visual.depthBits > 0) {
          /* Just regular depth testing */
-         ASSERT(ctx->Depth.Test);
-         ASSERT(span->arrayMask & SPAN_Z);
+         assert(ctx->Depth.Test);
+         assert(span->arrayMask & SPAN_Z);
          if (!_swrast_depth_test_span(ctx, span)) {
             /* all fragments failed test */
             goto end;
@@ -1272,7 +1273,7 @@ _swrast_write_rgba_span( struct gl_context *ctx, SWspan *span)
    }
 #endif
 
-   ASSERT(span->arrayMask & SPAN_RGBA);
+   assert(span->arrayMask & SPAN_RGBA);
 
    if (span->primitive == GL_BITMAP || !swrast->SpecularVertexAdd) {
       /* Add primary and specular (diffuse + specular) colors */
@@ -1353,7 +1354,7 @@ _swrast_write_rgba_span( struct gl_context *ctx, SWspan *span)
                       4 * span->end * sizeof(GLchan));
             }
 
-            ASSERT(rb->_BaseFormat == GL_RGBA ||
+            assert(rb->_BaseFormat == GL_RGBA ||
                    rb->_BaseFormat == GL_RGB ||
                    rb->_BaseFormat == GL_RED ||
                    rb->_BaseFormat == GL_RG ||
@@ -1457,8 +1458,8 @@ _swrast_read_rgba_span( struct gl_context *ctx, struct gl_renderbuffer *rb,
          length = (GLint) n;
       }
 
-      ASSERT(rb);
-      ASSERT(rb->_BaseFormat == GL_RGBA ||
+      assert(rb);
+      assert(rb->_BaseFormat == GL_RGBA ||
 	     rb->_BaseFormat == GL_RGB ||
 	     rb->_BaseFormat == GL_RG ||
 	     rb->_BaseFormat == GL_RED ||
@@ -1468,6 +1469,7 @@ _swrast_read_rgba_span( struct gl_context *ctx, struct gl_renderbuffer *rb,
 	     rb->_BaseFormat == GL_ALPHA);
 
       assert(srb->Map);
+      (void) srb; /* silence unused var warning */
 
       src = _swrast_pixel_address(rb, x + skip, y);
 
diff --git a/mesalib/src/mesa/swrast/s_stencil.c b/mesalib/src/mesa/swrast/s_stencil.c
index eba9da863..294b593a2 100644
--- a/mesalib/src/mesa/swrast/s_stencil.c
+++ b/mesalib/src/mesa/swrast/s_stencil.c
@@ -280,7 +280,7 @@ compute_pass_fail_masks(GLuint n, const GLubyte origMask[],
 {
    GLuint i;
    for (i = 0; i < n; i++) {
-      ASSERT(newMask[i] == 0 || newMask[i] == 1);
+      assert(newMask[i] == 0 || newMask[i] == 1);
       passMask[i] = origMask[i] & newMask[i];
       failMask[i] = origMask[i] & (newMask[i] ^ 1);
    }
diff --git a/mesalib/src/mesa/swrast/s_texcombine.c b/mesalib/src/mesa/swrast/s_texcombine.c
index def5eb19e..58ff16465 100644
--- a/mesalib/src/mesa/swrast/s_texcombine.c
+++ b/mesalib/src/mesa/swrast/s_texcombine.c
@@ -188,7 +188,7 @@ texture_combine( struct gl_context *ctx, GLuint unit,
             /* ARB_texture_env_crossbar source */
             {
                const GLuint srcUnit = srcRGB - GL_TEXTURE0;
-               ASSERT(srcUnit < ctx->Const.MaxTextureUnits);
+               assert(srcUnit < ctx->Const.MaxTextureUnits);
                if (!ctx->Texture.Unit[srcUnit]._Current)
                   goto end;
                argRGB[term] = get_texel_array(swrast, srcUnit);
@@ -278,7 +278,7 @@ texture_combine( struct gl_context *ctx, GLuint unit,
             /* ARB_texture_env_crossbar source */
             {
                const GLuint srcUnit = srcA - GL_TEXTURE0;
-               ASSERT(srcUnit < ctx->Const.MaxTextureUnits);
+               assert(srcUnit < ctx->Const.MaxTextureUnits);
                if (!ctx->Texture.Unit[srcUnit]._Current)
                   goto end;
                argA[term] = get_texel_array(swrast, srcUnit);
@@ -628,7 +628,7 @@ _swrast_texture_span( struct gl_context *ctx, SWspan *span )
       return;
    }
 
-   ASSERT(span->end <= SWRAST_MAX_WIDTH);
+   assert(span->end <= SWRAST_MAX_WIDTH);
 
    /*
     * Save copy of the incoming fragment colors (the GL_PRIMARY_COLOR)
diff --git a/mesalib/src/mesa/swrast/s_texfetch.c b/mesalib/src/mesa/swrast/s_texfetch.c
index 9629024b9..3c4ee15ba 100644
--- a/mesalib/src/mesa/swrast/s_texfetch.c
+++ b/mesalib/src/mesa/swrast/s_texfetch.c
@@ -573,7 +573,7 @@ set_fetch_functions(const struct gl_sampler_object *samp,
    }
 #endif
 
-   STATIC_ASSERT(Elements(texfetch_funcs) == MESA_FORMAT_COUNT);
+   STATIC_ASSERT(ARRAY_SIZE(texfetch_funcs) == MESA_FORMAT_COUNT);
 
    if (samp->sRGBDecode == GL_SKIP_DECODE_EXT &&
        _mesa_get_format_color_encoding(format) == GL_SRGB) {
@@ -598,7 +598,7 @@ set_fetch_functions(const struct gl_sampler_object *samp,
 
    texImage->FetchCompressedTexel = _mesa_get_compressed_fetch_func(format);
 
-   ASSERT(texImage->FetchTexel);
+   assert(texImage->FetchTexel);
 }
 
 void
diff --git a/mesalib/src/mesa/swrast/s_texfilter.c b/mesalib/src/mesa/swrast/s_texfilter.c
index fa79fdc5b..3ade99519 100644
--- a/mesalib/src/mesa/swrast/s_texfilter.c
+++ b/mesalib/src/mesa/swrast/s_texfilter.c
@@ -23,6 +23,7 @@
  */
 
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/colormac.h"
@@ -223,7 +224,7 @@ linear_texel_locations(GLenum wrapMode,
       }
       break;
    case GL_MIRROR_CLAMP_EXT:
-      u = FABSF(s);
+      u = fabsf(s);
       if (u >= 1.0F)
          u = (GLfloat) size;
       else
@@ -233,7 +234,7 @@ linear_texel_locations(GLenum wrapMode,
       *i1 = *i0 + 1;
       break;
    case GL_MIRROR_CLAMP_TO_EDGE_EXT:
-      u = FABSF(s);
+      u = fabsf(s);
       if (u >= 1.0F)
          u = (GLfloat) size;
       else
@@ -250,7 +251,7 @@ linear_texel_locations(GLenum wrapMode,
       {
          const GLfloat min = -1.0F / (2.0F * size);
          const GLfloat max = 1.0F - min;
-         u = FABSF(s);
+         u = fabsf(s);
          if (u <= min)
             u = min * size;
          else if (u >= max)
@@ -354,7 +355,7 @@ nearest_texel_location(GLenum wrapMode,
       {
          /* s limited to [0,1] */
          /* i limited to [0,size-1] */
-         const GLfloat u = FABSF(s);
+         const GLfloat u = fabsf(s);
          if (u <= 0.0F)
             i = 0;
          else if (u >= 1.0F)
@@ -369,7 +370,7 @@ nearest_texel_location(GLenum wrapMode,
          /* i limited to [0, size-1] */
          const GLfloat min = 1.0F / (2.0F * size);
          const GLfloat max = 1.0F - min;
-         const GLfloat u = FABSF(s);
+         const GLfloat u = fabsf(s);
          if (u < min)
             i = 0;
          else if (u > max)
@@ -384,7 +385,7 @@ nearest_texel_location(GLenum wrapMode,
          /* i limited to [0, size-1] */
          const GLfloat min = -1.0F / (2.0F * size);
          const GLfloat max = 1.0F - min;
-         const GLfloat u = FABSF(s);
+         const GLfloat u = fabsf(s);
          if (u < min)
             i = -1;
          else if (u > max)
@@ -668,7 +669,7 @@ compute_min_mag_ranges(const struct gl_sampler_object *samp,
    GLfloat minMagThresh;
 
    /* we shouldn't be here if minfilter == magfilter */
-   ASSERT(samp->MinFilter != samp->MagFilter);
+   assert(samp->MinFilter != samp->MagFilter);
 
    /* This bit comes from the OpenGL spec: */
    if (samp->MagFilter == GL_LINEAR
@@ -690,12 +691,12 @@ compute_min_mag_ranges(const struct gl_sampler_object *samp,
       printf("lambda delta = %g\n", lambda[0] - lambda[n-1]);
       if (lambda[0] >= lambda[n-1]) { /* decreasing */
          for (i = 0; i < n - 1; i++) {
-            ASSERT((GLint) (lambda[i] * 10) >= (GLint) (lambda[i+1] * 10));
+            assert((GLint) (lambda[i] * 10) >= (GLint) (lambda[i+1] * 10));
          }
       }
       else { /* increasing */
          for (i = 0; i < n - 1; i++) {
-            ASSERT((GLint) (lambda[i] * 10) <= (GLint) (lambda[i+1] * 10));
+            assert((GLint) (lambda[i] * 10) <= (GLint) (lambda[i+1] * 10));
          }
       }
    }
@@ -749,13 +750,13 @@ compute_min_mag_ranges(const struct gl_sampler_object *samp,
       for (i = 0; i < n; i++) {
          if (lambda[i] > minMagThresh) {
             /* minification */
-            ASSERT(i >= *minStart);
-            ASSERT(i < *minEnd);
+            assert(i >= *minStart);
+            assert(i < *minEnd);
          }
          else {
             /* magnification */
-            ASSERT(i >= *magStart);
-            ASSERT(i < *magEnd);
+            assert(i >= *magStart);
+            assert(i < *magEnd);
          }
       }
    }
@@ -920,7 +921,7 @@ sample_1d_nearest_mipmap_nearest(struct gl_context *ctx,
                                  const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = nearest_mipmap_level(tObj, lambda[i]);
       sample_1d_nearest(ctx, samp, tObj->Image[0][level], texcoord[i], rgba[i]);
@@ -936,7 +937,7 @@ sample_1d_linear_mipmap_nearest(struct gl_context *ctx,
                                 const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = nearest_mipmap_level(tObj, lambda[i]);
       sample_1d_linear(ctx, samp, tObj->Image[0][level], texcoord[i], rgba[i]);
@@ -952,7 +953,7 @@ sample_1d_nearest_mipmap_linear(struct gl_context *ctx,
                                 const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -978,7 +979,7 @@ sample_1d_linear_mipmap_linear(struct gl_context *ctx,
                                const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -1042,7 +1043,7 @@ sample_lambda_1d( struct gl_context *ctx,
    GLuint magStart, magEnd;  /* texels with magnification */
    GLuint i;
 
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    compute_min_mag_ranges(samp, n, lambda,
                           &minStart, &minEnd, &magStart, &magEnd);
 
@@ -1226,10 +1227,10 @@ sample_2d_linear_repeat(struct gl_context *ctx,
 
    (void) ctx;
 
-   ASSERT(samp->WrapS == GL_REPEAT);
-   ASSERT(samp->WrapT == GL_REPEAT);
-   ASSERT(img->Border == 0);
-   ASSERT(swImg->_IsPowerOfTwo);
+   assert(samp->WrapS == GL_REPEAT);
+   assert(samp->WrapT == GL_REPEAT);
+   assert(img->Border == 0);
+   assert(swImg->_IsPowerOfTwo);
 
    linear_repeat_texel_location(width,  texcoord[0], &i0, &i1, &wi);
    linear_repeat_texel_location(height, texcoord[1], &j0, &j1, &wj);
@@ -1266,7 +1267,7 @@ sample_2d_linear_mipmap_nearest(struct gl_context *ctx,
                                 const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = nearest_mipmap_level(tObj, lambda[i]);
       sample_2d_linear(ctx, samp, tObj->Image[0][level], texcoord[i], rgba[i]);
@@ -1282,7 +1283,7 @@ sample_2d_nearest_mipmap_linear(struct gl_context *ctx,
                                 const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -1308,7 +1309,7 @@ sample_2d_linear_mipmap_linear( struct gl_context *ctx,
                                 const GLfloat lambda[], GLfloat rgba[][4] )
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -1334,9 +1335,9 @@ sample_2d_linear_mipmap_linear_repeat(struct gl_context *ctx,
                                       const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
-   ASSERT(samp->WrapS == GL_REPEAT);
-   ASSERT(samp->WrapT == GL_REPEAT);
+   assert(lambda != NULL);
+   assert(samp->WrapS == GL_REPEAT);
+   assert(samp->WrapT == GL_REPEAT);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -1426,11 +1427,11 @@ opt_sample_rgb_2d(struct gl_context *ctx,
    GLuint k;
    (void) ctx;
    (void) lambda;
-   ASSERT(samp->WrapS==GL_REPEAT);
-   ASSERT(samp->WrapT==GL_REPEAT);
-   ASSERT(img->Border==0);
-   ASSERT(img->TexFormat == MESA_FORMAT_BGR_UNORM8);
-   ASSERT(swImg->_IsPowerOfTwo);
+   assert(samp->WrapS==GL_REPEAT);
+   assert(samp->WrapT==GL_REPEAT);
+   assert(img->Border==0);
+   assert(img->TexFormat == MESA_FORMAT_BGR_UNORM8);
+   assert(swImg->_IsPowerOfTwo);
    (void) swImg;
 
    for (k=0; k<n; k++) {
@@ -1471,11 +1472,11 @@ opt_sample_rgba_2d(struct gl_context *ctx,
    GLuint i;
    (void) ctx;
    (void) lambda;
-   ASSERT(samp->WrapS==GL_REPEAT);
-   ASSERT(samp->WrapT==GL_REPEAT);
-   ASSERT(img->Border==0);
-   ASSERT(img->TexFormat == MESA_FORMAT_A8B8G8R8_UNORM);
-   ASSERT(swImg->_IsPowerOfTwo);
+   assert(samp->WrapS==GL_REPEAT);
+   assert(samp->WrapT==GL_REPEAT);
+   assert(img->Border==0);
+   assert(img->TexFormat == MESA_FORMAT_A8B8G8R8_UNORM);
+   assert(swImg->_IsPowerOfTwo);
    (void) swImg;
 
    for (i = 0; i < n; i++) {
@@ -1511,7 +1512,7 @@ sample_lambda_2d(struct gl_context *ctx,
           swImg->RowStride)
       && swImg->_IsPowerOfTwo;
 
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    compute_min_mag_ranges(samp, n, lambda,
                           &minStart, &minEnd, &magStart, &magEnd);
 
@@ -1674,7 +1675,7 @@ sample_2d_ewa(struct gl_context *ctx,
    GLfloat F = A*C-B*B/4.0f;
 
    /* check if it is an ellipse */
-   /* ASSERT(F > 0.0); */
+   /* assert(F > 0.0); */
 
    /* Compute the ellipse's (u,v) bounding box in texture space */
    GLfloat d = -B*B+4.0f*C*A;
@@ -2164,7 +2165,7 @@ sample_3d_linear_mipmap_nearest(struct gl_context *ctx,
                                 const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = nearest_mipmap_level(tObj, lambda[i]);
       sample_3d_linear(ctx, samp, tObj->Image[0][level], texcoord[i], rgba[i]);
@@ -2180,7 +2181,7 @@ sample_3d_nearest_mipmap_linear(struct gl_context *ctx,
                                 const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -2206,7 +2207,7 @@ sample_3d_linear_mipmap_linear(struct gl_context *ctx,
                                const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -2270,7 +2271,7 @@ sample_lambda_3d(struct gl_context *ctx,
    GLuint magStart, magEnd;  /* texels with magnification */
    GLuint i;
 
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    compute_min_mag_ranges(samp, n, lambda,
                           &minStart, &minEnd, &magStart, &magEnd);
 
@@ -2358,7 +2359,7 @@ choose_cube_face(const struct gl_texture_object *texObj,
    const GLfloat rx = texcoord[0];
    const GLfloat ry = texcoord[1];
    const GLfloat rz = texcoord[2];
-   const GLfloat arx = FABSF(rx), ary = FABSF(ry), arz = FABSF(rz);
+   const GLfloat arx = fabsf(rx), ary = fabsf(ry), arz = fabsf(rz);
    GLuint face;
    GLfloat sc, tc, ma;
 
@@ -2471,7 +2472,7 @@ sample_cube_nearest_mipmap_nearest(struct gl_context *ctx,
                                    const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       const struct gl_texture_image **images;
       GLfloat newCoord[4];
@@ -2506,7 +2507,7 @@ sample_cube_linear_mipmap_nearest(struct gl_context *ctx,
                                   const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       const struct gl_texture_image **images;
       GLfloat newCoord[4];
@@ -2531,7 +2532,7 @@ sample_cube_nearest_mipmap_linear(struct gl_context *ctx,
                                   const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       const struct gl_texture_image **images;
       GLfloat newCoord[4];
@@ -2566,7 +2567,7 @@ sample_cube_linear_mipmap_linear(struct gl_context *ctx,
                                  const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       const struct gl_texture_image **images;
       GLfloat newCoord[4];
@@ -2604,7 +2605,7 @@ sample_lambda_cube(struct gl_context *ctx,
    GLuint minStart, minEnd;  /* texels with minification */
    GLuint magStart, magEnd;  /* texels with magnification */
 
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    compute_min_mag_ranges(samp, n, lambda,
                           &minStart, &minEnd, &magStart, &magEnd);
 
@@ -2687,10 +2688,10 @@ sample_nearest_rect(struct gl_context *ctx,
    (void) ctx;
    (void) lambda;
 
-   ASSERT(samp->WrapS == GL_CLAMP ||
+   assert(samp->WrapS == GL_CLAMP ||
           samp->WrapS == GL_CLAMP_TO_EDGE ||
           samp->WrapS == GL_CLAMP_TO_BORDER);
-   ASSERT(samp->WrapT == GL_CLAMP ||
+   assert(samp->WrapT == GL_CLAMP ||
           samp->WrapT == GL_CLAMP_TO_EDGE ||
           samp->WrapT == GL_CLAMP_TO_BORDER);
 
@@ -2722,10 +2723,10 @@ sample_linear_rect(struct gl_context *ctx,
    (void) ctx;
    (void) lambda;
 
-   ASSERT(samp->WrapS == GL_CLAMP ||
+   assert(samp->WrapS == GL_CLAMP ||
           samp->WrapS == GL_CLAMP_TO_EDGE ||
           samp->WrapS == GL_CLAMP_TO_BORDER);
-   ASSERT(samp->WrapT == GL_CLAMP ||
+   assert(samp->WrapT == GL_CLAMP ||
           samp->WrapT == GL_CLAMP_TO_EDGE ||
           samp->WrapT == GL_CLAMP_TO_BORDER);
 
@@ -2947,7 +2948,7 @@ sample_2d_array_linear_mipmap_nearest(struct gl_context *ctx,
                                       const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = nearest_mipmap_level(tObj, lambda[i]);
       sample_2d_array_linear(ctx, samp, tObj->Image[0][level],
@@ -2964,7 +2965,7 @@ sample_2d_array_nearest_mipmap_linear(struct gl_context *ctx,
                                       const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -2992,7 +2993,7 @@ sample_2d_array_linear_mipmap_linear(struct gl_context *ctx,
                                      const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -3059,7 +3060,7 @@ sample_lambda_2d_array(struct gl_context *ctx,
    GLuint magStart, magEnd;  /* texels with magnification */
    GLuint i;
 
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    compute_min_mag_ranges(samp, n, lambda,
                           &minStart, &minEnd, &magStart, &magEnd);
 
@@ -3242,7 +3243,7 @@ sample_1d_array_linear_mipmap_nearest(struct gl_context *ctx,
                                       const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = nearest_mipmap_level(tObj, lambda[i]);
       sample_1d_array_linear(ctx, samp, tObj->Image[0][level],
@@ -3259,7 +3260,7 @@ sample_1d_array_nearest_mipmap_linear(struct gl_context *ctx,
                                       const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -3285,7 +3286,7 @@ sample_1d_array_linear_mipmap_linear(struct gl_context *ctx,
                                      const GLfloat lambda[], GLfloat rgba[][4])
 {
    GLuint i;
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    for (i = 0; i < n; i++) {
       GLint level = linear_mipmap_level(tObj, lambda[i]);
       if (level >= tObj->_MaxLevel) {
@@ -3349,7 +3350,7 @@ sample_lambda_1d_array(struct gl_context *ctx,
    GLuint magStart, magEnd;  /* texels with magnification */
    GLuint i;
 
-   ASSERT(lambda != NULL);
+   assert(lambda != NULL);
    compute_min_mag_ranges(samp, n, lambda,
                           &minStart, &minEnd, &magStart, &magEnd);
 
@@ -3555,10 +3556,10 @@ sample_depth_texture( struct gl_context *ctx,
    GLenum function;
    GLfloat result;
 
-   ASSERT(img->_BaseFormat == GL_DEPTH_COMPONENT ||
+   assert(img->_BaseFormat == GL_DEPTH_COMPONENT ||
           img->_BaseFormat == GL_DEPTH_STENCIL_EXT);
 
-   ASSERT(tObj->Target == GL_TEXTURE_1D ||
+   assert(tObj->Target == GL_TEXTURE_1D ||
           tObj->Target == GL_TEXTURE_2D ||
           tObj->Target == GL_TEXTURE_RECTANGLE_NV ||
           tObj->Target == GL_TEXTURE_1D_ARRAY_EXT ||
@@ -3595,7 +3596,7 @@ sample_depth_texture( struct gl_context *ctx,
    }
    else {
       GLuint i;
-      ASSERT(samp->MagFilter == GL_LINEAR);
+      assert(samp->MagFilter == GL_LINEAR);
       for (i = 0; i < n; i++) {
          GLfloat depth00, depth01, depth10, depth11, depthRef;
          GLint i0, i1, j0, j1;
@@ -3730,7 +3731,7 @@ _swrast_choose_texture_sample_func( struct gl_context *ctx,
             return &sample_linear_1d;
          }
          else {
-            ASSERT(sampler->MinFilter == GL_NEAREST);
+            assert(sampler->MinFilter == GL_NEAREST);
             return &sample_nearest_1d;
          }
       case GL_TEXTURE_2D:
@@ -3755,7 +3756,7 @@ _swrast_choose_texture_sample_func( struct gl_context *ctx,
                swrast_texture_image_const(img);
             texture_sample_func func;
 
-            ASSERT(sampler->MinFilter == GL_NEAREST);
+            assert(sampler->MinFilter == GL_NEAREST);
             func = &sample_nearest_2d;
             if (sampler->WrapS == GL_REPEAT &&
                 sampler->WrapT == GL_REPEAT &&
@@ -3777,7 +3778,7 @@ _swrast_choose_texture_sample_func( struct gl_context *ctx,
             return &sample_linear_3d;
          }
          else {
-            ASSERT(sampler->MinFilter == GL_NEAREST);
+            assert(sampler->MinFilter == GL_NEAREST);
             return &sample_nearest_3d;
          }
       case GL_TEXTURE_CUBE_MAP:
@@ -3788,7 +3789,7 @@ _swrast_choose_texture_sample_func( struct gl_context *ctx,
             return &sample_linear_cube;
          }
          else {
-            ASSERT(sampler->MinFilter == GL_NEAREST);
+            assert(sampler->MinFilter == GL_NEAREST);
             return &sample_nearest_cube;
          }
       case GL_TEXTURE_RECTANGLE_NV:
@@ -3802,7 +3803,7 @@ _swrast_choose_texture_sample_func( struct gl_context *ctx,
             return &sample_linear_rect;
          }
          else {
-            ASSERT(sampler->MinFilter == GL_NEAREST);
+            assert(sampler->MinFilter == GL_NEAREST);
             return &sample_nearest_rect;
          }
       case GL_TEXTURE_1D_ARRAY_EXT:
@@ -3816,7 +3817,7 @@ _swrast_choose_texture_sample_func( struct gl_context *ctx,
             return &sample_linear_1d_array;
          }
          else {
-            ASSERT(sampler->MinFilter == GL_NEAREST);
+            assert(sampler->MinFilter == GL_NEAREST);
             return &sample_nearest_1d_array;
          }
       case GL_TEXTURE_2D_ARRAY_EXT:
@@ -3830,7 +3831,7 @@ _swrast_choose_texture_sample_func( struct gl_context *ctx,
             return &sample_linear_2d_array;
          }
          else {
-            ASSERT(sampler->MinFilter == GL_NEAREST);
+            assert(sampler->MinFilter == GL_NEAREST);
             return &sample_nearest_2d_array;
          }
       default:
diff --git a/mesalib/src/mesa/swrast/s_texrender.c b/mesalib/src/mesa/swrast/s_texrender.c
index d67e48ad3..29bb270d6 100644
--- a/mesalib/src/mesa/swrast/s_texrender.c
+++ b/mesalib/src/mesa/swrast/s_texrender.c
@@ -18,7 +18,7 @@
 static void
 delete_texture_wrapper(struct gl_context *ctx, struct gl_renderbuffer *rb)
 {
-   ASSERT(rb->RefCount == 0);
+   assert(rb->RefCount == 0);
    free(rb);
 }
 
diff --git a/mesalib/src/mesa/swrast/s_triangle.c b/mesalib/src/mesa/swrast/s_triangle.c
index 1d8e31c2e..af039c359 100644
--- a/mesalib/src/mesa/swrast/s_triangle.c
+++ b/mesalib/src/mesa/swrast/s_triangle.c
@@ -78,8 +78,8 @@ _swrast_culltriangle( struct gl_context *ctx,
 #define NAME flat_rgba_triangle
 #define INTERP_Z 1
 #define SETUP_CODE				\
-   ASSERT(ctx->Texture._EnabledCoordUnits == 0);\
-   ASSERT(ctx->Light.ShadeModel==GL_FLAT);	\
+   assert(ctx->Texture._EnabledCoordUnits == 0);\
+   assert(ctx->Light.ShadeModel==GL_FLAT);	\
    span.interpMask |= SPAN_RGBA;		\
    span.red = ChanToFixed(v2->color[0]);	\
    span.green = ChanToFixed(v2->color[1]);	\
@@ -104,8 +104,8 @@ _swrast_culltriangle( struct gl_context *ctx,
 #define SETUP_CODE				\
    {						\
       /* texturing must be off */		\
-      ASSERT(ctx->Texture._EnabledCoordUnits == 0);	\
-      ASSERT(ctx->Light.ShadeModel==GL_SMOOTH);	\
+      assert(ctx->Texture._EnabledCoordUnits == 0);	\
+      assert(ctx->Light.ShadeModel==GL_SMOOTH);	\
    }
 #define RENDER_SPAN( span )  _swrast_write_rgba_span(ctx, &span);
 #include "s_tritemp.h"
@@ -137,7 +137,7 @@ _swrast_culltriangle( struct gl_context *ctx,
    const GLubyte *texture = (const GLubyte *) swImg->ImageSlices[0];	\
    const GLint smask = texImg->Width - 1;				\
    const GLint tmask = texImg->Height - 1;				\
-   ASSERT(texImg->TexFormat == MESA_FORMAT_BGR_UNORM8);			\
+   assert(texImg->TexFormat == MESA_FORMAT_BGR_UNORM8);			\
    if (!rb || !texture) {						\
       return;								\
    }
@@ -195,7 +195,7 @@ _swrast_culltriangle( struct gl_context *ctx,
    const GLubyte *texture = (const GLubyte *) swImg->ImageSlices[0];	\
    const GLint smask = texImg->Width - 1;				\
    const GLint tmask = texImg->Height - 1;				\
-   ASSERT(texImg->TexFormat == MESA_FORMAT_BGR_UNORM8);			\
+   assert(texImg->TexFormat == MESA_FORMAT_BGR_UNORM8);			\
    if (!rb || !texture) {						\
       return;								\
    }
@@ -513,7 +513,7 @@ affine_span(struct gl_context *ctx, SWspan *span,
       break;
    }
    span->interpMask &= ~SPAN_RGBA;
-   ASSERT(span->arrayMask & SPAN_RGBA);
+   assert(span->arrayMask & SPAN_RGBA);
 
    _swrast_write_rgba_span(ctx, span);
 
@@ -783,7 +783,7 @@ fast_persp_span(struct gl_context *ctx, SWspan *span,
       break;
    }
    
-   ASSERT(span->arrayMask & SPAN_RGBA);
+   assert(span->arrayMask & SPAN_RGBA);
    _swrast_write_rgba_span(ctx, span);
 
 #undef SPAN_NEAREST
@@ -885,9 +885,9 @@ fast_persp_span(struct gl_context *ctx, SWspan *span,
    struct gl_renderbuffer *rb =                                         \
       ctx->DrawBuffer->Attachment[BUFFER_DEPTH].Renderbuffer;           \
    struct gl_query_object *q = ctx->Query.CurrentOcclusionObject;	\
-   ASSERT(ctx->Depth.Test);						\
-   ASSERT(!ctx->Depth.Mask);						\
-   ASSERT(ctx->Depth.Func == GL_LESS);					\
+   assert(ctx->Depth.Test);						\
+   assert(!ctx->Depth.Mask);						\
+   assert(ctx->Depth.Func == GL_LESS);					\
    assert(rb->Format == MESA_FORMAT_Z_UNORM16);                               \
    if (!q) {								\
       return;								\
@@ -1015,7 +1015,7 @@ _swrast_choose_triangle( struct gl_context *ctx )
 
       if (ctx->Polygon.SmoothFlag) {
          _swrast_set_aa_triangle_function(ctx);
-         ASSERT(swrast->Triangle);
+         assert(swrast->Triangle);
          return;
       }
 
@@ -1134,8 +1134,8 @@ _swrast_choose_triangle( struct gl_context *ctx )
          }
       }
       else {
-         ASSERT(!swrast->_FogEnabled);
-         ASSERT(!_mesa_need_secondary_color(ctx));
+         assert(!swrast->_FogEnabled);
+         assert(!_mesa_need_secondary_color(ctx));
 	 if (ctx->Light.ShadeModel==GL_SMOOTH) {
 	    /* smooth shaded, no texturing, stippled or some raster ops */
 #if CHAN_BITS != 8
diff --git a/mesalib/src/mesa/swrast/s_tritemp.h b/mesalib/src/mesa/swrast/s_tritemp.h
index 8a278dff6..fb73b2d59 100644
--- a/mesalib/src/mesa/swrast/s_tritemp.h
+++ b/mesalib/src/mesa/swrast/s_tritemp.h
@@ -91,6 +91,11 @@
  */
 
 
+#ifndef MAX_GLUINT
+#define MAX_GLUINT	0xffffffff
+#endif
+
+
 /*
  * Some code we unfortunately need to prevent negative interpolated colors.
  */
@@ -380,7 +385,7 @@ static void NAME(struct gl_context *ctx, const SWvertex *v0,
 #  endif /* INTERP_ALPHA */
       }
       else {
-         ASSERT(ctx->Light.ShadeModel == GL_FLAT);
+         assert(ctx->Light.ShadeModel == GL_FLAT);
          span.interpMask |= SPAN_FLAT;
          span.attrStepX[VARYING_SLOT_COL0][0] = span.attrStepY[VARYING_SLOT_COL0][0] = 0.0F;
          span.attrStepX[VARYING_SLOT_COL0][1] = span.attrStepY[VARYING_SLOT_COL0][1] = 0.0F;
@@ -662,7 +667,7 @@ static void NAME(struct gl_context *ctx, const SWvertex *v0,
 #  endif
                }
                else {
-                  ASSERT(ctx->Light.ShadeModel == GL_FLAT);
+                  assert(ctx->Light.ShadeModel == GL_FLAT);
                   rLeft = ChanToFixed(v2->color[RCOMP]);
                   gLeft = ChanToFixed(v2->color[GCOMP]);
                   bLeft = ChanToFixed(v2->color[BCOMP]);
diff --git a/mesalib/src/mesa/swrast/s_zoom.c b/mesalib/src/mesa/swrast/s_zoom.c
index 352c61ad2..ab22652c7 100644
--- a/mesalib/src/mesa/swrast/s_zoom.c
+++ b/mesalib/src/mesa/swrast/s_zoom.c
@@ -53,8 +53,8 @@ compute_zoomed_bounds(struct gl_context *ctx, GLint imageX, GLint imageY,
    const struct gl_framebuffer *fb = ctx->DrawBuffer;
    GLint c0, c1, r0, r1;
 
-   ASSERT(spanX >= imageX);
-   ASSERT(spanY >= imageY);
+   assert(spanX >= imageX);
+   assert(spanY >= imageY);
 
    /*
     * Compute destination columns: [c0, c1)
@@ -149,12 +149,12 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
    }
 
    zoomedWidth = x1 - x0;
-   ASSERT(zoomedWidth > 0);
-   ASSERT(zoomedWidth <= SWRAST_MAX_WIDTH);
+   assert(zoomedWidth > 0);
+   assert(zoomedWidth <= SWRAST_MAX_WIDTH);
 
    /* no pixel arrays! must be horizontal spans. */
-   ASSERT((span->arrayMask & SPAN_XY) == 0);
-   ASSERT(span->primitive == GL_BITMAP);
+   assert((span->arrayMask & SPAN_XY) == 0);
+   assert(span->primitive == GL_BITMAP);
 
    INIT_SPAN(zoomed, GL_BITMAP);
    zoomed.x = x0;
@@ -184,7 +184,7 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
       zoomed.interpMask = span->interpMask & ~SPAN_RGBA;
       zoomed.arrayMask |= SPAN_RGBA;
       zoomed.arrayAttribs |= VARYING_BIT_COL0;  /* we'll produce these values */
-      ASSERT(span->arrayMask & SPAN_RGBA);
+      assert(span->arrayMask & SPAN_RGBA);
    }
    else if (format == GL_DEPTH_COMPONENT) {
       /* Copy color info */
@@ -199,7 +199,7 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
       /* we'll generate an array of depth values */
       zoomed.interpMask = span->interpMask & ~SPAN_Z;
       zoomed.arrayMask |= SPAN_Z;
-      ASSERT(span->arrayMask & SPAN_Z);
+      assert(span->arrayMask & SPAN_Z);
    }
    else {
       _mesa_problem(ctx, "Bad format in zoom_span");
@@ -213,8 +213,8 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
          GLint i;
          for (i = 0; i < zoomedWidth; i++) {
             GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - span->x;
-            ASSERT(j >= 0);
-            ASSERT(j < (GLint) span->end);
+            assert(j >= 0);
+            assert(j < (GLint) span->end);
             COPY_4UBV(zoomed.array->rgba8[i], rgba[j]);
          }
       }
@@ -223,8 +223,8 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
          GLint i;
          for (i = 0; i < zoomedWidth; i++) {
             GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - span->x;
-            ASSERT(j >= 0);
-            ASSERT(j < (GLint) span->end);
+            assert(j >= 0);
+            assert(j < (GLint) span->end);
             COPY_4V(zoomed.array->rgba16[i], rgba[j]);
          }
       }
@@ -233,8 +233,8 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
          GLint i;
          for (i = 0; i < zoomedWidth; i++) {
             GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - span->x;
-            ASSERT(j >= 0);
-            ASSERT(j < (GLint) span->end);
+            assert(j >= 0);
+            assert(j < (GLint) span->end);
             COPY_4V(zoomed.array->attribs[VARYING_SLOT_COL0][i], rgba[j]);
          }
       }
@@ -245,8 +245,8 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
          GLint i;
          for (i = 0; i < zoomedWidth; i++) {
             GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - span->x;
-            ASSERT(j >= 0);
-            ASSERT(j < (GLint) span->end);
+            assert(j >= 0);
+            assert(j < (GLint) span->end);
             zoomed.array->rgba8[i][0] = rgb[j][0];
             zoomed.array->rgba8[i][1] = rgb[j][1];
             zoomed.array->rgba8[i][2] = rgb[j][2];
@@ -258,8 +258,8 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
          GLint i;
          for (i = 0; i < zoomedWidth; i++) {
             GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - span->x;
-            ASSERT(j >= 0);
-            ASSERT(j < (GLint) span->end);
+            assert(j >= 0);
+            assert(j < (GLint) span->end);
             zoomed.array->rgba16[i][0] = rgb[j][0];
             zoomed.array->rgba16[i][1] = rgb[j][1];
             zoomed.array->rgba16[i][2] = rgb[j][2];
@@ -271,8 +271,8 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
          GLint i;
          for (i = 0; i < zoomedWidth; i++) {
             GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - span->x;
-            ASSERT(j >= 0);
-            ASSERT(j < (GLint) span->end);
+            assert(j >= 0);
+            assert(j < (GLint) span->end);
             zoomed.array->attribs[VARYING_SLOT_COL0][i][0] = rgb[j][0];
             zoomed.array->attribs[VARYING_SLOT_COL0][i][1] = rgb[j][1];
             zoomed.array->attribs[VARYING_SLOT_COL0][i][2] = rgb[j][2];
@@ -285,8 +285,8 @@ zoom_span( struct gl_context *ctx, GLint imgX, GLint imgY, const SWspan *span,
       GLint i;
       for (i = 0; i < zoomedWidth; i++) {
          GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - span->x;
-         ASSERT(j >= 0);
-         ASSERT(j < (GLint) span->end);
+         assert(j >= 0);
+         assert(j < (GLint) span->end);
          zoomed.array->z[i] = zValues[j];
       }
       /* Now, fall into the RGB path below */
@@ -372,8 +372,8 @@ _swrast_write_zoomed_stencil_span(struct gl_context *ctx, GLint imgX, GLint imgY
    }
 
    zoomedWidth = x1 - x0;
-   ASSERT(zoomedWidth > 0);
-   ASSERT(zoomedWidth <= SWRAST_MAX_WIDTH);
+   assert(zoomedWidth > 0);
+   assert(zoomedWidth <= SWRAST_MAX_WIDTH);
 
    zoomedVals = malloc(zoomedWidth * sizeof(GLubyte));
    if (!zoomedVals)
@@ -382,8 +382,8 @@ _swrast_write_zoomed_stencil_span(struct gl_context *ctx, GLint imgX, GLint imgY
    /* zoom the span horizontally */
    for (i = 0; i < zoomedWidth; i++) {
       GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - spanX;
-      ASSERT(j >= 0);
-      ASSERT(j < width);
+      assert(j >= 0);
+      assert(j < width);
       zoomedVals[i] = stencil[j];
    }
 
@@ -417,8 +417,8 @@ _swrast_write_zoomed_z_span(struct gl_context *ctx, GLint imgX, GLint imgY,
    }
 
    zoomedWidth = x1 - x0;
-   ASSERT(zoomedWidth > 0);
-   ASSERT(zoomedWidth <= SWRAST_MAX_WIDTH);
+   assert(zoomedWidth > 0);
+   assert(zoomedWidth <= SWRAST_MAX_WIDTH);
 
    zoomedVals = malloc(zoomedWidth * sizeof(GLuint));
    if (!zoomedVals)
@@ -427,8 +427,8 @@ _swrast_write_zoomed_z_span(struct gl_context *ctx, GLint imgX, GLint imgY,
    /* zoom the span horizontally */
    for (i = 0; i < zoomedWidth; i++) {
       GLint j = unzoom_x(ctx->Pixel.ZoomX, imgX, x0 + i) - spanX;
-      ASSERT(j >= 0);
-      ASSERT(j < width);
+      assert(j >= 0);
+      assert(j < width);
       zoomedVals[i] = zVals[j];
    }
 
diff --git a/mesalib/src/mesa/swrast_setup/ss_triangle.c b/mesalib/src/mesa/swrast_setup/ss_triangle.c
index 42ba891ab..483c41592 100644
--- a/mesalib/src/mesa/swrast_setup/ss_triangle.c
+++ b/mesalib/src/mesa/swrast_setup/ss_triangle.c
@@ -25,6 +25,7 @@
  *    Keith Whitwell <keithw@vmware.com>
  */
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/colormac.h"
 #include "main/macros.h"
diff --git a/mesalib/src/mesa/swrast_setup/ss_tritmp.h b/mesalib/src/mesa/swrast_setup/ss_tritmp.h
index d4513c9ac..c38c76a4a 100644
--- a/mesalib/src/mesa/swrast_setup/ss_tritmp.h
+++ b/mesalib/src/mesa/swrast_setup/ss_tritmp.h
@@ -142,8 +142,8 @@ static void TAG(triangle)(struct gl_context *ctx, GLuint e0, GLuint e1, GLuint e
 	    const GLfloat ez = z[0] - z[2];
 	    const GLfloat fz = z[1] - z[2];
 	    const GLfloat oneOverArea = 1.0F / cc;
-	    const GLfloat dzdx = FABSF((ey * fz - ez * fy) * oneOverArea);
-	    const GLfloat dzdy = FABSF((ez * fx - ex * fz) * oneOverArea);
+	    const GLfloat dzdx = fabsf((ey * fz - ez * fy) * oneOverArea);
+	    const GLfloat dzdy = fabsf((ez * fx - ex * fz) * oneOverArea);
 	    offset += MAX2(dzdx, dzdy) * ctx->Polygon.OffsetFactor;
 	 }
          /* new Z values */
diff --git a/mesalib/src/mesa/tnl/t_context.c b/mesalib/src/mesa/tnl/t_context.c
index eb5bae41d..bc705d7a3 100644
--- a/mesalib/src/mesa/tnl/t_context.c
+++ b/mesalib/src/mesa/tnl/t_context.c
@@ -130,7 +130,7 @@ _tnl_InvalidateState( struct gl_context *ctx, GLuint new_state )
    GLuint i;
 
    if (new_state & (_NEW_HINT | _NEW_PROGRAM)) {
-      ASSERT(tnl->AllowVertexFog || tnl->AllowPixelFog);
+      assert(tnl->AllowVertexFog || tnl->AllowPixelFog);
       tnl->_DoVertexFog = ((tnl->AllowVertexFog && (ctx->Hint.Fog != GL_NICEST))
          || !tnl->AllowPixelFog) && !fp;
    }
diff --git a/mesalib/src/mesa/tnl/t_draw.c b/mesalib/src/mesa/tnl/t_draw.c
index be3f059bb..60265d60b 100644
--- a/mesalib/src/mesa/tnl/t_draw.c
+++ b/mesalib/src/mesa/tnl/t_draw.c
@@ -25,6 +25,8 @@
  *    Keith Whitwell <keithw@vmware.com>
  */
 
+#include <stdio.h>
+
 #include "main/glheader.h"
 #include "main/bufferobj.h"
 #include "main/condrender.h"
diff --git a/mesalib/src/mesa/tnl/t_pipeline.c b/mesalib/src/mesa/tnl/t_pipeline.c
index 78508aecf..0adbef012 100644
--- a/mesalib/src/mesa/tnl/t_pipeline.c
+++ b/mesalib/src/mesa/tnl/t_pipeline.c
@@ -112,6 +112,84 @@ static GLuint check_output_changes( struct gl_context *ctx )
 #endif
 }
 
+/**
+ * START/END_FAST_MATH macros:
+ *
+ * START_FAST_MATH: Set x86 FPU to faster, 32-bit precision mode (and save
+ *                  original mode to a temporary).
+ * END_FAST_MATH: Restore x86 FPU to original mode.
+ */
+#if defined(__GNUC__) && defined(__i386__)
+/*
+ * Set the x86 FPU control word to guarentee only 32 bits of precision
+ * are stored in registers.  Allowing the FPU to store more introduces
+ * differences between situations where numbers are pulled out of memory
+ * vs. situations where the compiler is able to optimize register usage.
+ *
+ * In the worst case, we force the compiler to use a memory access to
+ * truncate the float, by specifying the 'volatile' keyword.
+ */
+/* Hardware default: All exceptions masked, extended double precision,
+ * round to nearest (IEEE compliant):
+ */
+#define DEFAULT_X86_FPU		0x037f
+/* All exceptions masked, single precision, round to nearest:
+ */
+#define FAST_X86_FPU		0x003f
+/* The fldcw instruction will cause any pending FP exceptions to be
+ * raised prior to entering the block, and we clear any pending
+ * exceptions before exiting the block.  Hence, asm code has free
+ * reign over the FPU while in the fast math block.
+ */
+#if defined(NO_FAST_MATH)
+#define START_FAST_MATH(x)						\
+do {									\
+   static GLuint mask = DEFAULT_X86_FPU;				\
+   __asm__ ( "fnstcw %0" : "=m" (*&(x)) );				\
+   __asm__ ( "fldcw %0" : : "m" (mask) );				\
+} while (0)
+#else
+#define START_FAST_MATH(x)						\
+do {									\
+   static GLuint mask = FAST_X86_FPU;					\
+   __asm__ ( "fnstcw %0" : "=m" (*&(x)) );				\
+   __asm__ ( "fldcw %0" : : "m" (mask) );				\
+} while (0)
+#endif
+/* Restore original FPU mode, and clear any exceptions that may have
+ * occurred in the FAST_MATH block.
+ */
+#define END_FAST_MATH(x)						\
+do {									\
+   __asm__ ( "fnclex ; fldcw %0" : : "m" (*&(x)) );			\
+} while (0)
+
+#elif defined(_MSC_VER) && defined(_M_IX86)
+#define DEFAULT_X86_FPU		0x037f /* See GCC comments above */
+#define FAST_X86_FPU		0x003f /* See GCC comments above */
+#if defined(NO_FAST_MATH)
+#define START_FAST_MATH(x) do {\
+	static GLuint mask = DEFAULT_X86_FPU;\
+	__asm fnstcw word ptr [x]\
+	__asm fldcw word ptr [mask]\
+} while(0)
+#else
+#define START_FAST_MATH(x) do {\
+	static GLuint mask = FAST_X86_FPU;\
+	__asm fnstcw word ptr [x]\
+	__asm fldcw word ptr [mask]\
+} while(0)
+#endif
+#define END_FAST_MATH(x) do {\
+	__asm fnclex\
+	__asm fldcw word ptr [x]\
+} while(0)
+
+#else
+#define START_FAST_MATH(x)  x = 0
+#define END_FAST_MATH(x)  (void)(x)
+#endif
+
 
 void _tnl_run_pipeline( struct gl_context *ctx )
 {
diff --git a/mesalib/src/mesa/tnl/t_rasterpos.c b/mesalib/src/mesa/tnl/t_rasterpos.c
index 2f52bb306..9ecf947df 100644
--- a/mesalib/src/mesa/tnl/t_rasterpos.c
+++ b/mesalib/src/mesa/tnl/t_rasterpos.c
@@ -23,6 +23,7 @@
  */
 
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/colormac.h"
 #include "main/feedback.h"
@@ -271,7 +272,7 @@ compute_texgen(struct gl_context *ctx, const GLfloat vObj[4], const GLfloat vEye
    rz = u[2] - normal[2] * two_nu;
    m = rx * rx + ry * ry + (rz + 1.0F) * (rz + 1.0F);
    if (m > 0.0F)
-      mInv = 0.5F * INV_SQRTF(m);
+      mInv = 0.5F * (1.0f / sqrtf(m));
    else
       mInv = 0.0F;
 
diff --git a/mesalib/src/mesa/tnl/t_vb_cliptmp.h b/mesalib/src/mesa/tnl/t_vb_cliptmp.h
index 7dafb83cf..12181f085 100644
--- a/mesalib/src/mesa/tnl/t_vb_cliptmp.h
+++ b/mesalib/src/mesa/tnl/t_vb_cliptmp.h
@@ -155,7 +155,7 @@ TAG(clip_line)( struct gl_context *ctx, GLuint v0, GLuint v1, GLubyte mask )
       newvert++;
    }
    else {
-      ASSERT(t0 == 0.0);
+      assert(t0 == 0.0);
    }
 
    /* Note: we need to use vertex v0_orig when computing the new
@@ -174,7 +174,7 @@ TAG(clip_line)( struct gl_context *ctx, GLuint v0, GLuint v1, GLubyte mask )
       newvert++;
    }
    else {
-      ASSERT(t1 == 0.0);
+      assert(t1 == 0.0);
    }
 
    tnl->Driver.Render.ClippedLine( ctx, v0, v1 );
@@ -239,7 +239,7 @@ TAG(clip_tri)( struct gl_context *ctx, GLuint v0, GLuint v1, GLuint v2, GLubyte
 
    if (ctx->Light.ShadeModel == GL_FLAT) {
       if (pv != inlist[0]) {
-	 ASSERT( inlist[0] >= VB->Count );
+	 assert( inlist[0] >= VB->Count );
 	 tnl->Driver.Render.CopyPV( ctx, inlist[0], pv );
       }
    }
@@ -302,7 +302,7 @@ TAG(clip_quad)( struct gl_context *ctx, GLuint v0, GLuint v1, GLuint v2, GLuint
 
    if (ctx->Light.ShadeModel == GL_FLAT) {
       if (pv != inlist[0]) {
-	 ASSERT( inlist[0] >= VB->Count );
+	 assert( inlist[0] >= VB->Count );
 	 tnl->Driver.Render.CopyPV( ctx, inlist[0], pv );
       }
    }
diff --git a/mesalib/src/mesa/tnl/t_vb_fog.c b/mesalib/src/mesa/tnl/t_vb_fog.c
index ab7901edb..3626f1da3 100644
--- a/mesalib/src/mesa/tnl/t_vb_fog.c
+++ b/mesalib/src/mesa/tnl/t_vb_fog.c
@@ -26,6 +26,7 @@
  */
 
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/colormac.h"
 #include "main/macros.h"
@@ -78,7 +79,7 @@ init_static_data( void )
    GLfloat f = 0.0F;
    GLint i = 0;
    for ( ; i < FOG_EXP_TABLE_SIZE ; i++, f += FOG_INCR) {
-      exp_table[i] = EXPF(-f);
+      exp_table[i] = expf(-f);
    }
    inited = 1;
 }
@@ -186,7 +187,7 @@ run_fog_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage)
 	    NOTE should avoid going through array twice */
 	 coord = input->start;
 	 for (i = 0; i < input->count; i++) {
-	    *coord = FABSF(*coord);
+	    *coord = fabsf(*coord);
 	    STRIDE_F(coord, input->stride);
 	 }
       }
@@ -201,7 +202,7 @@ run_fog_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage)
 	 input->count = VB->EyePtr->count;
 	 coord = VB->EyePtr->start;
 	 for (i = 0 ; i < VB->EyePtr->count; i++) {
-	    input->data[i][0] = FABSF(coord[2]);
+	    input->data[i][0] = fabsf(coord[2]);
 	    STRIDE_F(coord, VB->EyePtr->stride);
 	 }
       }
diff --git a/mesalib/src/mesa/tnl/t_vb_light.c b/mesalib/src/mesa/tnl/t_vb_light.c
index f6884a464..7781b6a6c 100644
--- a/mesalib/src/mesa/tnl/t_vb_light.c
+++ b/mesalib/src/mesa/tnl/t_vb_light.c
@@ -23,7 +23,7 @@
  */
 
 
-
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/colormac.h"
 #include "main/light.h"
@@ -123,7 +123,7 @@ validate_shine_table( struct gl_context *ctx, GLuint side, GLfloat shininess )
    struct tnl_shine_tab *list = tnl->_ShineTabList;
    struct tnl_shine_tab *s;
 
-   ASSERT(side < 2);
+   assert(side < 2);
 
    foreach(s, list)
       if ( s->shininess == shininess )
diff --git a/mesalib/src/mesa/tnl/t_vb_points.c b/mesalib/src/mesa/tnl/t_vb_points.c
index 5f6c25857..0f8578daa 100644
--- a/mesalib/src/mesa/tnl/t_vb_points.c
+++ b/mesalib/src/mesa/tnl/t_vb_points.c
@@ -25,6 +25,7 @@
  *    Brian Paul
  */
 
+#include "c99_math.h"
 #include "main/glheader.h"
 #include "main/mtypes.h"
 #include "main/dd.h"
@@ -62,9 +63,9 @@ run_point_stage(struct gl_context *ctx, struct tnl_pipeline_stage *stage)
       GLuint i;
 
       for (i = 0; i < VB->Count; i++) {
-         const GLfloat dist = FABSF(*eyeCoord);
+         const GLfloat dist = fabsf(*eyeCoord);
          const GLfloat q = p0 + dist * (p1 + dist * p2);
-         const GLfloat atten = (q != 0.0F) ? INV_SQRTF(q) : 1.0F;
+         const GLfloat atten = (q != 0.0F) ? (1.0f / sqrtf(q)) : 1.0F;
          size[i][0] = pointSize * atten; /* clamping done in rasterization */
          eyeCoord += eyeCoordStride;
       }
diff --git a/mesalib/src/mesa/tnl/t_vb_program.c b/mesalib/src/mesa/tnl/t_vb_program.c
index d08abe7c2..464a4cddd 100644
--- a/mesalib/src/mesa/tnl/t_vb_program.c
+++ b/mesalib/src/mesa/tnl/t_vb_program.c
@@ -392,7 +392,7 @@ run_vp( struct gl_context *ctx, struct tnl_pipeline_stage *stage )
          store->results[VARYING_SLOT_FOGC].data[i][3] = 1.0;
       }
 #ifdef NAN_CHECK
-      ASSERT(machine->Outputs[0][3] != 0.0F);
+      assert(machine->Outputs[0][3] != 0.0F);
 #endif
 #if 0
       printf("HPOS: %f %f %f %f\n",
diff --git a/mesalib/src/mesa/tnl/t_vb_render.c b/mesalib/src/mesa/tnl/t_vb_render.c
index aff5b9a68..4960ac096 100644
--- a/mesalib/src/mesa/tnl/t_vb_render.c
+++ b/mesalib/src/mesa/tnl/t_vb_render.c
@@ -38,6 +38,7 @@
  */
 
 
+#include <stdio.h>
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/enums.h"
@@ -271,22 +272,22 @@ static GLboolean run_render( struct gl_context *ctx,
     * that window coordinates are guarenteed not to change before
     * rendering.
     */
-   ASSERT(tnl->Driver.Render.Start);
+   assert(tnl->Driver.Render.Start);
 
    tnl->Driver.Render.Start( ctx );
 
-   ASSERT(tnl->Driver.Render.BuildVertices);
-   ASSERT(tnl->Driver.Render.PrimitiveNotify);
-   ASSERT(tnl->Driver.Render.Points);
-   ASSERT(tnl->Driver.Render.Line);
-   ASSERT(tnl->Driver.Render.Triangle);
-   ASSERT(tnl->Driver.Render.Quad);
-   ASSERT(tnl->Driver.Render.ResetLineStipple);
-   ASSERT(tnl->Driver.Render.Interp);
-   ASSERT(tnl->Driver.Render.CopyPV);
-   ASSERT(tnl->Driver.Render.ClippedLine);
-   ASSERT(tnl->Driver.Render.ClippedPolygon);
-   ASSERT(tnl->Driver.Render.Finish);
+   assert(tnl->Driver.Render.BuildVertices);
+   assert(tnl->Driver.Render.PrimitiveNotify);
+   assert(tnl->Driver.Render.Points);
+   assert(tnl->Driver.Render.Line);
+   assert(tnl->Driver.Render.Triangle);
+   assert(tnl->Driver.Render.Quad);
+   assert(tnl->Driver.Render.ResetLineStipple);
+   assert(tnl->Driver.Render.Interp);
+   assert(tnl->Driver.Render.CopyPV);
+   assert(tnl->Driver.Render.ClippedLine);
+   assert(tnl->Driver.Render.ClippedPolygon);
+   assert(tnl->Driver.Render.Finish);
 
    tnl->Driver.Render.BuildVertices( ctx, 0, VB->Count, ~0 );
 
diff --git a/mesalib/src/mesa/tnl/t_vb_texgen.c b/mesalib/src/mesa/tnl/t_vb_texgen.c
index 8f527e343..9a61ef2fe 100644
--- a/mesalib/src/mesa/tnl/t_vb_texgen.c
+++ b/mesalib/src/mesa/tnl/t_vb_texgen.c
@@ -116,7 +116,7 @@ static void build_m3( GLfloat f[][3], GLfloat m[],
       fz = f[i][2] = u[2] - norm[2] * two_nu;
       m[i] = fx * fx + fy * fy + (fz + 1.0F) * (fz + 1.0F);
       if (m[i] != 0.0F) {
-	 m[i] = 0.5F * INV_SQRTF(m[i]);
+	 m[i] = 0.5F * (1.0f / sqrtf(m[i]));
       }
    }
 }
@@ -145,7 +145,7 @@ static void build_m2( GLfloat f[][3], GLfloat m[],
       fz = f[i][2] = u[2] - norm[2] * two_nu;
       m[i] = fx * fx + fy * fy + (fz + 1.0F) * (fz + 1.0F);
       if (m[i] != 0.0F) {
-	 m[i] = 0.5F * INV_SQRTF(m[i]);
+	 m[i] = 0.5F * (1.0f / sqrtf(m[i]));
       }
    }
 }
diff --git a/mesalib/src/mesa/tnl/t_vertex.c b/mesalib/src/mesa/tnl/t_vertex.c
index 90b97a092..369d6d945 100644
--- a/mesalib/src/mesa/tnl/t_vertex.c
+++ b/mesalib/src/mesa/tnl/t_vertex.c
@@ -25,6 +25,7 @@
  *    Keith Whitwell <keithw@vmware.com>
  */
 
+#include <stdio.h>
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/colormac.h"
diff --git a/mesalib/src/mesa/tnl/t_vertex_sse.c b/mesalib/src/mesa/tnl/t_vertex_sse.c
index 93128fbe6..963432c48 100644
--- a/mesalib/src/mesa/tnl/t_vertex_sse.c
+++ b/mesalib/src/mesa/tnl/t_vertex_sse.c
@@ -25,6 +25,8 @@
  *    Keith Whitwell <keithw@vmware.com>
  */
 
+#include <stdio.h>
+
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/colormac.h"
diff --git a/mesalib/src/mesa/vbo/vbo.h b/mesalib/src/mesa/vbo/vbo.h
index 6e1a4aa1a..54dee6c46 100644
--- a/mesalib/src/mesa/vbo/vbo.h
+++ b/mesalib/src/mesa/vbo/vbo.h
@@ -35,6 +35,10 @@
 #include <stdbool.h>
 #include "main/glheader.h"
 
+#ifdef __cplusplus
+extern "C" {
+#endif
+
 struct gl_client_array;
 struct gl_context;
 struct gl_transform_feedback_object;
@@ -228,4 +232,8 @@ _es_VertexAttrib3fv(GLuint indx, const GLfloat* values);
 void GLAPIENTRY
 _es_VertexAttrib4fv(GLuint indx, const GLfloat* values);
 
+#ifdef __cplusplus
+} // extern "C"
+#endif
+
 #endif
diff --git a/mesalib/src/mesa/vbo/vbo_attrib_tmp.h b/mesalib/src/mesa/vbo/vbo_attrib_tmp.h
index 0c44540fc..b1c3d9842 100644
--- a/mesalib/src/mesa/vbo/vbo_attrib_tmp.h
+++ b/mesalib/src/mesa/vbo/vbo_attrib_tmp.h
@@ -28,6 +28,22 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "util/u_format_r11g11b10f.h"
 #include "main/varray.h"
 
+
+/* ATTR */
+#define ATTR( A, N, T, V0, V1, V2, V3 ) \
+        ATTR_##T((A), (N), (T), (V0), (V1), (V2), (V3))
+
+#define ATTR_GL_UNSIGNED_INT( A, N, T, V0, V1, V2, V3 ) \
+    ATTR_UNION(A, N, T, UINT_AS_UNION(V0), UINT_AS_UNION(V1), \
+        UINT_AS_UNION(V2), UINT_AS_UNION(V3))
+#define ATTR_GL_INT( A, N, T, V0, V1, V2, V3 ) \
+    ATTR_UNION(A, N, T, INT_AS_UNION(V0), INT_AS_UNION(V1), \
+        INT_AS_UNION(V2), INT_AS_UNION(V3))
+#define ATTR_GL_FLOAT( A, N, T, V0, V1, V2, V3 ) \
+    ATTR_UNION(A, N, T, FLOAT_AS_UNION(V0), FLOAT_AS_UNION(V1),\
+        FLOAT_AS_UNION(V2), FLOAT_AS_UNION(V3))
+
+
 /* float */
 #define ATTR1FV( A, V ) ATTR( A, 1, GL_FLOAT, (V)[0], 0, 0, 1 )
 #define ATTR2FV( A, V ) ATTR( A, 2, GL_FLOAT, (V)[0], (V)[1], 0, 1 )
@@ -41,8 +57,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 /* int */
 #define ATTRI( A, N, X, Y, Z, W) ATTR( A, N, GL_INT, \
-                                       INT_AS_FLT(X), INT_AS_FLT(Y), \
-                                       INT_AS_FLT(Z), INT_AS_FLT(W) )
+                                       X, Y, Z, W )
 
 #define ATTR2IV( A, V ) ATTRI( A, 2, (V)[0], (V)[1], 0, 1 )
 #define ATTR3IV( A, V ) ATTRI( A, 3, (V)[0], (V)[1], (V)[2], 1 )
@@ -56,8 +71,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 /* uint */
 #define ATTRUI( A, N, X, Y, Z, W) ATTR( A, N, GL_UNSIGNED_INT, \
-                                        UINT_AS_FLT(X), UINT_AS_FLT(Y), \
-                                        UINT_AS_FLT(Z), UINT_AS_FLT(W) )
+                                        X, Y, Z, W )
 
 #define ATTR2UIV( A, V ) ATTRUI( A, 2, (V)[0], (V)[1], 0, 1 )
 #define ATTR3UIV( A, V ) ATTRUI( A, 3, (V)[0], (V)[1], (V)[2], 1 )
diff --git a/mesalib/src/mesa/vbo/vbo_context.c b/mesalib/src/mesa/vbo/vbo_context.c
index 19712aa58..fd1ffe2f7 100644
--- a/mesalib/src/mesa/vbo/vbo_context.c
+++ b/mesalib/src/mesa/vbo/vbo_context.c
@@ -108,7 +108,7 @@ static void init_mat_currval(struct gl_context *ctx)
       &vbo->currval[VBO_ATTRIB_MAT_FRONT_AMBIENT];
    GLuint i;
 
-   ASSERT(NR_MAT_ATTRIBS == MAT_ATTRIB_MAX);
+   assert(NR_MAT_ATTRIBS == MAT_ATTRIB_MAX);
 
    memset(arrays, 0, sizeof(*arrays) * NR_MAT_ATTRIBS);
 
@@ -172,14 +172,14 @@ GLboolean _vbo_CreateContext( struct gl_context *ctx )
       GLuint i;
 
       /* identity mapping */
-      for (i = 0; i < Elements(vbo->map_vp_none); i++) 
+      for (i = 0; i < ARRAY_SIZE(vbo->map_vp_none); i++) 
 	 vbo->map_vp_none[i] = i;
       /* map material attribs to generic slots */
       for (i = 0; i < NR_MAT_ATTRIBS; i++) 
 	 vbo->map_vp_none[VERT_ATTRIB_GENERIC(i)]
             = VBO_ATTRIB_MAT_FRONT_AMBIENT + i;
 
-      for (i = 0; i < Elements(vbo->map_vp_arb); i++)
+      for (i = 0; i < ARRAY_SIZE(vbo->map_vp_arb); i++)
 	 vbo->map_vp_arb[i] = i;
    }
 
diff --git a/mesalib/src/mesa/vbo/vbo_context.h b/mesalib/src/mesa/vbo/vbo_context.h
index e22451305..6099b56e6 100644
--- a/mesalib/src/mesa/vbo/vbo_context.h
+++ b/mesalib/src/mesa/vbo/vbo_context.h
@@ -56,6 +56,11 @@
 #include "vbo_exec.h"
 #include "vbo_save.h"
 
+#include "main/macros.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
 
 struct vbo_context {
    struct gl_client_array currval[VBO_ATTRIB_MAX];
@@ -124,7 +129,7 @@ vbo_draw_method(struct vbo_context *vbo, gl_draw_method method)
          ctx->Array._DrawArrays = vbo->save.inputs;
          break;
       default:
-         ASSERT(0);
+         assert(0);
       }
 
       ctx->NewDriverState |= ctx->DriverFlags.NewArray;
@@ -146,33 +151,36 @@ vbo_attrtype_to_integer_flag(GLenum format)
    case GL_UNSIGNED_INT:
       return GL_TRUE;
    default:
-      ASSERT(0);
+      assert(0);
       return GL_FALSE;
    }
 }
 
-
 /**
  * Return default component values for the given format.
- * The return type is an array of floats, because that's how we declare
- * the vertex storage despite the fact we sometimes store integers in there.
+ * The return type is an array of fi_types, because that's how we declare
+ * the vertex storage : floats , integers or unsigned integers.
  */
-static inline const GLfloat *
-vbo_get_default_vals_as_float(GLenum format)
+static inline const fi_type *
+vbo_get_default_vals_as_union(GLenum format)
 {
    static const GLfloat default_float[4] = { 0, 0, 0, 1 };
    static const GLint default_int[4] = { 0, 0, 0, 1 };
 
    switch (format) {
    case GL_FLOAT:
-      return default_float;
+      return (fi_type *)default_float;
    case GL_INT:
    case GL_UNSIGNED_INT:
-      return (const GLfloat*)default_int;
+      return (fi_type *)default_int;
    default:
-      ASSERT(0);
+      assert(0);
       return NULL;
    }
 }
 
+#ifdef __cplusplus
+} // extern "C"
+#endif
+
 #endif
diff --git a/mesalib/src/mesa/vbo/vbo_exec.h b/mesalib/src/mesa/vbo/vbo_exec.h
index bb265def8..f17fe684c 100644
--- a/mesalib/src/mesa/vbo/vbo_exec.h
+++ b/mesalib/src/mesa/vbo/vbo_exec.h
@@ -38,6 +38,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "vbo.h"
 #include "vbo_attrib.h"
 
+#include "main/imports.h"
 
 /**
  * Max number of primitives (number of glBegin/End pairs) per VBO.
@@ -71,7 +72,7 @@ struct vbo_exec_eval2_map {
 
 
 struct vbo_exec_copied_vtx {
-   GLfloat buffer[VBO_ATTRIB_MAX * 4 * VBO_MAX_COPIED_VERTS];
+   fi_type buffer[VBO_ATTRIB_MAX * 4 * VBO_MAX_COPIED_VERTS];
    GLuint nr;
 };
 
@@ -91,10 +92,10 @@ struct vbo_exec_context
       struct _mesa_prim prim[VBO_MAX_PRIM];
       GLuint prim_count;
 
-      GLfloat *buffer_map;
-      GLfloat *buffer_ptr;              /* cursor, points into buffer */
+      fi_type *buffer_map;
+      fi_type *buffer_ptr;              /* cursor, points into buffer */
       GLuint   buffer_used;             /* in bytes */
-      GLfloat vertex[VBO_ATTRIB_MAX*4]; /* current vertex */
+      fi_type vertex[VBO_ATTRIB_MAX*4]; /* current vertex */
 
       GLuint vert_count;
       GLuint max_vert;
@@ -104,7 +105,7 @@ struct vbo_exec_context
       GLenum attrtype[VBO_ATTRIB_MAX];
       GLubyte active_sz[VBO_ATTRIB_MAX];
 
-      GLfloat *attrptr[VBO_ATTRIB_MAX]; 
+      fi_type *attrptr[VBO_ATTRIB_MAX];
       struct gl_client_array arrays[VERT_ATTRIB_MAX];
 
       /* According to program mode, the values above plus current
diff --git a/mesalib/src/mesa/vbo/vbo_exec_api.c b/mesalib/src/mesa/vbo/vbo_exec_api.c
index 5f8250e78..9669abe7d 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_api.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_api.c
@@ -115,7 +115,7 @@ static void vbo_exec_wrap_buffers( struct vbo_exec_context *exec )
  */
 void vbo_exec_vtx_wrap( struct vbo_exec_context *exec )
 {
-   GLfloat *data = exec->vtx.copied.buffer;
+   fi_type *data = exec->vtx.copied.buffer;
    GLuint i;
 
    /* Run pipeline on current vertices, copy wrapped vertices
@@ -159,9 +159,9 @@ static void vbo_exec_copy_to_current( struct vbo_exec_context *exec )
           * ctx->Current.Attrib and ctx->Light.Material.Attrib arrays.
           */
 	 GLfloat *current = (GLfloat *)vbo->currval[i].Ptr;
-         GLfloat tmp[4];
+         fi_type tmp[4];
 
-         COPY_CLEAN_4V_TYPE_AS_FLOAT(tmp,
+         COPY_CLEAN_4V_TYPE_AS_UNION(tmp,
                                      exec->vtx.attrsz[i],
                                      exec->vtx.attrptr[i],
                                      exec->vtx.attrtype[i]);
@@ -214,7 +214,7 @@ vbo_exec_copy_from_current(struct vbo_exec_context *exec)
    GLint i;
 
    for (i = VBO_ATTRIB_POS + 1; i < VBO_ATTRIB_MAX; i++) {
-      const GLfloat *current = (GLfloat *) vbo->currval[i].Ptr;
+      const fi_type *current = (fi_type *) vbo->currval[i].Ptr;
       switch (exec->vtx.attrsz[i]) {
       case 4: exec->vtx.attrptr[i][3] = current[3];
       case 3: exec->vtx.attrptr[i][2] = current[2];
@@ -240,7 +240,7 @@ vbo_exec_wrap_upgrade_vertex(struct vbo_exec_context *exec,
    struct gl_context *ctx = exec->ctx;
    struct vbo_context *vbo = vbo_context(ctx);
    const GLint lastcount = exec->vtx.vert_count;
-   GLfloat *old_attrptr[VBO_ATTRIB_MAX];
+   fi_type *old_attrptr[VBO_ATTRIB_MAX];
    const GLuint old_vtx_size = exec->vtx.vertex_size; /* floats per vertex */
    const GLuint oldSize = exec->vtx.attrsz[attr];
    GLuint i;
@@ -287,7 +287,7 @@ vbo_exec_wrap_upgrade_vertex(struct vbo_exec_context *exec,
    if (unlikely(oldSize)) {
       /* Size changed, recalculate all the attrptr[] values
        */
-      GLfloat *tmp = exec->vtx.vertex;
+      fi_type *tmp = exec->vtx.vertex;
 
       for (i = 0 ; i < VBO_ATTRIB_MAX ; i++) {
 	 if (exec->vtx.attrsz[i]) {
@@ -306,7 +306,7 @@ vbo_exec_wrap_upgrade_vertex(struct vbo_exec_context *exec,
    else {
       /* Just have to append the new attribute at the end */
       exec->vtx.attrptr[attr] = exec->vtx.vertex +
-	 exec->vtx.vertex_size - newSize;
+        exec->vtx.vertex_size - newSize;
    }
 
    /* Replay stored vertices to translate them
@@ -315,8 +315,8 @@ vbo_exec_wrap_upgrade_vertex(struct vbo_exec_context *exec,
     * -- No need to replay - just copy piecewise
     */
    if (unlikely(exec->vtx.copied.nr)) {
-      GLfloat *data = exec->vtx.copied.buffer;
-      GLfloat *dest = exec->vtx.buffer_ptr;
+      fi_type *data = exec->vtx.copied.buffer;
+      fi_type *dest = exec->vtx.buffer_ptr;
       GLuint j;
 
       assert(exec->vtx.buffer_ptr == exec->vtx.buffer_map);
@@ -331,13 +331,13 @@ vbo_exec_wrap_upgrade_vertex(struct vbo_exec_context *exec,
 
 	       if (j == attr) {
 		  if (oldSize) {
-		     GLfloat tmp[4];
-                     COPY_CLEAN_4V_TYPE_AS_FLOAT(tmp, oldSize,
+		     fi_type tmp[4];
+		     COPY_CLEAN_4V_TYPE_AS_UNION(tmp, oldSize,
                                                  data + old_offset,
                                                  exec->vtx.attrtype[j]);
 		     COPY_SZ_4V(dest + new_offset, newSize, tmp);
 		  } else {
-		     GLfloat *current = (GLfloat *)vbo->currval[j].Ptr;
+		     fi_type *current = (fi_type *)vbo->currval[j].Ptr;
 		     COPY_SZ_4V(dest + new_offset, sz, current);
 		  }
 	       }
@@ -376,14 +376,14 @@ vbo_exec_fixup_vertex(struct gl_context *ctx, GLuint attr, GLuint newSize)
    }
    else if (newSize < exec->vtx.active_sz[attr]) {
       GLuint i;
-      const GLfloat *id =
-            vbo_get_default_vals_as_float(exec->vtx.attrtype[attr]);
+      const fi_type *id =
+            vbo_get_default_vals_as_union(exec->vtx.attrtype[attr]);
 
       /* New size is smaller - just need to fill in some
        * zeros.  Don't need to flush or wrap.
        */
       for (i = newSize; i <= exec->vtx.attrsz[attr]; i++)
-	 exec->vtx.attrptr[attr][i-1] = id[i-1];
+        exec->vtx.attrptr[attr][i-1] = id[i-1];
    }
 
    exec->vtx.active_sz[attr] = newSize;
@@ -401,7 +401,7 @@ vbo_exec_fixup_vertex(struct gl_context *ctx, GLuint attr, GLuint newSize)
  * This macro is used to implement all the glVertex, glColor, glTexCoord,
  * glVertexAttrib, etc functions.
  */
-#define ATTR( A, N, T, V0, V1, V2, V3 )					\
+#define ATTR_UNION( A, N, T, V0, V1, V2, V3 )				\
 do {									\
    struct vbo_exec_context *exec = &vbo_context(ctx)->exec;		\
 									\
@@ -412,12 +412,12 @@ do {									\
       vbo_exec_fixup_vertex(ctx, A, N);					\
    									\
    {									\
-      GLfloat *dest = exec->vtx.attrptr[A];				\
+      fi_type *dest = exec->vtx.attrptr[A];			\
       if (N>0) dest[0] = V0;						\
       if (N>1) dest[1] = V1;						\
       if (N>2) dest[2] = V2;						\
       if (N>3) dest[3] = V3;						\
-      exec->vtx.attrtype[A] = T;                                        \
+      exec->vtx.attrtype[A] = T;					\
    }									\
 									\
    if ((A) == 0) {							\
@@ -1036,7 +1036,7 @@ void vbo_exec_vtx_init( struct vbo_exec_context *exec )
                                  &exec->vtx.bufferobj,
                                  ctx->Shared->NullBufferObj);
 
-   ASSERT(!exec->vtx.buffer_map);
+   assert(!exec->vtx.buffer_map);
    exec->vtx.buffer_map = _mesa_align_malloc(VBO_VERT_BUFFER_SIZE, 64);
    exec->vtx.buffer_ptr = exec->vtx.buffer_map;
 
@@ -1044,16 +1044,16 @@ void vbo_exec_vtx_init( struct vbo_exec_context *exec )
    _mesa_noop_vtxfmt_init(&exec->vtxfmt_noop);
 
    for (i = 0 ; i < VBO_ATTRIB_MAX ; i++) {
-      ASSERT(i < Elements(exec->vtx.attrsz));
+      assert(i < ARRAY_SIZE(exec->vtx.attrsz));
       exec->vtx.attrsz[i] = 0;
-      ASSERT(i < Elements(exec->vtx.attrtype));
+      assert(i < ARRAY_SIZE(exec->vtx.attrtype));
       exec->vtx.attrtype[i] = GL_FLOAT;
-      ASSERT(i < Elements(exec->vtx.active_sz));
+      assert(i < ARRAY_SIZE(exec->vtx.active_sz));
       exec->vtx.active_sz[i] = 0;
    }
    for (i = 0 ; i < VERT_ATTRIB_MAX; i++) {
-      ASSERT(i < Elements(exec->vtx.inputs));
-      ASSERT(i < Elements(exec->vtx.arrays));
+      assert(i < ARRAY_SIZE(exec->vtx.inputs));
+      assert(i < ARRAY_SIZE(exec->vtx.arrays));
       exec->vtx.inputs[i] = &exec->vtx.arrays[i];
    }
    
@@ -1099,7 +1099,7 @@ void vbo_exec_vtx_destroy( struct vbo_exec_context *exec )
    /* True VBOs should already be unmapped
     */
    if (exec->vtx.buffer_map) {
-      ASSERT(exec->vtx.bufferobj->Name == 0 ||
+      assert(exec->vtx.bufferobj->Name == 0 ||
              exec->vtx.bufferobj->Name == IMM_BUFFER_NAME);
       if (exec->vtx.bufferobj->Name == 0) {
          _mesa_align_free(exec->vtx.buffer_map);
@@ -1110,7 +1110,7 @@ void vbo_exec_vtx_destroy( struct vbo_exec_context *exec )
 
    /* Drop any outstanding reference to the vertex buffer
     */
-   for (i = 0; i < Elements(exec->vtx.arrays); i++) {
+   for (i = 0; i < ARRAY_SIZE(exec->vtx.arrays); i++) {
       _mesa_reference_buffer_object(ctx,
                                     &exec->vtx.arrays[i].BufferObj,
                                     NULL);
diff --git a/mesalib/src/mesa/vbo/vbo_exec_array.c b/mesalib/src/mesa/vbo/vbo_exec_array.c
index c16fe77ee..3ea775c0e 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_array.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_array.c
@@ -26,6 +26,7 @@
  * 
  **************************************************************************/
 
+#include <stdio.h>
 #include "main/glheader.h"
 #include "main/context.h"
 #include "main/state.h"
@@ -181,8 +182,7 @@ vbo_get_minmax_index(struct gl_context *ctx,
       break;
    }
    default:
-      assert(0);
-      break;
+      unreachable("not reached");
    }
 
    if (_mesa_is_bufferobj(ib->obj)) {
@@ -331,7 +331,7 @@ check_draw_elements_data(struct gl_context *ctx, GLsizei count, GLenum elemType,
       }
 
       /* check element j of each enabled array */
-      for (k = 0; k < Elements(vao->_VertexAttrib); k++) {
+      for (k = 0; k < ARRAY_SIZE(vao->_VertexAttrib); k++) {
          check_array_data(ctx, &vao->_VertexAttrib[k], k, j);
       }
    }
@@ -341,7 +341,7 @@ check_draw_elements_data(struct gl_context *ctx, GLsizei count, GLenum elemType,
                               MAP_INTERNAL);
    }
 
-   for (k = 0; k < Elements(vao->_VertexAttrib); k++) {
+   for (k = 0; k < ARRAY_SIZE(vao->_VertexAttrib); k++) {
       unmap_array_buffer(ctx, &vao->_VertexAttrib[k]);
    }
 }
diff --git a/mesalib/src/mesa/vbo/vbo_exec_draw.c b/mesalib/src/mesa/vbo/vbo_exec_draw.c
index 362cc1040..91f2ca43a 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_draw.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_draw.c
@@ -25,6 +25,7 @@
  *    Keith Whitwell <keithw@vmware.com>
  */
 
+#include <stdio.h>
 #include "main/glheader.h"
 #include "main/bufferobj.h"
 #include "main/compiler.h"
@@ -73,8 +74,8 @@ vbo_copy_vertices( struct vbo_exec_context *exec )
    GLuint nr = exec->vtx.prim[exec->vtx.prim_count-1].count;
    GLuint ovf, i;
    GLuint sz = exec->vtx.vertex_size;
-   GLfloat *dst = exec->vtx.copied.buffer;
-   const GLfloat *src = (exec->vtx.buffer_map + 
+   fi_type *dst = exec->vtx.copied.buffer;
+   const fi_type *src = (exec->vtx.buffer_map +
                          exec->vtx.prim[exec->vtx.prim_count-1].start * 
                          exec->vtx.vertex_size);
 
@@ -172,7 +173,7 @@ vbo_exec_bind_arrays( struct gl_context *ctx )
          exec->vtx.inputs[attr] = &vbo->currval[VBO_ATTRIB_POS+attr];
       }
       for (attr = 0; attr < MAT_ATTRIB_MAX; attr++) {
-         ASSERT(VERT_ATTRIB_GENERIC(attr) < Elements(exec->vtx.inputs));
+         assert(VERT_ATTRIB_GENERIC(attr) < ARRAY_SIZE(exec->vtx.inputs));
          exec->vtx.inputs[VERT_ATTRIB_GENERIC(attr)] =
             &vbo->currval[VBO_ATTRIB_MAT_FRONT_AMBIENT+attr];
       }
@@ -183,7 +184,7 @@ vbo_exec_bind_arrays( struct gl_context *ctx )
          exec->vtx.inputs[attr] = &vbo->currval[VBO_ATTRIB_POS+attr];
       }
       for (attr = 0; attr < VERT_ATTRIB_GENERIC_MAX; attr++) {
-         ASSERT(VERT_ATTRIB_GENERIC(attr) < Elements(exec->vtx.inputs));
+         assert(VERT_ATTRIB_GENERIC(attr) < ARRAY_SIZE(exec->vtx.inputs));
          exec->vtx.inputs[VERT_ATTRIB_GENERIC(attr)] =
             &vbo->currval[VBO_ATTRIB_GENERIC0+attr];
       }
@@ -213,8 +214,8 @@ vbo_exec_bind_arrays( struct gl_context *ctx )
 	    (GLbyte *)exec->vtx.vertex;
 
          /* override the default array set above */
-         ASSERT(attr < Elements(exec->vtx.inputs));
-         ASSERT(attr < Elements(exec->vtx.arrays)); /* arrays[] */
+         assert(attr < ARRAY_SIZE(exec->vtx.inputs));
+         assert(attr < ARRAY_SIZE(exec->vtx.arrays)); /* arrays[] */
          exec->vtx.inputs[attr] = &arrays[attr];
 
          if (_mesa_is_bufferobj(exec->vtx.bufferobj)) {
@@ -309,7 +310,7 @@ vbo_exec_vtx_map( struct vbo_exec_context *exec )
       /* The VBO exists and there's room for more */
       if (exec->vtx.bufferobj->Size > 0) {
          exec->vtx.buffer_map =
-            (GLfloat *)ctx->Driver.MapBufferRange(ctx, 
+            (fi_type *)ctx->Driver.MapBufferRange(ctx,
                                                   exec->vtx.buffer_used,
                                                   (VBO_VERT_BUFFER_SIZE - 
                                                    exec->vtx.buffer_used),
@@ -336,7 +337,7 @@ vbo_exec_vtx_map( struct vbo_exec_context *exec )
                                  exec->vtx.bufferobj)) {
          /* buffer allocation worked, now map the buffer */
          exec->vtx.buffer_map =
-            (GLfloat *)ctx->Driver.MapBufferRange(ctx,
+            (fi_type *)ctx->Driver.MapBufferRange(ctx,
                                                   0, VBO_VERT_BUFFER_SIZE,
                                                   accessRange,
                                                   exec->vtx.bufferobj,
diff --git a/mesalib/src/mesa/vbo/vbo_exec_eval.c b/mesalib/src/mesa/vbo/vbo_exec_eval.c
index 82f89b95f..b26655bd5 100644
--- a/mesalib/src/mesa/vbo/vbo_exec_eval.c
+++ b/mesalib/src/mesa/vbo/vbo_exec_eval.c
@@ -35,20 +35,20 @@
 
 static void clear_active_eval1( struct vbo_exec_context *exec, GLuint attr ) 
 {
-   assert(attr < Elements(exec->eval.map1));
+   assert(attr < ARRAY_SIZE(exec->eval.map1));
    exec->eval.map1[attr].map = NULL;
 }
 
 static void clear_active_eval2( struct vbo_exec_context *exec, GLuint attr ) 
 {
-   assert(attr < Elements(exec->eval.map2));
+   assert(attr < ARRAY_SIZE(exec->eval.map2));
    exec->eval.map2[attr].map = NULL;
 }
 
 static void set_active_eval1( struct vbo_exec_context *exec, GLuint attr, GLuint dim, 
 			      struct gl_1d_map *map )
 {
-   assert(attr < Elements(exec->eval.map1));
+   assert(attr < ARRAY_SIZE(exec->eval.map1));
    if (!exec->eval.map1[attr].map) {
       exec->eval.map1[attr].map = map;
       exec->eval.map1[attr].sz = dim;
@@ -58,7 +58,7 @@ static void set_active_eval1( struct vbo_exec_context *exec, GLuint attr, GLuint
 static void set_active_eval2( struct vbo_exec_context *exec, GLuint attr, GLuint dim, 
 			      struct gl_2d_map *map )
 {
-   assert(attr < Elements(exec->eval.map2));
+   assert(attr < ARRAY_SIZE(exec->eval.map2));
    if (!exec->eval.map2[attr].map) {
       exec->eval.map2[attr].map = map;
       exec->eval.map2[attr].sz = dim;
@@ -130,11 +130,12 @@ void vbo_exec_do_EvalCoord1f(struct vbo_exec_context *exec, GLfloat u)
       struct gl_1d_map *map = exec->eval.map1[attr].map;
       if (map) {
 	 GLfloat uu = (u - map->u1) * map->du;
-	 GLfloat data[4];
+	 fi_type data[4];
 
-	 ASSIGN_4V(data, 0, 0, 0, 1);
+	 ASSIGN_4V(data, FLOAT_AS_UNION(0), FLOAT_AS_UNION(0),
+		   FLOAT_AS_UNION(0), FLOAT_AS_UNION(1));
 
-	 _math_horner_bezier_curve(map->Points, data, uu, 
+	 _math_horner_bezier_curve(map->Points, &data[0].f, uu,
 				   exec->eval.map1[attr].sz, 
 				   map->Order);
 
@@ -176,12 +177,13 @@ void vbo_exec_do_EvalCoord2f( struct vbo_exec_context *exec,
       if (map) {
 	 GLfloat uu = (u - map->u1) * map->du;
 	 GLfloat vv = (v - map->v1) * map->dv;
-	 GLfloat data[4];
+	 fi_type data[4];
 
-	 ASSIGN_4V(data, 0, 0, 0, 1);
+	 ASSIGN_4V(data, FLOAT_AS_UNION(0), FLOAT_AS_UNION(0),
+		   FLOAT_AS_UNION(0), FLOAT_AS_UNION(1));
 
 	 _math_horner_bezier_surf(map->Points, 
-				  data, 
+				  &data[0].f,
 				  uu, vv, 
 				  exec->eval.map2[attr].sz, 
 				  map->Uorder, map->Vorder);
@@ -203,7 +205,7 @@ void vbo_exec_do_EvalCoord2f( struct vbo_exec_context *exec,
       ASSIGN_4V(vertex, 0, 0, 0, 1);
 
       if (exec->ctx->Eval.AutoNormal) {
-	 GLfloat normal[4];
+	 fi_type normal[4];
          GLfloat du[4], dv[4];
 
          _math_de_casteljau_surf(map->Points, vertex, du, dv, uu, vv, 
@@ -221,11 +223,11 @@ void vbo_exec_do_EvalCoord2f( struct vbo_exec_context *exec,
 	 }
 
 
-         CROSS3(normal, du, dv);
-         NORMALIZE_3FV(normal);
-	 normal[3] = 1.0;
+         CROSS3(&normal[0].f, du, dv);
+         NORMALIZE_3FV(&normal[0].f);
+	 normal[3] = FLOAT_AS_UNION(1.0);
 
-	 COPY_SZ_4V( exec->vtx.attrptr[VBO_ATTRIB_NORMAL],
+ 	 COPY_SZ_4V( exec->vtx.attrptr[VBO_ATTRIB_NORMAL],
 		     exec->vtx.attrsz[VBO_ATTRIB_NORMAL],
 		     normal );
 
diff --git a/mesalib/src/mesa/vbo/vbo_rebase.c b/mesalib/src/mesa/vbo/vbo_rebase.c
index 82a0b8e2b..b06df4ab2 100644
--- a/mesalib/src/mesa/vbo/vbo_rebase.c
+++ b/mesalib/src/mesa/vbo/vbo_rebase.c
@@ -46,6 +46,7 @@
  * of zero.
  */
 
+#include <stdio.h>
 #include "main/glheader.h"
 #include "main/imports.h"
 #include "main/mtypes.h"
diff --git a/mesalib/src/mesa/vbo/vbo_save.h b/mesalib/src/mesa/vbo/vbo_save.h
index fd26b5f61..5b1ac8177 100644
--- a/mesalib/src/mesa/vbo/vbo_save.h
+++ b/mesalib/src/mesa/vbo/vbo_save.h
@@ -40,7 +40,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 
 struct vbo_save_copied_vtx {
-   GLfloat buffer[VBO_ATTRIB_MAX * 4 * VBO_MAX_COPIED_VERTS];
+   fi_type buffer[VBO_ATTRIB_MAX * 4 * VBO_MAX_COPIED_VERTS];
    GLuint nr;
 };
 
@@ -69,7 +69,7 @@ struct vbo_save_vertex_list {
     * Keep this in regular (non-VBO) memory to avoid repeated
     * map/unmap of the VBO when updating GL current data.
     */
-   GLfloat *current_data;
+   fi_type *current_data;
    GLuint current_size;
 
    GLuint buffer_offset;
@@ -107,7 +107,7 @@ struct vbo_save_vertex_list {
  */
 struct vbo_save_vertex_store {
    struct gl_buffer_object *bufferobj;
-   GLfloat *buffer;
+   fi_type *buffer;
    GLuint used;
    GLuint refcount;
 };
@@ -133,7 +133,7 @@ struct vbo_save_context {
 
    GLboolean out_of_memory;  /**< True if last VBO allocation failed */
 
-   GLfloat *buffer;
+   fi_type *buffer;
    GLuint count;
    GLuint wrap_count;
    GLuint replay_flags;
@@ -144,9 +144,9 @@ struct vbo_save_context {
    struct vbo_save_vertex_store *vertex_store;
    struct vbo_save_primitive_store *prim_store;
 
-   GLfloat *buffer_ptr;		   /* cursor, points into buffer */
-   GLfloat vertex[VBO_ATTRIB_MAX*4];	   /* current values */
-   GLfloat *attrptr[VBO_ATTRIB_MAX];
+   fi_type *buffer_ptr;		   /* cursor, points into buffer */
+   fi_type vertex[VBO_ATTRIB_MAX*4];	   /* current values */
+   fi_type *attrptr[VBO_ATTRIB_MAX];
    GLuint vert_count;
    GLuint max_vert;
    GLboolean dangling_attr_ref;
@@ -155,7 +155,7 @@ struct vbo_save_context {
 
    struct vbo_save_copied_vtx copied;
    
-   GLfloat *current[VBO_ATTRIB_MAX]; /* points into ctx->ListState */
+   fi_type *current[VBO_ATTRIB_MAX]; /* points into ctx->ListState */
    GLubyte *currentsz[VBO_ATTRIB_MAX];
 };
 
@@ -186,7 +186,7 @@ void vbo_save_playback_vertex_list( struct gl_context *ctx, void *data );
 
 void vbo_save_api_init( struct vbo_save_context *save );
 
-GLfloat *
+fi_type *
 vbo_save_map_vertex_store(struct gl_context *ctx,
                           struct vbo_save_vertex_store *vertex_store);
 
diff --git a/mesalib/src/mesa/vbo/vbo_save_api.c b/mesalib/src/mesa/vbo/vbo_save_api.c
index beef342be..fd9a5de3d 100644
--- a/mesalib/src/mesa/vbo/vbo_save_api.c
+++ b/mesalib/src/mesa/vbo/vbo_save_api.c
@@ -99,14 +99,14 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 static GLuint
 _save_copy_vertices(struct gl_context *ctx,
                     const struct vbo_save_vertex_list *node,
-                    const GLfloat * src_buffer)
+                    const fi_type * src_buffer)
 {
    struct vbo_save_context *save = &vbo_context(ctx)->save;
    const struct _mesa_prim *prim = &node->prim[node->prim_count - 1];
    GLuint nr = prim->count;
    GLuint sz = save->vertex_size;
-   const GLfloat *src = src_buffer + prim->start * sz;
-   GLfloat *dst = save->copied.buffer;
+   const fi_type *src = src_buffer + prim->start * sz;
+   fi_type *dst = save->copied.buffer;
    GLuint ovf, i;
 
    if (prim->end)
@@ -233,7 +233,7 @@ free_vertex_store(struct gl_context *ctx,
 }
 
 
-GLfloat *
+fi_type *
 vbo_save_map_vertex_store(struct gl_context *ctx,
                           struct vbo_save_vertex_store *vertex_store)
 {
@@ -249,7 +249,7 @@ vbo_save_map_vertex_store(struct gl_context *ctx,
       /* Map the remaining free space in the VBO */
       GLintptr offset = vertex_store->used * sizeof(GLfloat);
       GLsizeiptr size = vertex_store->bufferobj->Size - offset;
-      GLfloat *range = (GLfloat *)
+      fi_type *range = (fi_type *)
          ctx->Driver.MapBufferRange(ctx, offset, size, access,
                                     vertex_store->bufferobj,
                                     MAP_INTERNAL);
@@ -549,7 +549,7 @@ static void
 _save_wrap_filled_vertex(struct gl_context *ctx)
 {
    struct vbo_save_context *save = &vbo_context(ctx)->save;
-   GLfloat *data = save->copied.buffer;
+   fi_type *data = save->copied.buffer;
    GLuint i;
 
    /* Emit a glEnd to close off the last vertex list.
@@ -578,7 +578,7 @@ _save_copy_to_current(struct gl_context *ctx)
    for (i = VBO_ATTRIB_POS + 1; i < VBO_ATTRIB_MAX; i++) {
       if (save->attrsz[i]) {
          save->currentsz[i][0] = save->attrsz[i];
-         COPY_CLEAN_4V_TYPE_AS_FLOAT(save->current[i], save->attrsz[i],
+         COPY_CLEAN_4V_TYPE_AS_UNION(save->current[i], save->attrsz[i],
                                      save->attrptr[i], save->attrtype[i]);
       }
    }
@@ -620,7 +620,7 @@ _save_upgrade_vertex(struct gl_context *ctx, GLuint attr, GLuint newsz)
    struct vbo_save_context *save = &vbo_context(ctx)->save;
    GLuint oldsz;
    GLuint i;
-   GLfloat *tmp;
+   fi_type *tmp;
 
    /* Store the current run of vertices, and emit a GL_END.  Emit a
     * BEGIN in the new buffer.
@@ -669,8 +669,8 @@ _save_upgrade_vertex(struct gl_context *ctx, GLuint attr, GLuint newsz)
     * and will need fixup at runtime.
     */
    if (save->copied.nr) {
-      const GLfloat *data = save->copied.buffer;
-      GLfloat *dest = save->buffer;
+      const fi_type *data = save->copied.buffer;
+      fi_type *dest = save->buffer;
       GLuint j;
 
       /* Need to note this and fix up at runtime (or loopback):
@@ -685,7 +685,7 @@ _save_upgrade_vertex(struct gl_context *ctx, GLuint attr, GLuint newsz)
             if (save->attrsz[j]) {
                if (j == attr) {
                   if (oldsz) {
-                     COPY_CLEAN_4V_TYPE_AS_FLOAT(dest, oldsz, data,
+                     COPY_CLEAN_4V_TYPE_AS_UNION(dest, oldsz, data,
                                                  save->attrtype[j]);
                      data += oldsz;
                      dest += newsz;
@@ -729,7 +729,7 @@ save_fixup_vertex(struct gl_context *ctx, GLuint attr, GLuint sz)
    }
    else if (sz < save->active_sz[attr]) {
       GLuint i;
-      const GLfloat *id = vbo_get_default_vals_as_float(save->attrtype[attr]);
+      const fi_type *id = vbo_get_default_vals_as_union(save->attrtype[attr]);
 
       /* New size is equal or smaller - just need to fill in some
        * zeros.
@@ -772,7 +772,7 @@ _save_reset_vertex(struct gl_context *ctx)
  * 3f version won't otherwise set color[3] to 1.0 -- this is the job
  * of the chooser function when switching between Color4f and Color3f.
  */
-#define ATTR(A, N, T, V0, V1, V2, V3)				\
+#define ATTR_UNION(A, N, T, V0, V1, V2, V3)			\
 do {								\
    struct vbo_save_context *save = &vbo_context(ctx)->save;	\
 								\
@@ -780,12 +780,12 @@ do {								\
       save_fixup_vertex(ctx, A, N);				\
 								\
    {								\
-      GLfloat *dest = save->attrptr[A];				\
+      fi_type *dest = save->attrptr[A];		\
       if (N>0) dest[0] = V0;					\
       if (N>1) dest[1] = V1;					\
       if (N>2) dest[2] = V2;					\
       if (N>3) dest[3] = V3;					\
-      save->attrtype[A] = T;                                    \
+      save->attrtype[A] = T;					\
    }								\
 								\
    if ((A) == 0) {						\
@@ -1564,16 +1564,16 @@ _save_current_init(struct gl_context *ctx)
 
    for (i = VBO_ATTRIB_POS; i <= VBO_ATTRIB_GENERIC15; i++) {
       const GLuint j = i - VBO_ATTRIB_POS;
-      ASSERT(j < VERT_ATTRIB_MAX);
+      assert(j < VERT_ATTRIB_MAX);
       save->currentsz[i] = &ctx->ListState.ActiveAttribSize[j];
-      save->current[i] = ctx->ListState.CurrentAttrib[j];
+      save->current[i] = (fi_type *) ctx->ListState.CurrentAttrib[j];
    }
 
    for (i = VBO_ATTRIB_FIRST_MATERIAL; i <= VBO_ATTRIB_LAST_MATERIAL; i++) {
       const GLuint j = i - VBO_ATTRIB_FIRST_MATERIAL;
-      ASSERT(j < MAT_ATTRIB_MAX);
+      assert(j < MAT_ATTRIB_MAX);
       save->currentsz[i] = &ctx->ListState.ActiveMaterialSize[j];
-      save->current[i] = ctx->ListState.CurrentMaterial[j];
+      save->current[i] = (fi_type *) ctx->ListState.CurrentMaterial[j];
    }
 }
 
diff --git a/mesalib/src/mesa/vbo/vbo_save_draw.c b/mesalib/src/mesa/vbo/vbo_save_draw.c
index d0521d794..74e266d19 100644
--- a/mesalib/src/mesa/vbo/vbo_save_draw.c
+++ b/mesalib/src/mesa/vbo/vbo_save_draw.c
@@ -47,8 +47,8 @@ _playback_copy_to_current(struct gl_context *ctx,
                           const struct vbo_save_vertex_list *node)
 {
    struct vbo_context *vbo = vbo_context(ctx);
-   GLfloat vertex[VBO_ATTRIB_MAX * 4];
-   GLfloat *data;
+   fi_type vertex[VBO_ATTRIB_MAX * 4];
+   fi_type *data;
    GLuint i, offset;
 
    if (node->current_size == 0)
@@ -75,10 +75,10 @@ _playback_copy_to_current(struct gl_context *ctx,
 
    for (i = VBO_ATTRIB_POS+1 ; i < VBO_ATTRIB_MAX ; i++) {
       if (node->attrsz[i]) {
-	 GLfloat *current = (GLfloat *)vbo->currval[i].Ptr;
-         GLfloat tmp[4];
+	 fi_type *current = (fi_type *)vbo->currval[i].Ptr;
+         fi_type tmp[4];
 
-         COPY_CLEAN_4V_TYPE_AS_FLOAT(tmp,
+         COPY_CLEAN_4V_TYPE_AS_UNION(tmp,
                                      node->attrsz[i],
                                      data,
                                      node->attrtype[i]);
diff --git a/mesalib/src/mesa/vbo/vbo_save_loopback.c b/mesalib/src/mesa/vbo/vbo_save_loopback.c
index 8c93ae86a..7410f1875 100644
--- a/mesalib/src/mesa/vbo/vbo_save_loopback.c
+++ b/mesalib/src/mesa/vbo/vbo_save_loopback.c
@@ -25,6 +25,7 @@
  * 
  **************************************************************************/
 
+#include <stdio.h>
 #include "main/context.h"
 #include "main/glheader.h"
 #include "main/enums.h"
diff --git a/mesalib/src/mesa/vbo/vbo_split_copy.c b/mesalib/src/mesa/vbo/vbo_split_copy.c
index ca752e810..d1107dd84 100644
--- a/mesalib/src/mesa/vbo/vbo_split_copy.c
+++ b/mesalib/src/mesa/vbo/vbo_split_copy.c
@@ -29,6 +29,8 @@
 /* Split indexed primitives with per-vertex copying.
  */
 
+#include <stdio.h>
+
 #include "main/glheader.h"
 #include "main/bufferobj.h"
 #include "main/imports.h"
diff --git a/mesalib/src/mesa/x86/common_x86.c b/mesalib/src/mesa/x86/common_x86.c
index 25f5c40e2..14b497d06 100644
--- a/mesalib/src/mesa/x86/common_x86.c
+++ b/mesalib/src/mesa/x86/common_x86.c
@@ -344,13 +344,13 @@ _mesa_get_x86_features(void)
 
 #elif defined(USE_X86_64_ASM)
    {
-      unsigned int uninitialized_var(eax), uninitialized_var(ebx),
-                   uninitialized_var(ecx), uninitialized_var(edx);
+      unsigned int eax, ebx, ecx, edx;
 
       /* Always available on x86-64. */
       _mesa_x86_cpu_features |= X86_FEATURE_XMM | X86_FEATURE_XMM2;
 
-      __get_cpuid(1, &eax, &ebx, &ecx, &edx);
+      if (!__get_cpuid(1, &eax, &ebx, &ecx, &edx))
+         return;
 
       if (ecx & bit_SSE4_1)
          _mesa_x86_cpu_features |= X86_FEATURE_SSE4_1;
diff --git a/mesalib/src/mesa/x86/gen_matypes.c b/mesalib/src/mesa/x86/gen_matypes.c
index 1e904aa0b..18ffb728a 100644
--- a/mesalib/src/mesa/x86/gen_matypes.c
+++ b/mesalib/src/mesa/x86/gen_matypes.c
@@ -35,6 +35,8 @@
 #ifndef __STDC_FORMAT_MACROS
 #define __STDC_FORMAT_MACROS
 #endif
+
+#include <stdio.h>
 #include <inttypes.h>
 
 #include "main/glheader.h"
diff --git a/mesalib/src/util/Makefile.am b/mesalib/src/util/Makefile.am
index 9af233059..ec49dc6cf 100644
--- a/mesalib/src/util/Makefile.am
+++ b/mesalib/src/util/Makefile.am
@@ -25,6 +25,9 @@ include Makefile.sources
 
 noinst_LTLIBRARIES = libmesautil.la
 
+AM_CPPFLAGS = \
+	-I$(top_srcdir)/include
+
 libmesautil_la_CPPFLAGS = \
 	$(DEFINES) \
 	-I$(top_srcdir)/include \
@@ -34,7 +37,8 @@ libmesautil_la_CPPFLAGS = \
 	-I$(top_srcdir)/src/gallium/include \
 	-I$(top_srcdir)/src/gallium/auxiliary \
 	$(SHA1_CFLAGS) \
-	$(VISIBILITY_CFLAGS)
+	$(VISIBILITY_CFLAGS) \
+	$(MSVC2008_COMPAT_CFLAGS)
 
 libmesautil_la_SOURCES = \
 	$(MESA_UTIL_FILES) \
diff --git a/mesalib/src/util/SConscript b/mesalib/src/util/SConscript
index 84bd7a1e1..9e4d481f8 100644
--- a/mesalib/src/util/SConscript
+++ b/mesalib/src/util/SConscript
@@ -6,6 +6,8 @@ from sys import executable as python_cmd
 
 env = env.Clone()
 
+env.MSVC2008Compat()
+
 env.Prepend(CPPPATH = [
     '#include',
     '#src',
diff --git a/mesalib/src/util/bitset.h b/mesalib/src/util/bitset.h
index 17c5d5d25..febcddefd 100644
--- a/mesalib/src/util/bitset.h
+++ b/mesalib/src/util/bitset.h
@@ -94,6 +94,6 @@ __bitset_ffs(const BITSET_WORD *x, int n)
    return 0;
 }
 
-#define BITSET_FFS(x) __bitset_ffs(x, Elements(x))
+#define BITSET_FFS(x) __bitset_ffs(x, ARRAY_SIZE(x))
 
 #endif
diff --git a/mesalib/src/util/macros.h b/mesalib/src/util/macros.h
index eec8b9352..b862bfd5f 100644
--- a/mesalib/src/util/macros.h
+++ b/mesalib/src/util/macros.h
@@ -156,4 +156,27 @@ do {                       \
 #   endif
 #endif
 
+/**
+ * PUBLIC/USED macros
+ *
+ * If we build the library with gcc's -fvisibility=hidden flag, we'll
+ * use the PUBLIC macro to mark functions that are to be exported.
+ *
+ * We also need to define a USED attribute, so the optimizer doesn't
+ * inline a static function that we later use in an alias. - ajax
+ */
+#ifndef PUBLIC
+#  if defined(__GNUC__) || (defined(__SUNPRO_C) && (__SUNPRO_C >= 0x590))
+#    define PUBLIC __attribute__((visibility("default")))
+#    define USED __attribute__((used))
+#  elif defined(_MSC_VER)
+#    define PUBLIC __declspec(dllexport)
+#    define USED
+#  else
+#    define PUBLIC
+#    define USED
+#  endif
+#endif
+
+
 #endif /* UTIL_MACROS_H */
diff --git a/mesalib/src/util/register_allocate.c b/mesalib/src/util/register_allocate.c
index 684ee5d6c..2ad8c3ce1 100644
--- a/mesalib/src/util/register_allocate.c
+++ b/mesalib/src/util/register_allocate.c
@@ -168,6 +168,12 @@ struct ra_graph {
 
    unsigned int *stack;
    unsigned int stack_count;
+
+   /**
+    * Tracks the start of the set of optimistically-colored registers in the
+    * stack.
+    */
+   unsigned int stack_optimistic_start;
 };
 
 /**
@@ -454,6 +460,7 @@ static void
 ra_simplify(struct ra_graph *g)
 {
    bool progress = true;
+   unsigned int stack_optimistic_start = UINT_MAX;
    int i;
 
    while (progress) {
@@ -482,6 +489,9 @@ ra_simplify(struct ra_graph *g)
       }
 
       if (!progress && best_optimistic_node != ~0U) {
+         if (stack_optimistic_start == UINT_MAX)
+            stack_optimistic_start = g->stack_count;
+
 	 decrement_q(g, best_optimistic_node);
 	 g->stack[g->stack_count] = best_optimistic_node;
 	 g->stack_count++;
@@ -489,6 +499,8 @@ ra_simplify(struct ra_graph *g)
 	 progress = true;
       }
    }
+
+   g->stack_optimistic_start = stack_optimistic_start;
 }
 
 /**
@@ -542,7 +554,17 @@ ra_select(struct ra_graph *g)
       g->nodes[n].reg = r;
       g->stack_count--;
 
-      if (g->regs->round_robin)
+      /* Rotate the starting point except for any nodes above the lowest
+       * optimistically colorable node.  The likelihood that we will succeed
+       * at allocating optimistically colorable nodes is highly dependent on
+       * the way that the previous nodes popped off the stack are laid out.
+       * The round-robin strategy increases the fragmentation of the register
+       * file and decreases the number of nearby nodes assigned to the same
+       * color, what increases the likelihood of spilling with respect to the
+       * dense packing strategy.
+       */
+      if (g->regs->round_robin &&
+          g->stack_count - 1 <= g->stack_optimistic_start)
          start_search_reg = r + 1;
    }
 
diff --git a/mesalib/src/util/register_allocate.h b/mesalib/src/util/register_allocate.h
index dc6874495..61f182eff 100644
--- a/mesalib/src/util/register_allocate.h
+++ b/mesalib/src/util/register_allocate.h
@@ -27,6 +27,12 @@
 
 #include <stdbool.h>
 
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+
 struct ra_class;
 struct ra_regs;
 
@@ -77,3 +83,7 @@ void ra_set_node_spill_cost(struct ra_graph *g, unsigned int n, float cost);
 int ra_get_best_spill_node(struct ra_graph *g);
 /** @} */
 
+
+#ifdef __cplusplus
+}  // extern "C"
+#endif