Revert "Prefer intrinsics to handrolled atomic ops."

This reverts commit 5f66b340aa, quickly
fixing 30514.

Makefile

@@ -180,7 +180,7 @@ ultrix-gcc:
 
 # Rules for making release tarballs
 
-VERSION=7.8.1
+VERSION=7.8.3-rc1
 DIRECTORY = Mesa-$(VERSION)
 LIB_NAME = MesaLib-$(VERSION)
 DEMO_NAME = MesaDemos-$(VERSION)
@@ -285,9 +285,6 @@ MAIN_FILES = \
 	$(DIRECTORY)/src/mesa/x86-64/*.[chS]				\
 	$(DIRECTORY)/src/mesa/x86-64/Makefile				\
 	$(DIRECTORY)/progs/Makefile					\
-	$(DIRECTORY)/progs/util/README					\
-	$(DIRECTORY)/progs/util/*.[ch]					\
-	$(DIRECTORY)/progs/util/sampleMakefile				\
 	$(DIRECTORY)/windows/VC8/
 
 ES_FILES = \
@@ -438,7 +435,10 @@ DEMO_FILES = \
 	$(DIRECTORY)/progs/glsl/*.c			\
 	$(DIRECTORY)/progs/glsl/*.frag			\
 	$(DIRECTORY)/progs/glsl/*.vert			\
-	$(DIRECTORY)/progs/glsl/*.shtest
+	$(DIRECTORY)/progs/glsl/*.shtest		\
+	$(DIRECTORY)/progs/util/README			\
+	$(DIRECTORY)/progs/util/*.[ch]			\
+	$(DIRECTORY)/progs/util/sampleMakefile
 
 GLUT_FILES = \
 	$(DIRECTORY)/include/GL/glut.h			\

configs/autoconf.in

@@ -26,6 +26,11 @@ INTEL_LIBS = @INTEL_LIBS@
 INTEL_CFLAGS = @INTEL_CFLAGS@
 X11_LIBS = @X11_LIBS@
 X11_CFLAGS = @X11_CFLAGS@
+GLW_CFLAGS = @GLW_CFLAGS@
+GLUT_CFLAGS = @GLUT_CFLAGS@
+
+# dlopen
+DLOPEN_LIBS = @DLOPEN_LIBS@
 
 # Assembler
 MESA_ASM_SOURCES = @MESA_ASM_SOURCES@

configs/default

@@ -10,7 +10,7 @@ CONFIG_NAME = default
 # Version info
 MESA_MAJOR=7
 MESA_MINOR=8
-MESA_TINY=1
+MESA_TINY=2
 MESA_VERSION = $(MESA_MAJOR).$(MESA_MINOR).$(MESA_TINY)
 
 # external projects.  This should be useless now that we use libdrm.
@@ -121,6 +121,8 @@ APP_LIB_DEPS    = $(EXTRA_LIB_PATH) -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLU_LI
 APP_LIB_DEPS = -lm
 X11_LIBS = -lX11
 
+DLOPEN_LIBS = -ldl
+
 # Installation directories (for make install)
 INSTALL_DIR = /usr/local
 INSTALL_LIB_DIR = $(INSTALL_DIR)/$(LIB_DIR)

configure.ac

@@ -398,6 +398,7 @@ dnl Check to see if dlopen is in default libraries (like Solaris, which
 dnl has it in libc), or if libdl is needed to get it.
 AC_CHECK_FUNC([dlopen], [],
     [AC_CHECK_LIB([dl], [dlopen], [DLOPEN_LIBS="-ldl"])])
+AC_SUBST([DLOPEN_LIBS])
 
 dnl See if posix_memalign is available
 AC_CHECK_FUNC([posix_memalign], [DEFINES="$DEFINES -DHAVE_POSIX_MEMALIGN"])
@@ -699,7 +700,7 @@ AC_SUBST([DRI_DRIVER_SEARCH_DIR])
 dnl Direct rendering or just indirect rendering
 AC_ARG_ENABLE([driglx-direct],
     [AS_HELP_STRING([--disable-driglx-direct],
-        [enable direct rendering in GLX for DRI @<:@default=enabled@:>@])],
+        [enable direct rendering in GLX and EGL for DRI @<:@default=enabled@:>@])],
     [driglx_direct="$enableval"],
     [driglx_direct="yes"])
 dnl Which drivers to build - default is chosen by platform
@@ -865,7 +866,7 @@ else
 fi
 AC_ARG_ENABLE([gl-osmesa],
     [AS_HELP_STRING([--enable-gl-osmesa],
-        [enable OSMesa on libGL @<:@default=enabled for xlib driver@:>@])],
+        [enable OSMesa with libGL @<:@default=enabled for xlib driver@:>@])],
     [gl_osmesa="$enableval"],
     [gl_osmesa="$default_gl_osmesa"])
 if test "x$gl_osmesa" = xyes; then
@@ -900,8 +901,8 @@ x16|x32)
 esac
 AC_SUBST([OSMESA_LIB])
 
-case "$mesa_driver" in
-osmesa)
+case "$DRIVER_DIRS" in
+*osmesa*)
     # only link libraries with osmesa if shared
     if test "$enable_static" = no; then
         OSMESA_LIB_DEPS="-lm -lpthread $SELINUX_LIBS $DLOPEN_LIBS"
@@ -911,19 +912,7 @@ osmesa)
     OSMESA_MESA_DEPS=""
     OSMESA_PC_LIB_PRIV="-lm -lpthread $SELINUX_LIBS $DLOPEN_LIBS"
     ;;
-*)
-    # Link OSMesa to libGL otherwise
-    OSMESA_LIB_DEPS=""
-    # only link libraries with osmesa if shared
-    if test "$enable_static" = no; then
-        OSMESA_MESA_DEPS='-l$(GL_LIB)'
-    else
-        OSMESA_MESA_DEPS=""
-    fi
-    OSMESA_PC_REQ="gl"
-    ;;
 esac
-OSMESA_PC_LIB_PRIV="$OSMESA_PC_LIB_PRIV"
 AC_SUBST([OSMESA_LIB_DEPS])
 AC_SUBST([OSMESA_MESA_DEPS])
 AC_SUBST([OSMESA_PC_REQ])
@@ -1114,10 +1103,6 @@ fi
 
 if test "x$enable_glut" = xyes; then
     SRC_DIRS="$SRC_DIRS glut/glx"
-    GLUT_CFLAGS=""
-    if test "x$GCC" = xyes; then
-        GLUT_CFLAGS="-fexceptions"
-    fi
     if test "$x11_pkgconfig" = yes; then
         PKG_CHECK_MODULES([GLUT],[x11 xmu xi])
         GLUT_PC_REQ_PRIV="x11 xmu xi"
@@ -1128,6 +1113,9 @@ if test "x$enable_glut" = xyes; then
         GLUT_PC_LIB_PRIV="$GLUT_LIB_DEPS"
         GLUT_PC_CFLAGS="$X11_INCLUDES"
     fi
+    if test "x$GCC" = xyes; then
+        GLUT_CFLAGS="$GLUT_CFLAGS -fexceptions"
+    fi
     GLUT_LIB_DEPS="$GLUT_LIB_DEPS -lm"
     GLUT_PC_LIB_PRIV="$GLUT_PC_LIB_PRIV -lm"
 
@@ -1229,6 +1217,10 @@ yes)
             if test "x$enable_egl" != xyes; then
                 AC_MSG_ERROR([cannot build egl state tracker without EGL library])
             fi
+            # define GLX_DIRECT_RENDERING even when the driver is not dri
+            if test "x$mesa_driver" != xdri -a "x$driglx_direct" = xyes; then
+                DEFINES="$DEFINES -DGLX_DIRECT_RENDERING"
+            fi
             ;;
         xorg)
             PKG_CHECK_MODULES([LIBDRM_XORG], [libdrm >= $LIBDRM_XORG_REQUIRED])

docs/devinfo.html

@@ -145,7 +145,7 @@ Make sure the values in src/mesa/main/version.h are correct.
 </p>
 
 <p>
-Update the docs/news.html file and docs/download.html files.
+Update docs/news.html.
 </p>
 
 <p>
@@ -208,10 +208,11 @@ sftp USERNAME,mesa3d@web.sourceforge.net
 
 <p>
 Make an announcement on the mailing lists:
-<em>m</em><em>e</em><em>s</em><em>a</em><em>3</em><em>d</em><em>-</em><em>d</em><em>e</em><em>v</em><em>@</em><em>l</em><em>i</em><em>s</em><em>t</em><em>s</em><em>.</em><em>s</em><em>f</em><em>.</em><em>n</em><em>e</em><em>t</em>,
-<em>m</em><em>e</em><em>s</em><em>a</em><em>3</em><em>d</em><em>-</em><em>u</em><em>s</em><em>e</em><em>r</em><em>s</em><em>@</em><em>l</em><em>i</em><em>s</em><em>t</em><em>s</em><em>.</em><em>s</em><em>f</em><em>.</em><em>n</em><em>e</em><em>t</em>
+
+<em>m</em><em>e</em><em>s</em><em>a</em><em>-</em><em>d</em><em>e</em><em>v</em><em>@</em><em>l</em><em>i</em><em>s</em><em>t</em><em>s</em><em>.</em><em>f</em><em>r</em><em>e</em><em>e</em><em>d</em><em>e</em><em>s</em><em>k</em><em>t</em><em>o</em><em>p</em><em>.</em><em>o</em><em>r</em><em>g</em>,
+<em>m</em><em>e</em><em>s</em><em>a</em><em>-</em><em>u</em><em>s</em><em>e</em><em>r</em><em>s</em><em>@</em><em>l</em><em>i</em><em>s</em><em>t</em><em>s</em><em>.</em><em>f</em><em>r</em><em>e</em><em>e</em><em>d</em><em>e</em><em>s</em><em>k</em><em>t</em><em>o</em><em>p</em><em>.</em><em>o</em><em>r</em><em>g</em>
 and
-<em>m</em><em>e</em><em>s</em><em>a</em><em>3</em><em>d</em><em>-</em><em>a</em><em>n</em><em>n</em><em>o</em><em>u</em><em>n</em><em>c</em><em>e</em><em>@</em><em>l</em><em>i</em><em>s</em><em>t</em><em>s</em><em>.</em><em>s</em><em>f</em><em>.</em><em>n</em><em>e</em><em>t</em>
+<em>m</em><em>e</em><em>s</em><em>a</em><em>-</em><em>a</em><em>n</em><em>n</em><em>o</em><em>u</em><em>n</em><em>c</em><em>e</em><em>@</em><em>l</em><em>i</em><em>s</em><em>t</em><em>s</em><em>.</em><em>f</em><em>r</em><em>e</em><em>e</em><em>d</em><em>e</em><em>s</em><em>k</em><em>t</em><em>o</em><em>p</em><em>.</em><em>o</em><em>r</em><em>g</em>
 </p>
 
 

docs/news.html

@@ -11,6 +11,22 @@
 <H1>News</H1>
 
 
+<h2>TBD, 2010</h2>
+
+<p>
+<a href="relnotes-7.8.3.html">Mesa 7.8.3</a> is released.  This is a bug-fix
+release collecting fixes since the 7.8.2 release.
+</p>
+
+
+<h2>June 16, 2010</h2>
+
+<p>
+<a href="relnotes-7.8.2.html">Mesa 7.8.2</a> is released.  This is a bug-fix
+release collecting fixes since the 7.8.1 release.
+</p>
+
+
 <h2>April 5, 2010</h2>
 
 <p>

docs/relnotes-7.8.2.html

@@ -8,7 +8,7 @@
 
 <body bgcolor="#eeeeee">
 
-<H1>Mesa 7.8.2 Release Notes / April, 5, 2010</H1>
+<H1>Mesa 7.8.2 Release Notes / (date tbd)</H1>
 
 <p>
 Mesa 7.8.2 is a bug fix release which fixes bugs found since the 7.8.1 release.
@@ -26,7 +26,15 @@ for DRI hardware acceleration.
 
 <h2>MD5 checksums</h2>
 <pre>
-tbd
+c89b63d253605ed40e8ac370d25a833c  MesaLib-7.8.2.tar.gz
+6be2d343a0089bfd395ce02aaf8adb57  MesaLib-7.8.2.tar.bz2
+a04ad3b06ac5ff3969a003fa7bbf7d5b  MesaLib-7.8.2.zip
+7c213f92efeb471f0331670d5079d4c0  MesaDemos-7.8.2.tar.gz
+757d9e2e06f48b1a52848be9b0307ced  MesaDemos-7.8.2.tar.bz2
+8d0e5cfe68b8ebf90265d350ae2c48b1  MesaDemos-7.8.2.zip
+b74482e3f44f35ed395c4aada4fd8240  MesaGLUT-7.8.2.tar.gz
+a471807b65e49c325808ba4551be93ed  MesaGLUT-7.8.2.tar.bz2
+9f190268c42be582ef66e47365ee61e3  MesaGLUT-7.8.2.zip
 </pre>
 
 
@@ -43,10 +51,96 @@ tbd
 <h2>Bug fixes</h2>
 <ul>
 <li>Fixed Gallium glDrawPixels(GL_DEPTH_COMPONENT).
+<li>Fixed Gallium Cell driver to buildable, runable state
+<li>Fixed bad error checking for glFramebufferRenderbuffer(attachment=GL_DEPTH_STENCIL_ATTACHMENT).
+<li>Fixed incorrect Z coordinate handling in "meta" glDraw/CopyPixels.
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=23670">Bug
+    #23670</a>.</li>
+
+<li>Assorted i965 driver fixes.
+    Including but not limited to:
+  <ul>
+    <li>Fix scissoring when width or height is
+    0.  <a href="https://bugs.freedesktop.org/show_bug.cgi?id=27643">Bug
+    #27643</a>.
+    <li>Fix bit allocation for number of color regions for
+    ARB_draw_buffers.</li>
+    <li>Set the correct provoking vertex for clipped first-mode
+    trifans.  <a href="https://bugs.freedesktop.org/show_bug.cgi?id=24470">Bug
+    #24470</a>.</li>
+    <li>Use <code>R16G16B16A16_FLOAT</code> for 3-component half-float.</li>
+    <li>Fix assertion for surface tile offset usage on Ironlake.</li>
+    <li>Fix cube map layouts on Ironlake.</li>
+    <li>When an RB gets a new region, clear the old from the state
+    cache.  <a href="https://bugs.freedesktop.org/show_bug.cgi?id=24119">Bug
+    #24119</a>.</li>
+    <li>Reject shaders with uninlined function calls instead of hanging.</li>
+  </ul>
+</li>
+
+<li>Assorted i915 driver fixes.  Including but not limited to:
+  <ul>
+    <li>Fixed texture LOD clamping in i915 driver.
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=24846">Bug
+    #24846</a>.</li>
+    <li>Fix off-by-one for drawing rectangle.
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=27408">Bug
+    #27408</a>.</li>
+  </ul>
+</li>
+
+<li>Fixed hangs in etracer on 830 and 845
+    chipsets.  <a href="https://bugs.freedesktop.org/show_bug.cgi?id=26557">Bug
+    #26557</a>.</li>
+<li>Fixed tiling of small textures on all Intel drivers.</li>
+<li>Fixed crash in Savage driver when using <code>_mesa_CopyTexImage2D</code>.
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=27652">Bug
+    #27652</a>.</li>
+
+<li>Assorted GLX fixes.  Including but not limited to:
+  <ul>
+    <li>Fixed <code>__glXInitializeVisualConfigFromTags</code>'s handling of
+        unrecognized fbconfig tags.</li>
+    <li>Fixed regression with <code>GLX_USE_GL</code>.
+    <li>Fixed config chooser logic for 'mask' matching.</li>
+    <li>Report swap events correctly in direct rendered case (DRI2)</li>
+    <li>Fixed build with dri2proto which doesn't define
+        <code>X_DRI2SwapInterval</code>.</li>
+    <li>Get <code>GLX_SCREEN</code> first in <code>__glXQueryContextInfo</code>.
+        <a href="https://bugs.freedesktop.org/show_bug.cgi?id=14245">Bug
+	#14245</a>.</li>
+  </ul>
+</li>
+
+<li>Assorted GLSL fixes.  Including but not limited to:
+  <ul>
+    <li>Change variable declared assertion into conditional in GLSL
+        compiler.  <a href="https://bugs.freedesktop.org/show_bug.cgi?id=27921">Bug
+        #27921</a>.</li>
+    <li>Fix instruction indexing
+        bugs.  <a href="https://bugs.freedesktop.org/show_bug.cgi?id=27566">Bug
+        #27566</a>.</li>
+    <li>Updated uniform location / offset encoding to be more like
+        other implementations.</li>
+    <li>Don't overwrite a driver's shader infolog with generic failure
+        message.</li>
+  </ul>
+</li>
+
+<li>Fixed OSMesa build for 16 and 32-bit color channel depth.
+<li>Fixed OSMesa build with hidden symbol visibility. libOSMesa no longer links to libGL.
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=28305">Bug
+    #28305</a>.
+<li>Fixed handling of multiple render targets in fixed-function
+    texture envrionmnent programs.</li>
+<li>Fixed conversion errors in <code>signed_rgba8888[rev]</code> texel
+    fetch.</li>
+<li>Don't set srcLevel on <code>GL_TEXTURE_RECTANGLE_ARB</code> targets.</li>
+<li>Various build fixes for OpenBSD.</li>
+<li>Various build fixes for OS X.</li>
+<li>Various build fixes for GCC 3.3.</li>
 </ul>
 
 
-<h2>Changes</h2>
-<p>None.</p>
 </body>
 </html>

docs/relnotes-7.8.3.html

@@ -0,0 +1,89 @@
+<HTML>
+
+<TITLE>Mesa Release Notes</TITLE>
+
+<head><link rel="stylesheet" type="text/css" href="mesa.css"></head>
+
+<BODY>
+
+<body bgcolor="#eeeeee">
+
+<H1>Mesa 7.8.3 Release Notes / (date tbd)</H1>
+
+<p>
+Mesa 7.8.3 is a bug fix release which fixes bugs found since the 7.8.2 release.
+</p>
+<p>
+Mesa 7.8.3 implements the OpenGL 2.1 API, but the version reported by
+glGetString(GL_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 2.1.
+</p>
+<p>
+See the <a href="install.html">Compiling/Installing page</a> for prerequisites
+for DRI hardware acceleration.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+x  MesaLib-7.8.3.tar.gz
+x  MesaLib-7.8.3.tar.bz2
+x  MesaLib-7.8.3.zip
+x  MesaDemos-7.8.3.tar.gz
+x  MesaDemos-7.8.3.tar.bz2
+x  MesaDemos-7.8.3.zip
+x  MesaGLUT-7.8.3.tar.gz
+x  MesaGLUT-7.8.3.tar.bz2
+x  MesaGLUT-7.8.3.zip
+</pre>
+
+
+<h2>New features</h2>
+<p>None.</p>
+
+
+<h2>Changes</h2>
+<ul>
+<li>The radeon driver should use less memory when searching for a valid mip
+image.</li>
+</ul>
+
+
+<h2>Bug fixes</h2>
+<ul>
+<li>Fix unsupported FB with D24S8 (bug 
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=23670">29116</a>)</li>
+<li>Fix ReadPixels crash when reading depth/stencil from an FBO</li>
+<li>Fixed a bug rendering to 16-bit buffers using swrast.</li>
+<li>Fixed a state tracker/TGSI bug that caused crashes when using Windows'
+    memory debugging features.</li>
+<li>Fixed an issue rendering to 32-bit channels with swrast (bug
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=29487">29487</a>)</li>
+<li>GLSL: fix indirect <TT>gl_TextureMatrix</TT> addressing (bug
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=28967">28967</a>)</li>
+<li>GLSL: fix for bug
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=27216">27216</a></li>
+<li>GLSL: fix zw fragcoord entries in some cases (bug
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=29183">29183</a>)</li>
+<li>Fix texture env generation in some cases (bug
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=28169">28169</a>)</li>
+<li>osmesa: a fix for calling <TT>OSMesaMakeCurrent</TT> twice was applied (bug 
+    <a href="https://bugs.freedesktop.org/show_bug.cgi?id=10966">10966</a></li>
+<li>A bug was fixed which could cause Mesa to ignore the
+    <TT>MESA_EXTENSION_OVERRIDE</TT> environment variable.</li>
+<li>A bug related to specular highlights on backfaces was fixed.</li>
+<li>A radeon-specific issue with <TT>glCopyTex(Sub)Image</TT> was
+    corrected.</li>
+<li>radeon/wine: flush command stream in more cases, fixing wine d3d9
+    tests.</li>
+<li>r600: fix sin+cos normalization.</li>
+<li>r600: (properly) ignore <TT>GL_COORD_REPLACE</TT> when point sprites are
+    disabled.</li>
+<li>radeon: avoid flushing when the context is not current.</li>
+<li>r300c: a bug affecting unaligned BOs was fixed.</li>
+<li>r300c: a hardlock caused by ARB_half_float_vertex incorrectly advertised on some chipsets.</li>
+</ul>
+
+
+</body>
+</html>

docs/relnotes.html

@@ -13,6 +13,7 @@ The release notes summarize what's new or changed in each Mesa release.
 </p>
 
 <UL>
+<LI><A HREF="relnotes-7.8.3.html">7.8.3 release notes</A>
 <LI><A HREF="relnotes-7.8.2.html">7.8.2 release notes</A>
 <LI><A HREF="relnotes-7.8.1.html">7.8.1 release notes</A>
 <LI><A HREF="relnotes-7.8.html">7.8 release notes</A>

include/GL/gl_mangle.h

@@ -56,6 +56,7 @@
 #define glBeginOcclusionQueryNV		MANGLE(BeginOcclusionQueryNV)
 #define glBeginPerfMonitorAMD		MANGLE(BeginPerfMonitorAMD)
 #define glBeginQueryARB		MANGLE(BeginQueryARB)
+#define glBeginQueryIndexed		MANGLE(BeginQueryIndexed)
 #define glBeginQuery		MANGLE(BeginQuery)
 #define glBeginTransformFeedbackEXT		MANGLE(BeginTransformFeedbackEXT)
 #define glBeginTransformFeedback		MANGLE(BeginTransformFeedback)
@@ -75,6 +76,7 @@
 #define glBindBufferRange		MANGLE(BindBufferRange)
 #define glBindBufferRangeNV		MANGLE(BindBufferRangeNV)
 #define glBindFragDataLocationEXT		MANGLE(BindFragDataLocationEXT)
+#define glBindFragDataLocationIndexed		MANGLE(BindFragDataLocationIndexed)
 #define glBindFragDataLocation		MANGLE(BindFragDataLocation)
 #define glBindFragmentShaderATI		MANGLE(BindFragmentShaderATI)
 #define glBindFramebufferEXT		MANGLE(BindFramebufferEXT)
@@ -87,10 +89,12 @@
 #define glBindProgramNV		MANGLE(BindProgramNV)
 #define glBindRenderbufferEXT		MANGLE(BindRenderbufferEXT)
 #define glBindRenderbuffer		MANGLE(BindRenderbuffer)
+#define glBindSampler		MANGLE(BindSampler)
 #define glBindTexGenParameterEXT		MANGLE(BindTexGenParameterEXT)
 #define glBindTextureEXT		MANGLE(BindTextureEXT)
 #define glBindTexture		MANGLE(BindTexture)
 #define glBindTextureUnitParameterEXT		MANGLE(BindTextureUnitParameterEXT)
+#define glBindTransformFeedback		MANGLE(BindTransformFeedback)
 #define glBindTransformFeedbackNV		MANGLE(BindTransformFeedbackNV)
 #define glBindVertexArrayAPPLE		MANGLE(BindVertexArrayAPPLE)
 #define glBindVertexArray		MANGLE(BindVertexArray)
@@ -215,6 +219,10 @@
 #define glColorMaskIndexedEXT		MANGLE(ColorMaskIndexedEXT)
 #define glColorMask		MANGLE(ColorMask)
 #define glColorMaterial		MANGLE(ColorMaterial)
+#define glColorP3ui		MANGLE(ColorP3ui)
+#define glColorP3uiv		MANGLE(ColorP3uiv)
+#define glColorP4ui		MANGLE(ColorP4ui)
+#define glColorP4uiv		MANGLE(ColorP4uiv)
 #define glColorPointerEXT		MANGLE(ColorPointerEXT)
 #define glColorPointerListIBM		MANGLE(ColorPointerListIBM)
 #define glColorPointer		MANGLE(ColorPointer)
@@ -236,6 +244,7 @@
 #define glCombinerParameterivNV		MANGLE(CombinerParameterivNV)
 #define glCombinerStageParameterfvNV		MANGLE(CombinerStageParameterfvNV)
 #define glCompileShaderARB		MANGLE(CompileShaderARB)
+#define glCompileShaderIncludeARB		MANGLE(CompileShaderIncludeARB)
 #define glCompileShader		MANGLE(CompileShader)
 #define glCompressedMultiTexImage1DEXT		MANGLE(CompressedMultiTexImage1DEXT)
 #define glCompressedMultiTexImage2DEXT		MANGLE(CompressedMultiTexImage2DEXT)
@@ -326,6 +335,7 @@
 #define glDeleteFramebuffersEXT		MANGLE(DeleteFramebuffersEXT)
 #define glDeleteFramebuffers		MANGLE(DeleteFramebuffers)
 #define glDeleteLists		MANGLE(DeleteLists)
+#define glDeleteNamedStringARB		MANGLE(DeleteNamedStringARB)
 #define glDeleteObjectARB		MANGLE(DeleteObjectARB)
 #define glDeleteOcclusionQueriesNV		MANGLE(DeleteOcclusionQueriesNV)
 #define glDeletePerfMonitorsAMD		MANGLE(DeletePerfMonitorsAMD)
@@ -336,10 +346,12 @@
 #define glDeleteQueries		MANGLE(DeleteQueries)
 #define glDeleteRenderbuffersEXT		MANGLE(DeleteRenderbuffersEXT)
 #define glDeleteRenderbuffers		MANGLE(DeleteRenderbuffers)
+#define glDeleteSamplers		MANGLE(DeleteSamplers)
 #define glDeleteShader		MANGLE(DeleteShader)
 #define glDeleteSync		MANGLE(DeleteSync)
 #define glDeleteTexturesEXT		MANGLE(DeleteTexturesEXT)
 #define glDeleteTextures		MANGLE(DeleteTextures)
+#define glDeleteTransformFeedbacks		MANGLE(DeleteTransformFeedbacks)
 #define glDeleteTransformFeedbacksNV		MANGLE(DeleteTransformFeedbacksNV)
 #define glDeleteVertexArraysAPPLE		MANGLE(DeleteVertexArraysAPPLE)
 #define glDeleteVertexArrays		MANGLE(DeleteVertexArrays)
@@ -363,6 +375,7 @@
 #define glDisableVertexAttribArrayARB		MANGLE(DisableVertexAttribArrayARB)
 #define glDisableVertexAttribArray		MANGLE(DisableVertexAttribArray)
 #define glDrawArraysEXT		MANGLE(DrawArraysEXT)
+#define glDrawArraysIndirect		MANGLE(DrawArraysIndirect)
 #define glDrawArraysInstancedARB		MANGLE(DrawArraysInstancedARB)
 #define glDrawArraysInstancedEXT		MANGLE(DrawArraysInstancedEXT)
 #define glDrawArraysInstanced		MANGLE(DrawArraysInstanced)
@@ -374,6 +387,7 @@
 #define glDrawElementArrayAPPLE		MANGLE(DrawElementArrayAPPLE)
 #define glDrawElementArrayATI		MANGLE(DrawElementArrayATI)
 #define glDrawElementsBaseVertex		MANGLE(DrawElementsBaseVertex)
+#define glDrawElementsIndirect		MANGLE(DrawElementsIndirect)
 #define glDrawElementsInstancedARB		MANGLE(DrawElementsInstancedARB)
 #define glDrawElementsInstancedBaseVertex		MANGLE(DrawElementsInstancedBaseVertex)
 #define glDrawElementsInstancedEXT		MANGLE(DrawElementsInstancedEXT)
@@ -386,7 +400,9 @@
 #define glDrawRangeElementsBaseVertex		MANGLE(DrawRangeElementsBaseVertex)
 #define glDrawRangeElementsEXT		MANGLE(DrawRangeElementsEXT)
 #define glDrawRangeElements		MANGLE(DrawRangeElements)
+#define glDrawTransformFeedback		MANGLE(DrawTransformFeedback)
 #define glDrawTransformFeedbackNV		MANGLE(DrawTransformFeedbackNV)
+#define glDrawTransformFeedbackStream		MANGLE(DrawTransformFeedbackStream)
 #define glEdgeFlagFormatNV		MANGLE(EdgeFlagFormatNV)
 #define glEdgeFlag		MANGLE(EdgeFlag)
 #define glEdgeFlagPointerEXT		MANGLE(EdgeFlagPointerEXT)
@@ -414,6 +430,7 @@
 #define glEndOcclusionQueryNV		MANGLE(EndOcclusionQueryNV)
 #define glEndPerfMonitorAMD		MANGLE(EndPerfMonitorAMD)
 #define glEndQueryARB		MANGLE(EndQueryARB)
+#define glEndQueryIndexed		MANGLE(EndQueryIndexed)
 #define glEndQuery		MANGLE(EndQuery)
 #define glEndTransformFeedbackEXT		MANGLE(EndTransformFeedbackEXT)
 #define glEndTransformFeedback		MANGLE(EndTransformFeedback)
@@ -498,7 +515,6 @@
 #define glFramebufferTextureEXT		MANGLE(FramebufferTextureEXT)
 #define glFramebufferTextureFaceARB		MANGLE(FramebufferTextureFaceARB)
 #define glFramebufferTextureFaceEXT		MANGLE(FramebufferTextureFaceEXT)
-#define glFramebufferTextureFace		MANGLE(FramebufferTextureFace)
 #define glFramebufferTextureLayerARB		MANGLE(FramebufferTextureLayerARB)
 #define glFramebufferTextureLayerEXT		MANGLE(FramebufferTextureLayerEXT)
 #define glFramebufferTextureLayer		MANGLE(FramebufferTextureLayer)
@@ -529,15 +545,20 @@
 #define glGenQueries		MANGLE(GenQueries)
 #define glGenRenderbuffersEXT		MANGLE(GenRenderbuffersEXT)
 #define glGenRenderbuffers		MANGLE(GenRenderbuffers)
+#define glGenSamplers		MANGLE(GenSamplers)
 #define glGenSymbolsEXT		MANGLE(GenSymbolsEXT)
 #define glGenTexturesEXT		MANGLE(GenTexturesEXT)
 #define glGenTextures		MANGLE(GenTextures)
+#define glGenTransformFeedbacks		MANGLE(GenTransformFeedbacks)
 #define glGenTransformFeedbacksNV		MANGLE(GenTransformFeedbacksNV)
 #define glGenVertexArraysAPPLE		MANGLE(GenVertexArraysAPPLE)
 #define glGenVertexArrays		MANGLE(GenVertexArrays)
 #define glGenVertexShadersEXT		MANGLE(GenVertexShadersEXT)
 #define glGetActiveAttribARB		MANGLE(GetActiveAttribARB)
 #define glGetActiveAttrib		MANGLE(GetActiveAttrib)
+#define glGetActiveSubroutineName		MANGLE(GetActiveSubroutineName)
+#define glGetActiveSubroutineUniformiv		MANGLE(GetActiveSubroutineUniformiv)
+#define glGetActiveSubroutineUniformName		MANGLE(GetActiveSubroutineUniformName)
 #define glGetActiveUniformARB		MANGLE(GetActiveUniformARB)
 #define glGetActiveUniformBlockiv		MANGLE(GetActiveUniformBlockiv)
 #define glGetActiveUniformBlockName		MANGLE(GetActiveUniformBlockName)
@@ -599,6 +620,7 @@
 #define glGetFloatIndexedvEXT		MANGLE(GetFloatIndexedvEXT)
 #define glGetFloatv		MANGLE(GetFloatv)
 #define glGetFogFuncSGIS		MANGLE(GetFogFuncSGIS)
+#define glGetFragDataIndex		MANGLE(GetFragDataIndex)
 #define glGetFragDataLocationEXT		MANGLE(GetFragDataLocationEXT)
 #define glGetFragDataLocation		MANGLE(GetFragDataLocation)
 #define glGetFragmentLightfvSGIX		MANGLE(GetFragmentLightfvSGIX)
@@ -678,6 +700,8 @@
 #define glGetNamedProgramLocalParameterIuivEXT		MANGLE(GetNamedProgramLocalParameterIuivEXT)
 #define glGetNamedProgramStringEXT		MANGLE(GetNamedProgramStringEXT)
 #define glGetNamedRenderbufferParameterivEXT		MANGLE(GetNamedRenderbufferParameterivEXT)
+#define glGetNamedStringARB		MANGLE(GetNamedStringARB)
+#define glGetNamedStringivARB		MANGLE(GetNamedStringivARB)
 #define glGetObjectBufferfvATI		MANGLE(GetObjectBufferfvATI)
 #define glGetObjectBufferivATI		MANGLE(GetObjectBufferivATI)
 #define glGetObjectParameterfvARB		MANGLE(GetObjectParameterfvARB)
@@ -717,18 +741,26 @@
 #define glGetProgramParameterdvNV		MANGLE(GetProgramParameterdvNV)
 #define glGetProgramParameterfvNV		MANGLE(GetProgramParameterfvNV)
 #define glGetProgramRegisterfvMESA		MANGLE(GetProgramRegisterfvMESA)
+#define glGetProgramStageiv		MANGLE(GetProgramStageiv)
 #define glGetProgramStringARB		MANGLE(GetProgramStringARB)
 #define glGetProgramStringNV		MANGLE(GetProgramStringNV)
+#define glGetQueryIndexediv		MANGLE(GetQueryIndexediv)
 #define glGetQueryivARB		MANGLE(GetQueryivARB)
 #define glGetQueryiv		MANGLE(GetQueryiv)
 #define glGetQueryObjecti64vEXT		MANGLE(GetQueryObjecti64vEXT)
+#define glGetQueryObjecti64v		MANGLE(GetQueryObjecti64v)
 #define glGetQueryObjectivARB		MANGLE(GetQueryObjectivARB)
 #define glGetQueryObjectiv		MANGLE(GetQueryObjectiv)
 #define glGetQueryObjectui64vEXT		MANGLE(GetQueryObjectui64vEXT)
+#define glGetQueryObjectui64v		MANGLE(GetQueryObjectui64v)
 #define glGetQueryObjectuivARB		MANGLE(GetQueryObjectuivARB)
 #define glGetQueryObjectuiv		MANGLE(GetQueryObjectuiv)
 #define glGetRenderbufferParameterivEXT		MANGLE(GetRenderbufferParameterivEXT)
 #define glGetRenderbufferParameteriv		MANGLE(GetRenderbufferParameteriv)
+#define glGetSamplerParameterfv		MANGLE(GetSamplerParameterfv)
+#define glGetSamplerParameterIfv		MANGLE(GetSamplerParameterIfv)
+#define glGetSamplerParameterIiv		MANGLE(GetSamplerParameterIiv)
+#define glGetSamplerParameteriv		MANGLE(GetSamplerParameteriv)
 #define glGetSeparableFilterEXT		MANGLE(GetSeparableFilterEXT)
 #define glGetSeparableFilter		MANGLE(GetSeparableFilter)
 #define glGetShaderInfoLog		MANGLE(GetShaderInfoLog)
@@ -738,6 +770,8 @@
 #define glGetSharpenTexFuncSGIS		MANGLE(GetSharpenTexFuncSGIS)
 #define glGetStringi		MANGLE(GetStringi)
 #define glGetString		MANGLE(GetString)
+#define glGetSubroutineIndex		MANGLE(GetSubroutineIndex)
+#define glGetSubroutineUniformLocation		MANGLE(GetSubroutineUniformLocation)
 #define glGetSynciv		MANGLE(GetSynciv)
 #define glGetTexBumpParameterfvATI		MANGLE(GetTexBumpParameterfvATI)
 #define glGetTexBumpParameterivATI		MANGLE(GetTexBumpParameterivATI)
@@ -770,6 +804,7 @@
 #define glGetTransformFeedbackVaryingNV		MANGLE(GetTransformFeedbackVaryingNV)
 #define glGetUniformBlockIndex		MANGLE(GetUniformBlockIndex)
 #define glGetUniformBufferSizeEXT		MANGLE(GetUniformBufferSizeEXT)
+#define glGetUniformdv		MANGLE(GetUniformdv)
 #define glGetUniformfvARB		MANGLE(GetUniformfvARB)
 #define glGetUniformfv		MANGLE(GetUniformfv)
 #define glGetUniformIndices		MANGLE(GetUniformIndices)
@@ -778,6 +813,7 @@
 #define glGetUniformLocationARB		MANGLE(GetUniformLocationARB)
 #define glGetUniformLocation		MANGLE(GetUniformLocation)
 #define glGetUniformOffsetEXT		MANGLE(GetUniformOffsetEXT)
+#define glGetUniformSubroutineuiv		MANGLE(GetUniformSubroutineuiv)
 #define glGetUniformui64vNV		MANGLE(GetUniformui64vNV)
 #define glGetUniformuivEXT		MANGLE(GetUniformuivEXT)
 #define glGetUniformuiv		MANGLE(GetUniformuiv)
@@ -865,6 +901,7 @@
 #define glIsFramebuffer		MANGLE(IsFramebuffer)
 #define glIsList		MANGLE(IsList)
 #define glIsNamedBufferResidentNV		MANGLE(IsNamedBufferResidentNV)
+#define glIsNamedStringARB		MANGLE(IsNamedStringARB)
 #define glIsObjectBufferATI		MANGLE(IsObjectBufferATI)
 #define glIsOcclusionQueryNV		MANGLE(IsOcclusionQueryNV)
 #define glIsProgramARB		MANGLE(IsProgramARB)
@@ -874,10 +911,12 @@
 #define glIsQuery		MANGLE(IsQuery)
 #define glIsRenderbufferEXT		MANGLE(IsRenderbufferEXT)
 #define glIsRenderbuffer		MANGLE(IsRenderbuffer)
+#define glIsSampler		MANGLE(IsSampler)
 #define glIsShader		MANGLE(IsShader)
 #define glIsSync		MANGLE(IsSync)
 #define glIsTextureEXT		MANGLE(IsTextureEXT)
 #define glIsTexture		MANGLE(IsTexture)
+#define glIsTransformFeedback		MANGLE(IsTransformFeedback)
 #define glIsTransformFeedbackNV		MANGLE(IsTransformFeedbackNV)
 #define glIsVariantEnabledEXT		MANGLE(IsVariantEnabledEXT)
 #define glIsVertexArrayAPPLE		MANGLE(IsVertexArrayAPPLE)
@@ -915,6 +954,8 @@
 #define glLogicOp		MANGLE(LogicOp)
 #define glMakeBufferNonResidentNV		MANGLE(MakeBufferNonResidentNV)
 #define glMakeBufferResidentNV		MANGLE(MakeBufferResidentNV)
+#define glMakeNamedBufferNonResidentNV		MANGLE(MakeNamedBufferNonResidentNV)
+#define glMakeNamedBufferResidentNV		MANGLE(MakeNamedBufferResidentNV)
 #define glMap1d		MANGLE(Map1d)
 #define glMap1f		MANGLE(Map1f)
 #define glMap2d		MANGLE(Map2d)
@@ -1048,6 +1089,14 @@
 #define glMultiTexCoord4s		MANGLE(MultiTexCoord4s)
 #define glMultiTexCoord4svARB		MANGLE(MultiTexCoord4svARB)
 #define glMultiTexCoord4sv		MANGLE(MultiTexCoord4sv)
+#define glMultiTexCoordP1ui		MANGLE(MultiTexCoordP1ui)
+#define glMultiTexCoordP1uiv		MANGLE(MultiTexCoordP1uiv)
+#define glMultiTexCoordP2ui		MANGLE(MultiTexCoordP2ui)
+#define glMultiTexCoordP2uiv		MANGLE(MultiTexCoordP2uiv)
+#define glMultiTexCoordP3ui		MANGLE(MultiTexCoordP3ui)
+#define glMultiTexCoordP3uiv		MANGLE(MultiTexCoordP3uiv)
+#define glMultiTexCoordP4ui		MANGLE(MultiTexCoordP4ui)
+#define glMultiTexCoordP4uiv		MANGLE(MultiTexCoordP4uiv)
 #define glMultiTexCoordPointerEXT		MANGLE(MultiTexCoordPointerEXT)
 #define glMultiTexEnvfEXT		MANGLE(MultiTexEnvfEXT)
 #define glMultiTexEnvfvEXT		MANGLE(MultiTexEnvfvEXT)
@@ -1087,8 +1136,6 @@
 #define glNamedFramebufferTextureEXT		MANGLE(NamedFramebufferTextureEXT)
 #define glNamedFramebufferTextureFaceEXT		MANGLE(NamedFramebufferTextureFaceEXT)
 #define glNamedFramebufferTextureLayerEXT		MANGLE(NamedFramebufferTextureLayerEXT)
-#define glNamedMakeBufferNonResidentNV		MANGLE(NamedMakeBufferNonResidentNV)
-#define glNamedMakeBufferResidentNV		MANGLE(NamedMakeBufferResidentNV)
 #define glNamedProgramLocalParameter4dEXT		MANGLE(NamedProgramLocalParameter4dEXT)
 #define glNamedProgramLocalParameter4dvEXT		MANGLE(NamedProgramLocalParameter4dvEXT)
 #define glNamedProgramLocalParameter4fEXT		MANGLE(NamedProgramLocalParameter4fEXT)
@@ -1104,6 +1151,7 @@
 #define glNamedRenderbufferStorageEXT		MANGLE(NamedRenderbufferStorageEXT)
 #define glNamedRenderbufferStorageMultisampleCoverageEXT		MANGLE(NamedRenderbufferStorageMultisampleCoverageEXT)
 #define glNamedRenderbufferStorageMultisampleEXT		MANGLE(NamedRenderbufferStorageMultisampleEXT)
+#define glNamedStringARB		MANGLE(NamedStringARB)
 #define glNewList		MANGLE(NewList)
 #define glNewObjectBufferATI		MANGLE(NewObjectBufferATI)
 #define glNormal3b		MANGLE(Normal3b)
@@ -1121,6 +1169,8 @@
 #define glNormal3s		MANGLE(Normal3s)
 #define glNormal3sv		MANGLE(Normal3sv)
 #define glNormalFormatNV		MANGLE(NormalFormatNV)
+#define glNormalP3ui		MANGLE(NormalP3ui)
+#define glNormalP3uiv		MANGLE(NormalP3uiv)
 #define glNormalPointerEXT		MANGLE(NormalPointerEXT)
 #define glNormalPointerListIBM		MANGLE(NormalPointerListIBM)
 #define glNormalPointer		MANGLE(NormalPointer)
@@ -1140,6 +1190,9 @@
 #define glOrtho		MANGLE(Ortho)
 #define glPassTexCoordATI		MANGLE(PassTexCoordATI)
 #define glPassThrough		MANGLE(PassThrough)
+#define glPatchParameterfv		MANGLE(PatchParameterfv)
+#define glPatchParameteri		MANGLE(PatchParameteri)
+#define glPauseTransformFeedback		MANGLE(PauseTransformFeedback)
 #define glPauseTransformFeedbackNV		MANGLE(PauseTransformFeedbackNV)
 #define glPixelDataRangeNV		MANGLE(PixelDataRangeNV)
 #define glPixelMapfv		MANGLE(PixelMapfv)
@@ -1231,38 +1284,55 @@
 #define glProgramParameters4dvNV		MANGLE(ProgramParameters4dvNV)
 #define glProgramParameters4fvNV		MANGLE(ProgramParameters4fvNV)
 #define glProgramStringARB		MANGLE(ProgramStringARB)
+#define glProgramUniform1dEXT		MANGLE(ProgramUniform1dEXT)
+#define glProgramUniform1dvEXT		MANGLE(ProgramUniform1dvEXT)
 #define glProgramUniform1fEXT		MANGLE(ProgramUniform1fEXT)
 #define glProgramUniform1fvEXT		MANGLE(ProgramUniform1fvEXT)
 #define glProgramUniform1iEXT		MANGLE(ProgramUniform1iEXT)
 #define glProgramUniform1ivEXT		MANGLE(ProgramUniform1ivEXT)
 #define glProgramUniform1uiEXT		MANGLE(ProgramUniform1uiEXT)
 #define glProgramUniform1uivEXT		MANGLE(ProgramUniform1uivEXT)
+#define glProgramUniform2dEXT		MANGLE(ProgramUniform2dEXT)
+#define glProgramUniform2dvEXT		MANGLE(ProgramUniform2dvEXT)
 #define glProgramUniform2fEXT		MANGLE(ProgramUniform2fEXT)
 #define glProgramUniform2fvEXT		MANGLE(ProgramUniform2fvEXT)
 #define glProgramUniform2iEXT		MANGLE(ProgramUniform2iEXT)
 #define glProgramUniform2ivEXT		MANGLE(ProgramUniform2ivEXT)
 #define glProgramUniform2uiEXT		MANGLE(ProgramUniform2uiEXT)
 #define glProgramUniform2uivEXT		MANGLE(ProgramUniform2uivEXT)
+#define glProgramUniform3dEXT		MANGLE(ProgramUniform3dEXT)
+#define glProgramUniform3dvEXT		MANGLE(ProgramUniform3dvEXT)
 #define glProgramUniform3fEXT		MANGLE(ProgramUniform3fEXT)
 #define glProgramUniform3fvEXT		MANGLE(ProgramUniform3fvEXT)
 #define glProgramUniform3iEXT		MANGLE(ProgramUniform3iEXT)
 #define glProgramUniform3ivEXT		MANGLE(ProgramUniform3ivEXT)
 #define glProgramUniform3uiEXT		MANGLE(ProgramUniform3uiEXT)
 #define glProgramUniform3uivEXT		MANGLE(ProgramUniform3uivEXT)
+#define glProgramUniform4dEXT		MANGLE(ProgramUniform4dEXT)
+#define glProgramUniform4dvEXT		MANGLE(ProgramUniform4dvEXT)
 #define glProgramUniform4fEXT		MANGLE(ProgramUniform4fEXT)
 #define glProgramUniform4fvEXT		MANGLE(ProgramUniform4fvEXT)
 #define glProgramUniform4iEXT		MANGLE(ProgramUniform4iEXT)
 #define glProgramUniform4ivEXT		MANGLE(ProgramUniform4ivEXT)
 #define glProgramUniform4uiEXT		MANGLE(ProgramUniform4uiEXT)
 #define glProgramUniform4uivEXT		MANGLE(ProgramUniform4uivEXT)
+#define glProgramUniformMatrix2dvEXT		MANGLE(ProgramUniformMatrix2dvEXT)
 #define glProgramUniformMatrix2fvEXT		MANGLE(ProgramUniformMatrix2fvEXT)
+#define glProgramUniformMatrix2x3dvEXT		MANGLE(ProgramUniformMatrix2x3dvEXT)
 #define glProgramUniformMatrix2x3fvEXT		MANGLE(ProgramUniformMatrix2x3fvEXT)
+#define glProgramUniformMatrix2x4dvEXT		MANGLE(ProgramUniformMatrix2x4dvEXT)
 #define glProgramUniformMatrix2x4fvEXT		MANGLE(ProgramUniformMatrix2x4fvEXT)
+#define glProgramUniformMatrix3dvEXT		MANGLE(ProgramUniformMatrix3dvEXT)
 #define glProgramUniformMatrix3fvEXT		MANGLE(ProgramUniformMatrix3fvEXT)
+#define glProgramUniformMatrix3x2dvEXT		MANGLE(ProgramUniformMatrix3x2dvEXT)
 #define glProgramUniformMatrix3x2fvEXT		MANGLE(ProgramUniformMatrix3x2fvEXT)
+#define glProgramUniformMatrix3x4dvEXT		MANGLE(ProgramUniformMatrix3x4dvEXT)
 #define glProgramUniformMatrix3x4fvEXT		MANGLE(ProgramUniformMatrix3x4fvEXT)
+#define glProgramUniformMatrix4dvEXT		MANGLE(ProgramUniformMatrix4dvEXT)
 #define glProgramUniformMatrix4fvEXT		MANGLE(ProgramUniformMatrix4fvEXT)
+#define glProgramUniformMatrix4x2dvEXT		MANGLE(ProgramUniformMatrix4x2dvEXT)
 #define glProgramUniformMatrix4x2fvEXT		MANGLE(ProgramUniformMatrix4x2fvEXT)
+#define glProgramUniformMatrix4x3dvEXT		MANGLE(ProgramUniformMatrix4x3dvEXT)
 #define glProgramUniformMatrix4x3fvEXT		MANGLE(ProgramUniformMatrix4x3fvEXT)
 #define glProgramUniformui64NV		MANGLE(ProgramUniformui64NV)
 #define glProgramUniformui64vNV		MANGLE(ProgramUniformui64vNV)
@@ -1274,6 +1344,7 @@
 #define glPushClientAttrib		MANGLE(PushClientAttrib)
 #define glPushMatrix		MANGLE(PushMatrix)
 #define glPushName		MANGLE(PushName)
+#define glQueryCounter		MANGLE(QueryCounter)
 #define glRasterPos2d		MANGLE(RasterPos2d)
 #define glRasterPos2dv		MANGLE(RasterPos2dv)
 #define glRasterPos2f		MANGLE(RasterPos2f)
@@ -1345,6 +1416,7 @@
 #define glResetMinmaxEXT		MANGLE(ResetMinmaxEXT)
 #define glResetMinmax		MANGLE(ResetMinmax)
 #define glResizeBuffersMESA		MANGLE(ResizeBuffersMESA)
+#define glResumeTransformFeedback		MANGLE(ResumeTransformFeedback)
 #define glResumeTransformFeedbackNV		MANGLE(ResumeTransformFeedbackNV)
 #define glRotated		MANGLE(Rotated)
 #define glRotatef		MANGLE(Rotatef)
@@ -1357,6 +1429,12 @@
 #define glSampleMaskSGIS		MANGLE(SampleMaskSGIS)
 #define glSamplePatternEXT		MANGLE(SamplePatternEXT)
 #define glSamplePatternSGIS		MANGLE(SamplePatternSGIS)
+#define glSamplerParameterf		MANGLE(SamplerParameterf)
+#define glSamplerParameterfv		MANGLE(SamplerParameterfv)
+#define glSamplerParameterIiv		MANGLE(SamplerParameterIiv)
+#define glSamplerParameteri		MANGLE(SamplerParameteri)
+#define glSamplerParameterIuiv		MANGLE(SamplerParameterIuiv)
+#define glSamplerParameteriv		MANGLE(SamplerParameteriv)
 #define glScaled		MANGLE(Scaled)
 #define glScalef		MANGLE(Scalef)
 #define glScissor		MANGLE(Scissor)
@@ -1395,6 +1473,8 @@
 #define glSecondaryColor3usvEXT		MANGLE(SecondaryColor3usvEXT)
 #define glSecondaryColor3usv		MANGLE(SecondaryColor3usv)
 #define glSecondaryColorFormatNV		MANGLE(SecondaryColorFormatNV)
+#define glSecondaryColorP3ui		MANGLE(SecondaryColorP3ui)
+#define glSecondaryColorP3uiv		MANGLE(SecondaryColorP3uiv)
 #define glSecondaryColorPointerEXT		MANGLE(SecondaryColorPointerEXT)
 #define glSecondaryColorPointerListIBM		MANGLE(SecondaryColorPointerListIBM)
 #define glSecondaryColorPointer		MANGLE(SecondaryColorPointer)
@@ -1509,6 +1589,14 @@
 #define glTexCoord4s		MANGLE(TexCoord4s)
 #define glTexCoord4sv		MANGLE(TexCoord4sv)
 #define glTexCoordFormatNV		MANGLE(TexCoordFormatNV)
+#define glTexCoordP1ui		MANGLE(TexCoordP1ui)
+#define glTexCoordP1uiv		MANGLE(TexCoordP1uiv)
+#define glTexCoordP2ui		MANGLE(TexCoordP2ui)
+#define glTexCoordP2uiv		MANGLE(TexCoordP2uiv)
+#define glTexCoordP3ui		MANGLE(TexCoordP3ui)
+#define glTexCoordP3uiv		MANGLE(TexCoordP3uiv)
+#define glTexCoordP4ui		MANGLE(TexCoordP4ui)
+#define glTexCoordP4uiv		MANGLE(TexCoordP4uiv)
 #define glTexCoordPointerEXT		MANGLE(TexCoordPointerEXT)
 #define glTexCoordPointerListIBM		MANGLE(TexCoordPointerListIBM)
 #define glTexCoordPointer		MANGLE(TexCoordPointer)
@@ -1574,6 +1662,8 @@
 #define glTransformFeedbackVaryingsNV		MANGLE(TransformFeedbackVaryingsNV)
 #define glTranslated		MANGLE(Translated)
 #define glTranslatef		MANGLE(Translatef)
+#define glUniform1d		MANGLE(Uniform1d)
+#define glUniform1dv		MANGLE(Uniform1dv)
 #define glUniform1fARB		MANGLE(Uniform1fARB)
 #define glUniform1f		MANGLE(Uniform1f)
 #define glUniform1fvARB		MANGLE(Uniform1fvARB)
@@ -1586,6 +1676,8 @@
 #define glUniform1ui		MANGLE(Uniform1ui)
 #define glUniform1uivEXT		MANGLE(Uniform1uivEXT)
 #define glUniform1uiv		MANGLE(Uniform1uiv)
+#define glUniform2d		MANGLE(Uniform2d)
+#define glUniform2dv		MANGLE(Uniform2dv)
 #define glUniform2fARB		MANGLE(Uniform2fARB)
 #define glUniform2f		MANGLE(Uniform2f)
 #define glUniform2fvARB		MANGLE(Uniform2fvARB)
@@ -1598,6 +1690,8 @@
 #define glUniform2ui		MANGLE(Uniform2ui)
 #define glUniform2uivEXT		MANGLE(Uniform2uivEXT)
 #define glUniform2uiv		MANGLE(Uniform2uiv)
+#define glUniform3d		MANGLE(Uniform3d)
+#define glUniform3dv		MANGLE(Uniform3dv)
 #define glUniform3fARB		MANGLE(Uniform3fARB)
 #define glUniform3f		MANGLE(Uniform3f)
 #define glUniform3fvARB		MANGLE(Uniform3fvARB)
@@ -1610,6 +1704,8 @@
 #define glUniform3ui		MANGLE(Uniform3ui)
 #define glUniform3uivEXT		MANGLE(Uniform3uivEXT)
 #define glUniform3uiv		MANGLE(Uniform3uiv)
+#define glUniform4d		MANGLE(Uniform4d)
+#define glUniform4dv		MANGLE(Uniform4dv)
 #define glUniform4fARB		MANGLE(Uniform4fARB)
 #define glUniform4f		MANGLE(Uniform4f)
 #define glUniform4fvARB		MANGLE(Uniform4fvARB)
@@ -1624,18 +1720,28 @@
 #define glUniform4uiv		MANGLE(Uniform4uiv)
 #define glUniformBlockBinding		MANGLE(UniformBlockBinding)
 #define glUniformBufferEXT		MANGLE(UniformBufferEXT)
+#define glUniformMatrix2dv		MANGLE(UniformMatrix2dv)
 #define glUniformMatrix2fvARB		MANGLE(UniformMatrix2fvARB)
 #define glUniformMatrix2fv		MANGLE(UniformMatrix2fv)
+#define glUniformMatrix2x3dv		MANGLE(UniformMatrix2x3dv)
 #define glUniformMatrix2x3fv		MANGLE(UniformMatrix2x3fv)
+#define glUniformMatrix2x4dv		MANGLE(UniformMatrix2x4dv)
 #define glUniformMatrix2x4fv		MANGLE(UniformMatrix2x4fv)
+#define glUniformMatrix3dv		MANGLE(UniformMatrix3dv)
 #define glUniformMatrix3fvARB		MANGLE(UniformMatrix3fvARB)
 #define glUniformMatrix3fv		MANGLE(UniformMatrix3fv)
+#define glUniformMatrix3x2dv		MANGLE(UniformMatrix3x2dv)
 #define glUniformMatrix3x2fv		MANGLE(UniformMatrix3x2fv)
+#define glUniformMatrix3x4dv		MANGLE(UniformMatrix3x4dv)
 #define glUniformMatrix3x4fv		MANGLE(UniformMatrix3x4fv)
+#define glUniformMatrix4dv		MANGLE(UniformMatrix4dv)
 #define glUniformMatrix4fvARB		MANGLE(UniformMatrix4fvARB)
 #define glUniformMatrix4fv		MANGLE(UniformMatrix4fv)
+#define glUniformMatrix4x2dv		MANGLE(UniformMatrix4x2dv)
 #define glUniformMatrix4x2fv		MANGLE(UniformMatrix4x2fv)
+#define glUniformMatrix4x3dv		MANGLE(UniformMatrix4x3dv)
 #define glUniformMatrix4x3fv		MANGLE(UniformMatrix4x3fv)
+#define glUniformSubroutinesuiv		MANGLE(UniformSubroutinesuiv)
 #define glUniformui64NV		MANGLE(Uniformui64NV)
 #define glUniformui64vNV		MANGLE(Uniformui64vNV)
 #define glUnlockArraysEXT		MANGLE(UnlockArraysEXT)
@@ -1844,6 +1950,14 @@
 #define glVertexAttribIFormatNV		MANGLE(VertexAttribIFormatNV)
 #define glVertexAttribIPointerEXT		MANGLE(VertexAttribIPointerEXT)
 #define glVertexAttribIPointer		MANGLE(VertexAttribIPointer)
+#define glVertexAttribP1ui		MANGLE(VertexAttribP1ui)
+#define glVertexAttribP1uiv		MANGLE(VertexAttribP1uiv)
+#define glVertexAttribP2ui		MANGLE(VertexAttribP2ui)
+#define glVertexAttribP2uiv		MANGLE(VertexAttribP2uiv)
+#define glVertexAttribP3ui		MANGLE(VertexAttribP3ui)
+#define glVertexAttribP3uiv		MANGLE(VertexAttribP3uiv)
+#define glVertexAttribP4ui		MANGLE(VertexAttribP4ui)
+#define glVertexAttribP4uiv		MANGLE(VertexAttribP4uiv)
 #define glVertexAttribPointerARB		MANGLE(VertexAttribPointerARB)
 #define glVertexAttribPointer		MANGLE(VertexAttribPointer)
 #define glVertexAttribPointerNV		MANGLE(VertexAttribPointerNV)
@@ -1868,6 +1982,12 @@
 #define glVertexBlendEnvfATI		MANGLE(VertexBlendEnvfATI)
 #define glVertexBlendEnviATI		MANGLE(VertexBlendEnviATI)
 #define glVertexFormatNV		MANGLE(VertexFormatNV)
+#define glVertexP2ui		MANGLE(VertexP2ui)
+#define glVertexP2uiv		MANGLE(VertexP2uiv)
+#define glVertexP3ui		MANGLE(VertexP3ui)
+#define glVertexP3uiv		MANGLE(VertexP3uiv)
+#define glVertexP4ui		MANGLE(VertexP4ui)
+#define glVertexP4uiv		MANGLE(VertexP4uiv)
 #define glVertexPointerEXT		MANGLE(VertexPointerEXT)
 #define glVertexPointerListIBM		MANGLE(VertexPointerListIBM)
 #define glVertexPointer		MANGLE(VertexPointer)

progs/egl/Makefile

@@ -60,7 +60,7 @@ xeglgears: xeglgears.o $(HEADERS) $(LIB_DEP)
 	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $< $(LIBS) -lm $(X11_LIBS)
 
 xeglthreads: xeglthreads.o $(HEADERS) $(LIB_DEP)
-	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $< $(LIBS) -lm $(X11_LIBS)
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $< $(LIBS) -lm -lpthread $(X11_LIBS)
 
 xegl_tri: xegl_tri.o $(HEADERS) $(LIB_DEP)
 	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $< $(LIBS) -lm $(X11_LIBS)

src/egl/drivers/dri2/egl_dri2.c

@@ -1331,6 +1331,7 @@ _eglMain(const char *args)
    if (!dri2_drv)
       return NULL;
 
+   memset(dri2_drv, 0, sizeof *dri2_drv);
    _eglInitDriverFallbacks(&dri2_drv->base);
    dri2_drv->base.API.Initialize = dri2_initialize;
    dri2_drv->base.API.Terminate = dri2_terminate;

src/gallium/auxiliary/draw/draw_vs_ppc.c

@@ -125,7 +125,7 @@ vs_ppc_run_linear( struct draw_vertex_shader *base,
        */
       shader->func(inputs_soa, outputs_soa, temps_soa,
 		   (float (*)[4]) shader->base.immediates,
-                   (const float (*)[4])constants[0],
+                   (float (*)[4]) constants[0],
                    ppc_builtin_constants);
 
       /* convert (up to) four output verts from SoA back to AoS format */

src/gallium/auxiliary/indices/.gitignore

@@ -0,0 +1,2 @@
+u_indices_gen.c
+u_unfilled_gen.c

src/gallium/auxiliary/indices/u_unfilled_gen.c

@@ -1,992 +0,0 @@
-/* File automatically generated by u_unfilled_gen.py */
-
-/*
- * Copyright 2009 VMware, Inc.
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * on the rights to use, copy, modify, merge, publish, distribute, sub
- * license, and/or sell copies of the Software, and to permit persons to whom
- * the Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.  IN NO EVENT SHALL
- * VMWARE AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-
-
-/**
- * @file
- * Functions to translate and generate index lists
- */
-
-#include "indices/u_indices.h"
-#include "indices/u_indices_priv.h"
-#include "pipe/p_compiler.h"
-#include "util/u_debug.h"
-#include "pipe/p_defines.h"
-#include "util/u_memory.h"
-
-
-static unsigned out_size_idx( unsigned index_size )
-{
-   switch (index_size) {
-   case 4: return OUT_UINT;
-   case 2: return OUT_USHORT;
-   default: assert(0); return OUT_USHORT;
-   }
-}
-
-static unsigned in_size_idx( unsigned index_size )
-{
-   switch (index_size) {
-   case 4: return IN_UINT;
-   case 2: return IN_USHORT;
-   case 1: return IN_UBYTE;
-   default: assert(0); return IN_UBYTE;
-   }
-}
-
-
-static u_generate_func generate_line[OUT_COUNT][PRIM_COUNT];
-static u_translate_func translate_line[IN_COUNT][OUT_COUNT][PRIM_COUNT];
-
-
-static void generate_tris_ushort(
-    unsigned nr,
-    void *_out )
-{
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (ushort)(i);
-      (out+j)[1] = (ushort)(i+1);
-      (out+j+2)[0] = (ushort)(i+1);
-      (out+j+2)[1] = (ushort)(i+2);
-      (out+j+4)[0] = (ushort)(i+2);
-      (out+j+4)[1] = (ushort)(i);
-   }
-}
-static void generate_tristrip_ushort(
-    unsigned nr,
-    void *_out )
-{
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)(i);
-      (out+j)[1] = (ushort)(i+1/*+(i&1)*/);
-      (out+j+2)[0] = (ushort)(i+1/*+(i&1)*/);
-      (out+j+2)[1] = (ushort)(i+2/*-(i&1)*/);
-      (out+j+4)[0] = (ushort)(i+2/*-(i&1)*/);
-      (out+j+4)[1] = (ushort)(i);
-   }
-}
-static void generate_trifan_ushort(
-    unsigned nr,
-    void *_out )
-{
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)(0);
-      (out+j)[1] = (ushort)(i+1);
-      (out+j+2)[0] = (ushort)(i+1);
-      (out+j+2)[1] = (ushort)(i+2);
-      (out+j+4)[0] = (ushort)(i+2);
-      (out+j+4)[1] = (ushort)(0);
-   }
-}
-static void generate_quads_ushort(
-    unsigned nr,
-    void *_out )
-{
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (ushort)(i+0);
-      (out+j)[1] = (ushort)(i+1);
-      (out+j+2)[0] = (ushort)(i+1);
-      (out+j+2)[1] = (ushort)(i+2);
-      (out+j+4)[0] = (ushort)(i+2);
-      (out+j+4)[1] = (ushort)(i+3);
-      (out+j+6)[0] = (ushort)(i+3);
-      (out+j+6)[1] = (ushort)(i+0);
-   }
-}
-static void generate_quadstrip_ushort(
-    unsigned nr,
-    void *_out )
-{
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (ushort)(i+2);
-      (out+j)[1] = (ushort)(i+0);
-      (out+j+2)[0] = (ushort)(i+0);
-      (out+j+2)[1] = (ushort)(i+1);
-      (out+j+4)[0] = (ushort)(i+1);
-      (out+j+4)[1] = (ushort)(i+3);
-      (out+j+6)[0] = (ushort)(i+3);
-      (out+j+6)[1] = (ushort)(i+2);
-   }
-}
-static void generate_polygon_ushort(
-    unsigned nr,
-    void *_out )
-{
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)(0);
-      (out+j)[1] = (ushort)(i+1);
-      (out+j+2)[0] = (ushort)(i+1);
-      (out+j+2)[1] = (ushort)(i+2);
-      (out+j+4)[0] = (ushort)(i+2);
-      (out+j+4)[1] = (ushort)(0);
-   }
-}
-static void generate_tris_uint(
-    unsigned nr,
-    void *_out )
-{
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (uint)(i);
-      (out+j)[1] = (uint)(i+1);
-      (out+j+2)[0] = (uint)(i+1);
-      (out+j+2)[1] = (uint)(i+2);
-      (out+j+4)[0] = (uint)(i+2);
-      (out+j+4)[1] = (uint)(i);
-   }
-}
-static void generate_tristrip_uint(
-    unsigned nr,
-    void *_out )
-{
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)(i);
-      (out+j)[1] = (uint)(i+1/*+(i&1)*/);
-      (out+j+2)[0] = (uint)(i+1/*+(i&1)*/);
-      (out+j+2)[1] = (uint)(i+2/*-(i&1)*/);
-      (out+j+4)[0] = (uint)(i+2/*-(i&1)*/);
-      (out+j+4)[1] = (uint)(i);
-   }
-}
-static void generate_trifan_uint(
-    unsigned nr,
-    void *_out )
-{
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)(0);
-      (out+j)[1] = (uint)(i+1);
-      (out+j+2)[0] = (uint)(i+1);
-      (out+j+2)[1] = (uint)(i+2);
-      (out+j+4)[0] = (uint)(i+2);
-      (out+j+4)[1] = (uint)(0);
-   }
-}
-static void generate_quads_uint(
-    unsigned nr,
-    void *_out )
-{
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (uint)(i+0);
-      (out+j)[1] = (uint)(i+1);
-      (out+j+2)[0] = (uint)(i+1);
-      (out+j+2)[1] = (uint)(i+2);
-      (out+j+4)[0] = (uint)(i+2);
-      (out+j+4)[1] = (uint)(i+3);
-      (out+j+6)[0] = (uint)(i+3);
-      (out+j+6)[1] = (uint)(i+0);
-   }
-}
-static void generate_quadstrip_uint(
-    unsigned nr,
-    void *_out )
-{
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (uint)(i+2);
-      (out+j)[1] = (uint)(i+0);
-      (out+j+2)[0] = (uint)(i+0);
-      (out+j+2)[1] = (uint)(i+1);
-      (out+j+4)[0] = (uint)(i+1);
-      (out+j+4)[1] = (uint)(i+3);
-      (out+j+6)[0] = (uint)(i+3);
-      (out+j+6)[1] = (uint)(i+2);
-   }
-}
-static void generate_polygon_uint(
-    unsigned nr,
-    void *_out )
-{
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)(0);
-      (out+j)[1] = (uint)(i+1);
-      (out+j+2)[0] = (uint)(i+1);
-      (out+j+2)[1] = (uint)(i+2);
-      (out+j+4)[0] = (uint)(i+2);
-      (out+j+4)[1] = (uint)(0);
-   }
-}
-static void translate_tris_ubyte2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (ushort)in[i];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[i];
-   }
-}
-static void translate_tristrip_ubyte2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[i];
-      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
-      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
-      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
-      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
-      (out+j+4)[1] = (ushort)in[i];
-   }
-}
-static void translate_trifan_ubyte2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[0];
-   }
-}
-static void translate_quads_ubyte2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (ushort)in[i+0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[i+3];
-      (out+j+6)[0] = (ushort)in[i+3];
-      (out+j+6)[1] = (ushort)in[i+0];
-   }
-}
-static void translate_quadstrip_ubyte2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (ushort)in[i+2];
-      (out+j)[1] = (ushort)in[i+0];
-      (out+j+2)[0] = (ushort)in[i+0];
-      (out+j+2)[1] = (ushort)in[i+1];
-      (out+j+4)[0] = (ushort)in[i+1];
-      (out+j+4)[1] = (ushort)in[i+3];
-      (out+j+6)[0] = (ushort)in[i+3];
-      (out+j+6)[1] = (ushort)in[i+2];
-   }
-}
-static void translate_polygon_ubyte2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[0];
-   }
-}
-static void translate_tris_ubyte2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (uint)in[i];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[i];
-   }
-}
-static void translate_tristrip_ubyte2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[i];
-      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
-      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
-      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
-      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
-      (out+j+4)[1] = (uint)in[i];
-   }
-}
-static void translate_trifan_ubyte2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[0];
-   }
-}
-static void translate_quads_ubyte2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (uint)in[i+0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[i+3];
-      (out+j+6)[0] = (uint)in[i+3];
-      (out+j+6)[1] = (uint)in[i+0];
-   }
-}
-static void translate_quadstrip_ubyte2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (uint)in[i+2];
-      (out+j)[1] = (uint)in[i+0];
-      (out+j+2)[0] = (uint)in[i+0];
-      (out+j+2)[1] = (uint)in[i+1];
-      (out+j+4)[0] = (uint)in[i+1];
-      (out+j+4)[1] = (uint)in[i+3];
-      (out+j+6)[0] = (uint)in[i+3];
-      (out+j+6)[1] = (uint)in[i+2];
-   }
-}
-static void translate_polygon_ubyte2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ubyte*in = (const ubyte*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[0];
-   }
-}
-static void translate_tris_ushort2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (ushort)in[i];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[i];
-   }
-}
-static void translate_tristrip_ushort2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[i];
-      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
-      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
-      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
-      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
-      (out+j+4)[1] = (ushort)in[i];
-   }
-}
-static void translate_trifan_ushort2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[0];
-   }
-}
-static void translate_quads_ushort2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (ushort)in[i+0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[i+3];
-      (out+j+6)[0] = (ushort)in[i+3];
-      (out+j+6)[1] = (ushort)in[i+0];
-   }
-}
-static void translate_quadstrip_ushort2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (ushort)in[i+2];
-      (out+j)[1] = (ushort)in[i+0];
-      (out+j+2)[0] = (ushort)in[i+0];
-      (out+j+2)[1] = (ushort)in[i+1];
-      (out+j+4)[0] = (ushort)in[i+1];
-      (out+j+4)[1] = (ushort)in[i+3];
-      (out+j+6)[0] = (ushort)in[i+3];
-      (out+j+6)[1] = (ushort)in[i+2];
-   }
-}
-static void translate_polygon_ushort2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[0];
-   }
-}
-static void translate_tris_ushort2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (uint)in[i];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[i];
-   }
-}
-static void translate_tristrip_ushort2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[i];
-      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
-      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
-      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
-      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
-      (out+j+4)[1] = (uint)in[i];
-   }
-}
-static void translate_trifan_ushort2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[0];
-   }
-}
-static void translate_quads_ushort2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (uint)in[i+0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[i+3];
-      (out+j+6)[0] = (uint)in[i+3];
-      (out+j+6)[1] = (uint)in[i+0];
-   }
-}
-static void translate_quadstrip_ushort2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (uint)in[i+2];
-      (out+j)[1] = (uint)in[i+0];
-      (out+j+2)[0] = (uint)in[i+0];
-      (out+j+2)[1] = (uint)in[i+1];
-      (out+j+4)[0] = (uint)in[i+1];
-      (out+j+4)[1] = (uint)in[i+3];
-      (out+j+6)[0] = (uint)in[i+3];
-      (out+j+6)[1] = (uint)in[i+2];
-   }
-}
-static void translate_polygon_ushort2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const ushort*in = (const ushort*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[0];
-   }
-}
-static void translate_tris_uint2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (ushort)in[i];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[i];
-   }
-}
-static void translate_tristrip_uint2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[i];
-      (out+j)[1] = (ushort)in[i+1/*+(i&1)*/];
-      (out+j+2)[0] = (ushort)in[i+1/*+(i&1)*/];
-      (out+j+2)[1] = (ushort)in[i+2/*-(i&1)*/];
-      (out+j+4)[0] = (ushort)in[i+2/*-(i&1)*/];
-      (out+j+4)[1] = (ushort)in[i];
-   }
-}
-static void translate_trifan_uint2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[0];
-   }
-}
-static void translate_quads_uint2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (ushort)in[i+0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[i+3];
-      (out+j+6)[0] = (ushort)in[i+3];
-      (out+j+6)[1] = (ushort)in[i+0];
-   }
-}
-static void translate_quadstrip_uint2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (ushort)in[i+2];
-      (out+j)[1] = (ushort)in[i+0];
-      (out+j+2)[0] = (ushort)in[i+0];
-      (out+j+2)[1] = (ushort)in[i+1];
-      (out+j+4)[0] = (ushort)in[i+1];
-      (out+j+4)[1] = (ushort)in[i+3];
-      (out+j+6)[0] = (ushort)in[i+3];
-      (out+j+6)[1] = (ushort)in[i+2];
-   }
-}
-static void translate_polygon_uint2ushort(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  ushort *out = (ushort*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (ushort)in[0];
-      (out+j)[1] = (ushort)in[i+1];
-      (out+j+2)[0] = (ushort)in[i+1];
-      (out+j+2)[1] = (ushort)in[i+2];
-      (out+j+4)[0] = (ushort)in[i+2];
-      (out+j+4)[1] = (ushort)in[0];
-   }
-}
-static void translate_tris_uint2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i+=3) { 
-      (out+j)[0] = (uint)in[i];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[i];
-   }
-}
-static void translate_tristrip_uint2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[i];
-      (out+j)[1] = (uint)in[i+1/*+(i&1)*/];
-      (out+j+2)[0] = (uint)in[i+1/*+(i&1)*/];
-      (out+j+2)[1] = (uint)in[i+2/*-(i&1)*/];
-      (out+j+4)[0] = (uint)in[i+2/*-(i&1)*/];
-      (out+j+4)[1] = (uint)in[i];
-   }
-}
-static void translate_trifan_uint2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[0];
-   }
-}
-static void translate_quads_uint2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=4) { 
-      (out+j)[0] = (uint)in[i+0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[i+3];
-      (out+j+6)[0] = (uint)in[i+3];
-      (out+j+6)[1] = (uint)in[i+0];
-   }
-}
-static void translate_quadstrip_uint2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=8, i+=2) { 
-      (out+j)[0] = (uint)in[i+2];
-      (out+j)[1] = (uint)in[i+0];
-      (out+j+2)[0] = (uint)in[i+0];
-      (out+j+2)[1] = (uint)in[i+1];
-      (out+j+4)[0] = (uint)in[i+1];
-      (out+j+4)[1] = (uint)in[i+3];
-      (out+j+6)[0] = (uint)in[i+3];
-      (out+j+6)[1] = (uint)in[i+2];
-   }
-}
-static void translate_polygon_uint2uint(
-    const void * _in,
-    unsigned nr,
-    void *_out )
-{
-  const uint*in = (const uint*)_in;
-  uint *out = (uint*)_out;
-  unsigned i, j;
-  (void)j;
-  for (j = i = 0; j < nr; j+=6, i++) { 
-      (out+j)[0] = (uint)in[0];
-      (out+j)[1] = (uint)in[i+1];
-      (out+j+2)[0] = (uint)in[i+1];
-      (out+j+2)[1] = (uint)in[i+2];
-      (out+j+4)[0] = (uint)in[i+2];
-      (out+j+4)[1] = (uint)in[0];
-   }
-}
-void u_unfilled_init( void )
-{
-  static int firsttime = 1;
-  if (!firsttime) return;
-  firsttime = 0;
-generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLES] = generate_tris_ushort;
-generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_ushort;
-generate_line[OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_ushort;
-generate_line[OUT_USHORT][PIPE_PRIM_QUADS] = generate_quads_ushort;
-generate_line[OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_ushort;
-generate_line[OUT_USHORT][PIPE_PRIM_POLYGON] = generate_polygon_ushort;
-generate_line[OUT_UINT][PIPE_PRIM_TRIANGLES] = generate_tris_uint;
-generate_line[OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = generate_trifan_uint;
-generate_line[OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = generate_tristrip_uint;
-generate_line[OUT_UINT][PIPE_PRIM_QUADS] = generate_quads_uint;
-generate_line[OUT_UINT][PIPE_PRIM_QUAD_STRIP] = generate_quadstrip_uint;
-generate_line[OUT_UINT][PIPE_PRIM_POLYGON] = generate_polygon_uint;
-translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2ushort;
-translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2ushort;
-translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2ushort;
-translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_ubyte2ushort;
-translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2ushort;
-translate_line[IN_UBYTE][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2ushort;
-translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_ubyte2uint;
-translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ubyte2uint;
-translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ubyte2uint;
-translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_ubyte2uint;
-translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ubyte2uint;
-translate_line[IN_UBYTE][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_ubyte2uint;
-translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2ushort;
-translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2ushort;
-translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2ushort;
-translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_ushort2ushort;
-translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2ushort;
-translate_line[IN_USHORT][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_ushort2ushort;
-translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_ushort2uint;
-translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_ushort2uint;
-translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_ushort2uint;
-translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_ushort2uint;
-translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_ushort2uint;
-translate_line[IN_USHORT][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_ushort2uint;
-translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLES] = translate_tris_uint2ushort;
-translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2ushort;
-translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2ushort;
-translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_QUADS] = translate_quads_uint2ushort;
-translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2ushort;
-translate_line[IN_UINT][OUT_USHORT][PIPE_PRIM_POLYGON] = translate_polygon_uint2ushort;
-translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLES] = translate_tris_uint2uint;
-translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLE_FAN] = translate_trifan_uint2uint;
-translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_TRIANGLE_STRIP] = translate_tristrip_uint2uint;
-translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_QUADS] = translate_quads_uint2uint;
-translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_QUAD_STRIP] = translate_quadstrip_uint2uint;
-translate_line[IN_UINT][OUT_UINT][PIPE_PRIM_POLYGON] = translate_polygon_uint2uint;
-}
-#include "indices/u_unfilled_indices.c"

src/gallium/auxiliary/rtasm/rtasm_x86sse.c

@@ -1757,6 +1757,8 @@ void (*x86_get_func( struct x86_function *p ))(void)
 
 #else
 
+void x86sse_dummy( void );
+
 void x86sse_dummy( void )
 {
 }

src/gallium/auxiliary/tgsi/tgsi_ureg.c

@@ -1489,6 +1489,12 @@ const struct tgsi_token *ureg_get_tokens( struct ureg_program *ureg,
 }
 
 
+void ureg_free_tokens( const struct tgsi_token *tokens )
+{
+   FREE((struct tgsi_token *)tokens);
+}
+
+
 struct ureg_program *ureg_create( unsigned processor )
 {
    struct ureg_program *ureg = CALLOC_STRUCT( ureg_program );

src/gallium/auxiliary/tgsi/tgsi_ureg.h

@@ -104,6 +104,10 @@ ureg_get_tokens( struct ureg_program *ureg,
                  unsigned *nr_tokens );
 
 
+/* Free the tokens created by ureg_get_tokens() */
+void ureg_free_tokens( const struct tgsi_token *tokens );
+
+
 void 
 ureg_destroy( struct ureg_program * );
 

src/gallium/auxiliary/util/u_atomic.h

@@ -29,6 +29,8 @@
 #define PIPE_ATOMIC_ASM_MSVC_X86                
 #elif (defined(PIPE_CC_GCC) && defined(PIPE_ARCH_X86))
 #define PIPE_ATOMIC_ASM_GCC_X86
+#elif (defined(PIPE_CC_GCC) && defined(PIPE_ARCH_X86_64))
+#define PIPE_ATOMIC_ASM_GCC_X86_64
 #elif defined(PIPE_CC_GCC) && (PIPE_CC_GCC_VERSION >= 401)
 #define PIPE_ATOMIC_GCC_INTRINSIC
 #else
@@ -36,6 +38,51 @@
 #endif
 
 
+#if defined(PIPE_ATOMIC_ASM_GCC_X86_64)
+#define PIPE_ATOMIC "GCC x86_64 assembly"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#define p_atomic_set(_v, _i) (*(_v) = (_i))
+#define p_atomic_read(_v) (*(_v))
+
+static INLINE boolean
+p_atomic_dec_zero(int32_t *v)
+{
+   unsigned char c;
+
+   __asm__ __volatile__("lock; decl %0; sete %1":"+m"(*v), "=qm"(c)
+			::"memory");
+
+   return c != 0;
+}
+
+static INLINE void
+p_atomic_inc(int32_t *v)
+{
+   __asm__ __volatile__("lock; incl %0":"+m"(*v));
+}
+
+static INLINE void
+p_atomic_dec(int32_t *v)
+{
+   __asm__ __volatile__("lock; decl %0":"+m"(*v));
+}
+
+static INLINE int32_t
+p_atomic_cmpxchg(int32_t *v, int32_t old, int32_t _new)
+{
+   return __sync_val_compare_and_swap(v, old, _new);
+}
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* PIPE_ATOMIC_ASM_GCC_X86_64 */
+
 
 #if defined(PIPE_ATOMIC_ASM_GCC_X86)
 

src/gallium/drivers/cell/ppu/cell_context.c

@@ -74,7 +74,7 @@ cell_destroy_context( struct pipe_context *pipe )
 static struct draw_context *
 cell_draw_create(struct cell_context *cell)
 {
-   struct draw_context *draw = draw_create();
+   struct draw_context *draw = draw_create(&cell->pipe);
 
 #if 0 /* broken */
    if (getenv("GALLIUM_CELL_VS")) {

src/gallium/drivers/cell/ppu/cell_fence.c

@@ -92,7 +92,6 @@ cell_add_buffer_to_list(struct cell_context *cell,
                         struct cell_buffer_list *list,
                         struct pipe_buffer *buffer)
 {
-   struct pipe_screen *ps = cell->pipe.screen;
    struct cell_buffer_node *node = CALLOC_STRUCT(cell_buffer_node);
    /* create new list node which references the buffer, insert at head */
    if (node) {

src/gallium/drivers/cell/ppu/cell_pipe_state.c

@@ -197,7 +197,7 @@ cell_bind_rasterizer_state(struct pipe_context *pipe, void *rast)
    struct cell_context *cell = cell_context(pipe);
 
    /* pass-through to draw module */
-   draw_set_rasterizer_state(cell->draw, rasterizer);
+   draw_set_rasterizer_state(cell->draw, rasterizer, rast);
 
    cell->rasterizer = rasterizer;
 

src/gallium/state_trackers/egl/x11/glxinit.c

@@ -16,6 +16,8 @@
 
 #include "glxinit.h"
 
+#ifdef GLX_DIRECT_RENDERING
+
 typedef struct GLXGenericGetString
 {
    CARD8 reqType;
@@ -680,3 +682,5 @@ __glXInitialize(Display * dpy)
 
    return dpyPriv;
 }
+
+#endif /* GLX_DIRECT_RENDERING */

src/gallium/state_trackers/egl/x11/native_dri2.c

@@ -37,6 +37,8 @@
 #include "native_x11.h"
 #include "x11_screen.h"
 
+#ifdef GLX_DIRECT_RENDERING
+
 enum dri2_surface_type {
    DRI2_SURFACE_TYPE_WINDOW,
    DRI2_SURFACE_TYPE_PIXMAP,
@@ -878,3 +880,15 @@ x11_create_dri2_display(EGLNativeDisplayType dpy,
 
    return &dri2dpy->base;
 }
+
+#else /* GLX_DIRECT_RENDERING */
+
+struct native_display *
+x11_create_dri2_display(EGLNativeDisplayType dpy,
+                        struct native_event_handler *event_handler,
+                        struct drm_api *api)
+{
+   return NULL;
+}
+
+#endif /* GLX_DIRECT_RENDERING */

src/gallium/state_trackers/egl/x11/native_x11.c

@@ -70,7 +70,9 @@ native_create_probe(EGLNativeDisplayType dpy)
    xscr = x11_screen_create(xdpy, scr);
    if (xscr) {
       if (x11_screen_support(xscr, X11_SCREEN_EXTENSION_DRI2)) {
+#ifdef GLX_DIRECT_RENDERING
          driver_name = x11_screen_probe_dri2(xscr, NULL, NULL);
+#endif
          if (driver_name)
             nprobe->data = strdup(driver_name);
       }

src/gallium/state_trackers/egl/x11/x11_screen.c

@@ -39,8 +39,10 @@
 #include "glxinit.h"
 
 struct x11_screen {
+#ifdef GLX_DIRECT_RENDERING
    /* dummy base class */
    struct __GLXDRIdisplayRec base;
+#endif
 
    Display *dpy;
    int number;
@@ -103,15 +105,19 @@ x11_screen_destroy(struct x11_screen *xscr)
    if (xscr->dri_device)
       Xfree(xscr->dri_device);
 
+#ifdef GLX_DIRECT_RENDERING
    /* xscr->glx_dpy will be destroyed with the X display */
    if (xscr->glx_dpy)
       xscr->glx_dpy->dri2Display = NULL;
+#endif
 
    if (xscr->visuals)
       XFree(xscr->visuals);
    free(xscr);
 }
 
+#ifdef GLX_DIRECT_RENDERING
+
 static boolean
 x11_screen_init_dri2(struct x11_screen *xscr)
 {
@@ -133,6 +139,8 @@ x11_screen_init_glx(struct x11_screen *xscr)
    return (xscr->glx_dpy != NULL);
 }
 
+#endif /* GLX_DIRECT_RENDERING */
+
 /**
  * Return true if the screen supports the extension.
  */
@@ -145,12 +153,14 @@ x11_screen_support(struct x11_screen *xscr, enum x11_screen_extension ext)
    case X11_SCREEN_EXTENSION_XSHM:
       supported = XShmQueryExtension(xscr->dpy);
       break;
+#ifdef GLX_DIRECT_RENDERING
    case X11_SCREEN_EXTENSION_GLX:
       supported = x11_screen_init_glx(xscr);
       break;
    case X11_SCREEN_EXTENSION_DRI2:
       supported = x11_screen_init_dri2(xscr);
       break;
+#endif
    default:
       break;
    }
@@ -233,6 +243,39 @@ x11_screen_convert_visual(struct x11_screen *xscr, const XVisualInfo *visual,
    mode->xRenderable = TRUE;
 }
 
+/**
+ * Return the depth of a drawable.
+ *
+ * Unlike other drawable functions, the drawable needs not be a DRI2 drawable.
+ */
+uint
+x11_drawable_get_depth(struct x11_screen *xscr, Drawable drawable)
+{
+   unsigned int depth;
+
+   if (drawable != xscr->last_drawable) {
+      Window root;
+      int x, y;
+      unsigned int w, h, border;
+      Status ok;
+
+      ok = XGetGeometry(xscr->dpy, drawable, &root,
+            &x, &y, &w, &h, &border, &depth);
+      if (!ok)
+         depth = 0;
+
+      xscr->last_drawable = drawable;
+      xscr->last_depth = depth;
+   }
+   else {
+      depth = xscr->last_depth;
+   }
+
+   return depth;
+}
+
+#ifdef GLX_DIRECT_RENDERING
+
 /**
  * Return the GLX fbconfigs.
  */
@@ -391,37 +434,6 @@ x11_drawable_get_buffers(struct x11_screen *xscr, Drawable drawable,
    return (struct x11_drawable_buffer *) dri2bufs;
 }
 
-/**
- * Return the depth of a drawable.
- *
- * Unlike other drawable functions, the drawable needs not be a DRI2 drawable.
- */
-uint
-x11_drawable_get_depth(struct x11_screen *xscr, Drawable drawable)
-{
-   unsigned int depth;
-
-   if (drawable != xscr->last_drawable) {
-      Window root;
-      int x, y;
-      unsigned int w, h, border;
-      Status ok;
-
-      ok = XGetGeometry(xscr->dpy, drawable, &root,
-            &x, &y, &w, &h, &border, &depth);
-      if (!ok)
-         depth = 0;
-
-      xscr->last_drawable = drawable;
-      xscr->last_depth = depth;
-   }
-   else {
-      depth = xscr->last_depth;
-   }
-
-   return depth;
-}
-
 /**
  * Create a mode list of the given size.
  */
@@ -489,3 +501,5 @@ dri2InvalidateBuffers(Display *dpy, XID drawable)
 
    xscr->dri_invalidate_buffers(xscr, drawable, xscr->dri_user_data);
 }
+
+#endif /* GLX_DIRECT_RENDERING */

src/gallium/state_trackers/egl/x11/x11_screen.h

@@ -68,20 +68,18 @@ void
 x11_screen_convert_visual(struct x11_screen *xscr, const XVisualInfo *visual,
                           __GLcontextModes *mode);
 
+uint
+x11_drawable_get_depth(struct x11_screen *xscr, Drawable drawable);
+
+#ifdef GLX_DIRECT_RENDERING
+
+/* GLX */
 const __GLcontextModes *
 x11_screen_get_glx_configs(struct x11_screen *xscr);
 
 const __GLcontextModes *
 x11_screen_get_glx_visuals(struct x11_screen *xscr);
 
-const char *
-x11_screen_probe_dri2(struct x11_screen *xscr, int *major, int *minor);
-
-int
-x11_screen_enable_dri2(struct x11_screen *xscr,
-                       x11_drawable_invalidate_buffers invalidate_buffers,
-                       void *user_data);
-
 __GLcontextModes *
 x11_context_modes_create(unsigned count);
 
@@ -91,6 +89,15 @@ x11_context_modes_destroy(__GLcontextModes *modes);
 unsigned
 x11_context_modes_count(const __GLcontextModes *modes);
 
+/* DRI2 */
+const char *
+x11_screen_probe_dri2(struct x11_screen *xscr, int *major, int *minor);
+
+int
+x11_screen_enable_dri2(struct x11_screen *xscr,
+                       x11_drawable_invalidate_buffers invalidate_buffers,
+                       void *user_data);
+
 void
 x11_drawable_enable_dri2(struct x11_screen *xscr,
                          Drawable drawable, boolean on);
@@ -105,7 +112,6 @@ x11_drawable_get_buffers(struct x11_screen *xscr, Drawable drawable,
                          int *width, int *height, unsigned int *attachments,
                          boolean with_format, int num_ins, int *num_outs);
 
-uint
-x11_drawable_get_depth(struct x11_screen *xscr, Drawable drawable);
+#endif /* GLX_DIRECT_RENDERING */
 
 #endif /* _X11_SCREEN_H_ */

src/gallium/winsys/drm/Makefile.egl

@@ -11,7 +11,7 @@
 
 EGL_DRIVER_OBJECTS = $(EGL_DRIVER_SOURCES:.c=.o)
 
-common_LIBS = -ldrm -lm -ldl
+common_LIBS = -ldrm -lm $(DLOPEN_LIBS)
 
 x11_ST = $(TOP)/src/gallium/state_trackers/egl/libeglx11.a
 x11_LIBS = $(common_LIBS) -lX11 -lXext -lXfixes

src/glw/Makefile

@@ -17,7 +17,7 @@ OBJECTS = $(GLW_SOURCES:.c=.o)
 ##### RULES #####
 
 .c.o:
-	$(CC) -c $(INCDIRS) $(CFLAGS) $<
+	$(CC) -c $(INCDIRS) $(CFLAGS) $(GLW_CFLAGS) $<
 
 
 

src/glx/glxcmds.c

@@ -887,9 +887,6 @@ glXCopyContext(Display * dpy, GLXContext source,
 static Bool
 __glXIsDirect(Display * dpy, GLXContextID contextID)
 {
-#ifdef GLX_USE_APPLEGL /* TODO: apple indirect */
-   return GC_IS_DIRECT(gc);
-#else
 #if !defined(USE_XCB)
    xGLXIsDirectReq *req;
    xGLXIsDirectReply reply;
@@ -925,7 +922,6 @@ __glXIsDirect(Display * dpy, GLXContextID contextID)
 
    return reply.isDirect;
 #endif /* USE_XCB */
-#endif /* GLX_USE_APPLEGL */
 }
 
 /**
@@ -943,7 +939,11 @@ glXIsDirect(Display * dpy, GLXContext gc)
    else if (GC_IS_DIRECT(gc)) {
       return GL_TRUE;
    }
+#ifdef GLX_USE_APPLEGL  /* TODO: indirect on darwin */
+      return GL_FALSE;
+#else
    return __glXIsDirect(dpy, gc->xid);
+#endif
 }
 
 PUBLIC GLXPixmap
@@ -1766,6 +1766,15 @@ static int __glXQueryContextInfo(Display * dpy, GLXContext ctx)
 	 unsigned i;
 
          _XRead(dpy, (char *) propList, nPropListBytes);
+
+	 /* Look up screen first so we can look up visuals/fbconfigs later */
+         pProp = propList;
+         for (i = 0; i < numValues; i++, pProp += 2)
+	     if (pProp[0] == GLX_SCREEN) {
+		 ctx->screen = pProp[1];
+		 ctx->psc = GetGLXScreenConfigs(dpy, ctx->screen);
+	     }
+
          pProp = propList;
          for (i = 0; i < numValues; i++) {
             switch (*pProp++) {
@@ -1776,9 +1785,6 @@ static int __glXQueryContextInfo(Display * dpy, GLXContext ctx)
                ctx->mode =
                   _gl_context_modes_find_visual(ctx->psc->visuals, *pProp++);
                break;
-            case GLX_SCREEN:
-               ctx->screen = *pProp++;
-               break;
             case GLX_FBCONFIG_ID:
                ctx->mode =
                   _gl_context_modes_find_fbconfig(ctx->psc->configs,
@@ -1787,6 +1793,7 @@ static int __glXQueryContextInfo(Display * dpy, GLXContext ctx)
             case GLX_RENDER_TYPE:
                ctx->renderType = *pProp++;
                break;
+            case GLX_SCREEN:
             default:
                pProp++;
                continue;

src/glx/glxext.c

@@ -584,6 +584,10 @@ __glXInitializeVisualConfigFromTags(__GLcontextModes * config, int count,
          config->yInverted = *bp++;
          break;
 #endif
+      case GLX_USE_GL:
+         if (fbconfig_style_tags)
+            bp++;
+         break;
       case None:
          i = count;
          break;

src/mesa/drivers/common/meta.c

@@ -812,6 +812,21 @@ _mesa_meta_end(GLcontext *ctx)
 }
 
 
+/**
+ * Convert Z from a normalized value in the range [0, 1] to an object-space
+ * Z coordinate in [-1, +1] so that drawing at the new Z position with the
+ * default/identity ortho projection results in the original Z value.
+ * Used by the meta-Clear, Draw/CopyPixels and Bitmap functions where the Z
+ * value comes from the clear value or raster position.
+ */
+static INLINE GLfloat
+invert_z(GLfloat normZ)
+{
+   GLfloat objZ = 1.0 - 2.0 * normZ;
+   return objZ;
+}
+
+
 /**
  * One-time init for a temp_texture object.
  * Choose tex target, compute max tex size, etc.
@@ -1433,7 +1448,7 @@ _mesa_meta_Clear(GLcontext *ctx, GLbitfield buffers)
       const GLfloat y0 = (GLfloat) ctx->DrawBuffer->_Ymin;
       const GLfloat x1 = (GLfloat) ctx->DrawBuffer->_Xmax;
       const GLfloat y1 = (GLfloat) ctx->DrawBuffer->_Ymax;
-      const GLfloat z = 1.0 - 2.0 * ctx->Depth.Clear;
+      const GLfloat z = invert_z(ctx->Depth.Clear);
       GLuint i;
 
       verts[0].x = x0;
@@ -1539,7 +1554,7 @@ _mesa_meta_CopyPixels(GLcontext *ctx, GLint srcX, GLint srcY,
       const GLfloat dstY0 = (GLfloat) dstY;
       const GLfloat dstX1 = dstX + width * ctx->Pixel.ZoomX;
       const GLfloat dstY1 = dstY + height * ctx->Pixel.ZoomY;
-      const GLfloat z = ctx->Current.RasterPos[2];
+      const GLfloat z = invert_z(ctx->Current.RasterPos[2]);
 
       verts[0].x = dstX0;
       verts[0].y = dstY0;
@@ -1828,7 +1843,7 @@ _mesa_meta_DrawPixels(GLcontext *ctx,
       const GLfloat y0 = (GLfloat) y;
       const GLfloat x1 = x + width * ctx->Pixel.ZoomX;
       const GLfloat y1 = y + height * ctx->Pixel.ZoomY;
-      const GLfloat z = ctx->Current.RasterPos[2];
+      const GLfloat z = invert_z(ctx->Current.RasterPos[2]);
 
       verts[0].x = x0;
       verts[0].y = y0;
@@ -2031,7 +2046,7 @@ _mesa_meta_Bitmap(GLcontext *ctx,
       const GLfloat y0 = (GLfloat) y;
       const GLfloat x1 = (GLfloat) (x + width);
       const GLfloat y1 = (GLfloat) (y + height);
-      const GLfloat z = ctx->Current.RasterPos[2];
+      const GLfloat z = invert_z(ctx->Current.RasterPos[2]);
       GLuint i;
 
       verts[0].x = x0;

src/mesa/drivers/dri/i915/i915_tex_layout.c

@@ -484,7 +484,7 @@ i945_miptree_layout(struct intel_context *intel, struct intel_mipmap_tree * mt,
    case GL_TEXTURE_1D:
    case GL_TEXTURE_2D:
    case GL_TEXTURE_RECTANGLE_ARB:
-      i945_miptree_layout_2d(intel, mt, tiling);
+      i945_miptree_layout_2d(intel, mt, tiling, 1);
       break;
    default:
       _mesa_problem(NULL, "Unexpected tex target in i945_miptree_layout()");

src/mesa/drivers/dri/i915/i915_texstate.c

@@ -281,6 +281,7 @@ i915_update_tex_unit(struct intel_context *intel, GLuint unit, GLuint ss3)
       GLenum ws = tObj->WrapS;
       GLenum wt = tObj->WrapT;
       GLenum wr = tObj->WrapR;
+      float minlod;
 
       /* We program 1D textures as 2D textures, so the 2D texcoord could
        * result in sampling border values if we don't set the T wrap to
@@ -321,8 +322,9 @@ i915_update_tex_unit(struct intel_context *intel, GLuint unit, GLuint ss3)
           (translate_wrap_mode(wt) << SS3_TCY_ADDR_MODE_SHIFT) |
           (translate_wrap_mode(wr) << SS3_TCZ_ADDR_MODE_SHIFT));
 
+      minlod = MIN2(tObj->MinLod, tObj->_MaxLevel - tObj->BaseLevel);
       state[I915_TEXREG_SS3] |= (unit << SS3_TEXTUREMAP_INDEX_SHIFT);
-      state[I915_TEXREG_SS3] |= (U_FIXED(CLAMP(tObj->MinLod, 0.0, 11.0), 4) <<
+      state[I915_TEXREG_SS3] |= (U_FIXED(CLAMP(minlod, 0.0, 11.0), 4) <<
 				 SS3_MIN_LOD_SHIFT);
 
    }

src/mesa/drivers/dri/i915/i915_vtbl.c

@@ -627,8 +627,8 @@ i915_set_draw_region(struct intel_context *intel,
    state->Buffer[I915_DESTREG_DRAWRECT2] = 0;
    state->Buffer[I915_DESTREG_DRAWRECT3] = (draw_y << 16) | draw_x;
    state->Buffer[I915_DESTREG_DRAWRECT4] =
-      ((ctx->DrawBuffer->Width + draw_x) & 0xffff) |
-      ((ctx->DrawBuffer->Height + draw_y) << 16);
+      ((ctx->DrawBuffer->Width + draw_x - 1) & 0xffff) |
+      ((ctx->DrawBuffer->Height + draw_y - 1) << 16);
    state->Buffer[I915_DESTREG_DRAWRECT5] = (draw_y << 16) | draw_x;
 
    I915_STATECHANGE(i915, I915_UPLOAD_BUFFERS);

src/mesa/drivers/dri/i965/brw_clip.c

@@ -72,7 +72,7 @@ static void compile_clip_prog( struct brw_context *brw,
     */
    c.header_position_offset = ATTR_SIZE;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        delta = 3 * REG_SIZE;
    else
        delta = REG_SIZE;
@@ -85,7 +85,7 @@ static void compile_clip_prog( struct brw_context *brw,
 
    c.nr_attrs = brw_count_bits(c.key.attrs);
    
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        c.nr_regs = (c.nr_attrs + 1) / 2 + 3;  /* are vertices packed, or reg-aligned? */
    else
        c.nr_regs = (c.nr_attrs + 1) / 2 + 1;  /* are vertices packed, or reg-aligned? */
@@ -162,7 +162,7 @@ static void upload_clip_prog(struct brw_context *brw)
    /* _NEW_TRANSFORM */
    key.nr_userclip = brw_count_bits(ctx->Transform.ClipPlanesEnabled);
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        key.clip_mode = BRW_CLIPMODE_KERNEL_CLIP;
    else
        key.clip_mode = BRW_CLIPMODE_NORMAL;

src/mesa/drivers/dri/i965/brw_clip_state.c

@@ -102,10 +102,10 @@ clip_unit_create_from_key(struct brw_context *brw,
        */
       assert(key->nr_urb_entries % 2 == 0);
       
-      /* Although up to 16 concurrent Clip threads are allowed on IGDNG, 
+      /* Although up to 16 concurrent Clip threads are allowed on Ironlake,
        * only 2 threads can output VUEs at a time.
        */
-      if (intel->is_ironlake)
+      if (intel->gen == 5)
          clip.thread4.max_threads = 16 - 1;        
       else
          clip.thread4.max_threads = 2 - 1;

src/mesa/drivers/dri/i965/brw_clip_tri.c

@@ -78,7 +78,7 @@ void brw_clip_tri_alloc_regs( struct brw_clip_compile *c,
       for (j = 0; j < 3; j++) {
 	 GLuint delta = c->nr_attrs*16 + 32;
 
-         if (intel->is_ironlake)
+         if (intel->gen == 5)
              delta = c->nr_attrs * 16 + 32 * 3;
 
 	 brw_MOV(&c->func, byte_offset(c->reg.vertex[j], delta), brw_imm_f(0));
@@ -177,7 +177,7 @@ void brw_clip_tri_init_vertices( struct brw_clip_compile *c )
 void brw_clip_tri_flat_shade( struct brw_clip_compile *c )
 {
    struct brw_compile *p = &c->func;
-   struct brw_instruction *is_poly;
+   struct brw_instruction *is_poly, *is_trifan;
    struct brw_reg tmp0 = c->reg.loopcount; /* handy temporary */
 
    brw_AND(p, tmp0, get_element_ud(c->reg.R0, 2), brw_imm_ud(PRIM_MASK)); 
@@ -195,8 +195,22 @@ void brw_clip_tri_flat_shade( struct brw_clip_compile *c )
    is_poly = brw_ELSE(p, is_poly);
    {
       if (c->key.pv_first) {
-         brw_clip_copy_colors(c, 1, 0);
-         brw_clip_copy_colors(c, 2, 0);
+	 brw_CMP(p,
+		 vec1(brw_null_reg()),
+		 BRW_CONDITIONAL_EQ,
+		 tmp0,
+		 brw_imm_ud(_3DPRIM_TRIFAN));
+	 is_trifan = brw_IF(p, BRW_EXECUTE_1);
+	 {
+	    brw_clip_copy_colors(c, 0, 1);
+	    brw_clip_copy_colors(c, 2, 1);
+	 }
+	 is_trifan = brw_ELSE(p, is_trifan);
+	 {
+	    brw_clip_copy_colors(c, 1, 0);
+	    brw_clip_copy_colors(c, 2, 0);
+	 }
+	 brw_ENDIF(p, is_trifan);
       }
       else {
          brw_clip_copy_colors(c, 0, 2);

src/mesa/drivers/dri/i965/brw_clip_util.c

@@ -151,7 +151,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile *c,
    for (i = 0; i < c->nr_attrs; i++) {
       GLuint delta = i*16 + 32;
 
-      if (intel->is_ironlake)
+      if (intel->gen == 5)
           delta = i * 16 + 32 * 3;
 
       if (delta == c->offset[VERT_RESULT_EDGE]) {
@@ -185,7 +185,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile *c,
    if (i & 1) {
       GLuint delta = i*16 + 32;
 
-      if (intel->is_ironlake)
+      if (intel->gen == 5)
           delta = i * 16 + 32 * 3;
 
       brw_MOV(p, deref_4f(dest_ptr, delta), brw_imm_f(0));

src/mesa/drivers/dri/i965/brw_context.c

@@ -150,7 +150,7 @@ GLboolean brwCreateContext( const __GLcontextModes *mesaVis,
       MIN2(ctx->Const.FragmentProgram.MaxNativeParameters,
 	   ctx->Const.FragmentProgram.MaxEnvParams);
 
-   if (intel->is_ironlake || intel->is_g4x || intel->gen >= 6) {
+   if (intel->is_g4x || intel->gen >= 5) {
       brw->CMD_VF_STATISTICS = CMD_VF_STATISTICS_GM45;
       brw->CMD_PIPELINE_SELECT = CMD_PIPELINE_SELECT_GM45;
       brw->has_surface_tile_offset = GL_TRUE;
@@ -162,7 +162,7 @@ GLboolean brwCreateContext( const __GLcontextModes *mesaVis,
    }
 
    /* WM maximum threads is number of EUs times number of threads per EU. */
-   if (intel->is_ironlake) {
+   if (intel->gen == 5) {
       brw->urb.size = 1024;
       brw->vs_max_threads = 72;
       brw->wm_max_threads = 12 * 6;

src/mesa/drivers/dri/i965/brw_defines.h

@@ -674,12 +674,12 @@
 #define BRW_SAMPLER_MESSAGE_SIMD8_LD                  3
 #define BRW_SAMPLER_MESSAGE_SIMD16_LD                 3
 
-#define BRW_SAMPLER_MESSAGE_SAMPLE_IGDNG            0
-#define BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_IGDNG       1
-#define BRW_SAMPLER_MESSAGE_SAMPLE_LOD_IGDNG        2
-#define BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_IGDNG    3
+#define BRW_SAMPLER_MESSAGE_SAMPLE_GEN5            0
+#define BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_GEN5       1
+#define BRW_SAMPLER_MESSAGE_SAMPLE_LOD_GEN5        2
+#define BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_GEN5    3
 
-/* for IGDNG only */
+/* for GEN5 only */
 #define BRW_SAMPLER_SIMD_MODE_SIMD4X2                   0
 #define BRW_SAMPLER_SIMD_MODE_SIMD8                     1
 #define BRW_SAMPLER_SIMD_MODE_SIMD16                    2

src/mesa/drivers/dri/i965/brw_draw_upload.c

@@ -59,7 +59,7 @@ static GLuint half_float_types[5] = {
    0,
    BRW_SURFACEFORMAT_R16_FLOAT,
    BRW_SURFACEFORMAT_R16G16_FLOAT,
-   0, /* can't seem to render this one */
+   BRW_SURFACEFORMAT_R16G16B16A16_FLOAT,
    BRW_SURFACEFORMAT_R16G16B16A16_FLOAT
 };
 
@@ -522,7 +522,7 @@ static void brw_emit_vertices(struct brw_context *brw)
       OUT_RELOC(input->bo,
 		I915_GEM_DOMAIN_VERTEX, 0,
 		input->offset);
-      if (intel->is_ironlake || intel->gen >= 6) {
+      if (intel->gen >= 5) {
 	 OUT_RELOC(input->bo,
 		   I915_GEM_DOMAIN_VERTEX, 0,
 		   input->bo->size - 1);
@@ -565,7 +565,7 @@ static void brw_emit_vertices(struct brw_context *brw)
 		   (0 << BRW_VE0_SRC_OFFSET_SHIFT));
       }
 
-      if (intel->is_ironlake || intel->gen >= 6)
+      if (intel->gen >= 5)
           OUT_BATCH((comp0 << BRW_VE1_COMPONENT_0_SHIFT) |
                     (comp1 << BRW_VE1_COMPONENT_1_SHIFT) |
                     (comp2 << BRW_VE1_COMPONENT_2_SHIFT) |

src/mesa/drivers/dri/i965/brw_eu_emit.c

@@ -253,19 +253,19 @@ static void brw_set_math_message( struct brw_context *brw,
    struct intel_context *intel = &brw->intel;
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (intel->is_ironlake) {
-       insn->bits3.math_igdng.function = function;
-       insn->bits3.math_igdng.int_type = integer_type;
-       insn->bits3.math_igdng.precision = low_precision;
-       insn->bits3.math_igdng.saturate = saturate;
-       insn->bits3.math_igdng.data_type = dataType;
-       insn->bits3.math_igdng.snapshot = 0;
-       insn->bits3.math_igdng.header_present = 0;
-       insn->bits3.math_igdng.response_length = response_length;
-       insn->bits3.math_igdng.msg_length = msg_length;
-       insn->bits3.math_igdng.end_of_thread = 0;
-       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_MATH;
-       insn->bits2.send_igdng.end_of_thread = 0;
+   if (intel->gen == 5) {
+       insn->bits3.math_gen5.function = function;
+       insn->bits3.math_gen5.int_type = integer_type;
+       insn->bits3.math_gen5.precision = low_precision;
+       insn->bits3.math_gen5.saturate = saturate;
+       insn->bits3.math_gen5.data_type = dataType;
+       insn->bits3.math_gen5.snapshot = 0;
+       insn->bits3.math_gen5.header_present = 0;
+       insn->bits3.math_gen5.response_length = response_length;
+       insn->bits3.math_gen5.msg_length = msg_length;
+       insn->bits3.math_gen5.end_of_thread = 0;
+       insn->bits2.send_gen5.sfid = BRW_MESSAGE_TARGET_MATH;
+       insn->bits2.send_gen5.end_of_thread = 0;
    } else {
        insn->bits3.math.function = function;
        insn->bits3.math.int_type = integer_type;
@@ -293,18 +293,18 @@ static void brw_set_ff_sync_message( struct brw_context *brw,
 {
 	brw_set_src1(insn, brw_imm_d(0));
 
-	insn->bits3.urb_igdng.opcode = 1;
-	insn->bits3.urb_igdng.offset = offset;
-	insn->bits3.urb_igdng.swizzle_control = swizzle_control;
-	insn->bits3.urb_igdng.allocate = allocate;
-	insn->bits3.urb_igdng.used = used;
-	insn->bits3.urb_igdng.complete = complete;
-	insn->bits3.urb_igdng.header_present = 1;
-	insn->bits3.urb_igdng.response_length = response_length;
-	insn->bits3.urb_igdng.msg_length = msg_length;
-	insn->bits3.urb_igdng.end_of_thread = end_of_thread;
-	insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_URB;
-	insn->bits2.send_igdng.end_of_thread = end_of_thread;
+	insn->bits3.urb_gen5.opcode = 1;
+	insn->bits3.urb_gen5.offset = offset;
+	insn->bits3.urb_gen5.swizzle_control = swizzle_control;
+	insn->bits3.urb_gen5.allocate = allocate;
+	insn->bits3.urb_gen5.used = used;
+	insn->bits3.urb_gen5.complete = complete;
+	insn->bits3.urb_gen5.header_present = 1;
+	insn->bits3.urb_gen5.response_length = response_length;
+	insn->bits3.urb_gen5.msg_length = msg_length;
+	insn->bits3.urb_gen5.end_of_thread = end_of_thread;
+	insn->bits2.send_gen5.sfid = BRW_MESSAGE_TARGET_URB;
+	insn->bits2.send_gen5.end_of_thread = end_of_thread;
 }
 
 static void brw_set_urb_message( struct brw_context *brw,
@@ -321,17 +321,17 @@ static void brw_set_urb_message( struct brw_context *brw,
     struct intel_context *intel = &brw->intel;
     brw_set_src1(insn, brw_imm_d(0));
 
-    if (intel->is_ironlake || intel->gen >= 6) {
-        insn->bits3.urb_igdng.opcode = 0;	/* ? */
-        insn->bits3.urb_igdng.offset = offset;
-        insn->bits3.urb_igdng.swizzle_control = swizzle_control;
-        insn->bits3.urb_igdng.allocate = allocate;
-        insn->bits3.urb_igdng.used = used;	/* ? */
-        insn->bits3.urb_igdng.complete = complete;
-        insn->bits3.urb_igdng.header_present = 1;
-        insn->bits3.urb_igdng.response_length = response_length;
-        insn->bits3.urb_igdng.msg_length = msg_length;
-        insn->bits3.urb_igdng.end_of_thread = end_of_thread;
+    if (intel->gen >= 5) {
+        insn->bits3.urb_gen5.opcode = 0;	/* ? */
+        insn->bits3.urb_gen5.offset = offset;
+        insn->bits3.urb_gen5.swizzle_control = swizzle_control;
+        insn->bits3.urb_gen5.allocate = allocate;
+        insn->bits3.urb_gen5.used = used;	/* ? */
+        insn->bits3.urb_gen5.complete = complete;
+        insn->bits3.urb_gen5.header_present = 1;
+        insn->bits3.urb_gen5.response_length = response_length;
+        insn->bits3.urb_gen5.msg_length = msg_length;
+        insn->bits3.urb_gen5.end_of_thread = end_of_thread;
 	if (intel->gen >= 6) {
 	   /* For SNB, the SFID bits moved to the condmod bits, and
 	    * EOT stayed in bits3 above.  Does the EOT bit setting
@@ -339,8 +339,8 @@ static void brw_set_urb_message( struct brw_context *brw,
 	    */
 	   insn->header.destreg__conditionalmod = BRW_MESSAGE_TARGET_URB;
 	} else {
-	   insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_URB;
-	   insn->bits2.send_igdng.end_of_thread = end_of_thread;
+	   insn->bits2.send_gen5.sfid = BRW_MESSAGE_TARGET_URB;
+	   insn->bits2.send_gen5.end_of_thread = end_of_thread;
 	}
     } else {
         insn->bits3.urb.opcode = 0;	/* ? */
@@ -369,18 +369,18 @@ static void brw_set_dp_write_message( struct brw_context *brw,
    struct intel_context *intel = &brw->intel;
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (intel->is_ironlake) {
-       insn->bits3.dp_write_igdng.binding_table_index = binding_table_index;
-       insn->bits3.dp_write_igdng.msg_control = msg_control;
-       insn->bits3.dp_write_igdng.pixel_scoreboard_clear = pixel_scoreboard_clear;
-       insn->bits3.dp_write_igdng.msg_type = msg_type;
-       insn->bits3.dp_write_igdng.send_commit_msg = 0;
-       insn->bits3.dp_write_igdng.header_present = 1;
-       insn->bits3.dp_write_igdng.response_length = response_length;
-       insn->bits3.dp_write_igdng.msg_length = msg_length;
-       insn->bits3.dp_write_igdng.end_of_thread = end_of_thread;
-       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_DATAPORT_WRITE;
-       insn->bits2.send_igdng.end_of_thread = end_of_thread;
+   if (intel->gen == 5) {
+       insn->bits3.dp_write_gen5.binding_table_index = binding_table_index;
+       insn->bits3.dp_write_gen5.msg_control = msg_control;
+       insn->bits3.dp_write_gen5.pixel_scoreboard_clear = pixel_scoreboard_clear;
+       insn->bits3.dp_write_gen5.msg_type = msg_type;
+       insn->bits3.dp_write_gen5.send_commit_msg = 0;
+       insn->bits3.dp_write_gen5.header_present = 1;
+       insn->bits3.dp_write_gen5.response_length = response_length;
+       insn->bits3.dp_write_gen5.msg_length = msg_length;
+       insn->bits3.dp_write_gen5.end_of_thread = end_of_thread;
+       insn->bits2.send_gen5.sfid = BRW_MESSAGE_TARGET_DATAPORT_WRITE;
+       insn->bits2.send_gen5.end_of_thread = end_of_thread;
    } else {
        insn->bits3.dp_write.binding_table_index = binding_table_index;
        insn->bits3.dp_write.msg_control = msg_control;
@@ -407,18 +407,18 @@ static void brw_set_dp_read_message( struct brw_context *brw,
    struct intel_context *intel = &brw->intel;
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (intel->is_ironlake) {
-       insn->bits3.dp_read_igdng.binding_table_index = binding_table_index;
-       insn->bits3.dp_read_igdng.msg_control = msg_control;
-       insn->bits3.dp_read_igdng.msg_type = msg_type;
-       insn->bits3.dp_read_igdng.target_cache = target_cache;
-       insn->bits3.dp_read_igdng.header_present = 1;
-       insn->bits3.dp_read_igdng.response_length = response_length;
-       insn->bits3.dp_read_igdng.msg_length = msg_length;
-       insn->bits3.dp_read_igdng.pad1 = 0;
-       insn->bits3.dp_read_igdng.end_of_thread = end_of_thread;
-       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_DATAPORT_READ;
-       insn->bits2.send_igdng.end_of_thread = end_of_thread;
+   if (intel->gen == 5) {
+       insn->bits3.dp_read_gen5.binding_table_index = binding_table_index;
+       insn->bits3.dp_read_gen5.msg_control = msg_control;
+       insn->bits3.dp_read_gen5.msg_type = msg_type;
+       insn->bits3.dp_read_gen5.target_cache = target_cache;
+       insn->bits3.dp_read_gen5.header_present = 1;
+       insn->bits3.dp_read_gen5.response_length = response_length;
+       insn->bits3.dp_read_gen5.msg_length = msg_length;
+       insn->bits3.dp_read_gen5.pad1 = 0;
+       insn->bits3.dp_read_gen5.end_of_thread = end_of_thread;
+       insn->bits2.send_gen5.sfid = BRW_MESSAGE_TARGET_DATAPORT_READ;
+       insn->bits2.send_gen5.end_of_thread = end_of_thread;
    } else {
        insn->bits3.dp_read.binding_table_index = binding_table_index; /*0:7*/
        insn->bits3.dp_read.msg_control = msg_control;  /*8:11*/
@@ -447,17 +447,17 @@ static void brw_set_sampler_message(struct brw_context *brw,
    assert(eot == 0);
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (intel->is_ironlake) {
-      insn->bits3.sampler_igdng.binding_table_index = binding_table_index;
-      insn->bits3.sampler_igdng.sampler = sampler;
-      insn->bits3.sampler_igdng.msg_type = msg_type;
-      insn->bits3.sampler_igdng.simd_mode = simd_mode;
-      insn->bits3.sampler_igdng.header_present = header_present;
-      insn->bits3.sampler_igdng.response_length = response_length;
-      insn->bits3.sampler_igdng.msg_length = msg_length;
-      insn->bits3.sampler_igdng.end_of_thread = eot;
-      insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_SAMPLER;
-      insn->bits2.send_igdng.end_of_thread = eot;
+   if (intel->gen == 5) {
+      insn->bits3.sampler_gen5.binding_table_index = binding_table_index;
+      insn->bits3.sampler_gen5.sampler = sampler;
+      insn->bits3.sampler_gen5.msg_type = msg_type;
+      insn->bits3.sampler_gen5.simd_mode = simd_mode;
+      insn->bits3.sampler_gen5.header_present = header_present;
+      insn->bits3.sampler_gen5.response_length = response_length;
+      insn->bits3.sampler_gen5.msg_length = msg_length;
+      insn->bits3.sampler_gen5.end_of_thread = eot;
+      insn->bits2.send_gen5.sfid = BRW_MESSAGE_TARGET_SAMPLER;
+      insn->bits2.send_gen5.end_of_thread = eot;
    } else if (intel->is_g4x) {
       insn->bits3.sampler_g4x.binding_table_index = binding_table_index;
       insn->bits3.sampler_g4x.sampler = sampler;
@@ -663,7 +663,7 @@ struct brw_instruction *brw_ELSE(struct brw_compile *p,
    struct brw_instruction *insn;
    GLuint br = 1;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       br = 2;
 
    if (p->single_program_flow) {
@@ -705,7 +705,7 @@ void brw_ENDIF(struct brw_compile *p,
    struct intel_context *intel = &p->brw->intel;
    GLuint br = 1;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       br = 2; 
  
    if (p->single_program_flow) {
@@ -820,7 +820,7 @@ struct brw_instruction *brw_WHILE(struct brw_compile *p,
    struct brw_instruction *insn;
    GLuint br = 1;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       br = 2;
 
    if (p->single_program_flow)
@@ -864,7 +864,7 @@ void brw_land_fwd_jump(struct brw_compile *p,
    struct brw_instruction *landing = &p->store[p->nr_insn];
    GLuint jmpi = 1;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        jmpi = 2;
 
    assert(jmp_insn->header.opcode == BRW_OPCODE_JMPI);

src/mesa/drivers/dri/i965/brw_gs.c

@@ -60,7 +60,7 @@ static void compile_gs_prog( struct brw_context *brw,
     */
    c.nr_attrs = brw_count_bits(c.key.attrs);
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        c.nr_regs = (c.nr_attrs + 1) / 2 + 3;  /* are vertices packed, or reg-aligned? */
    else
        c.nr_regs = (c.nr_attrs + 1) / 2 + 1;  /* are vertices packed, or reg-aligned? */

src/mesa/drivers/dri/i965/brw_gs_state.c

@@ -98,7 +98,7 @@ gs_unit_create_from_key(struct brw_context *brw, struct brw_gs_unit_key *key)
    else
       gs.thread4.max_threads = 0;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       gs.thread4.rendering_enable = 1;
 
    if (INTEL_DEBUG & DEBUG_STATS)

src/mesa/drivers/dri/i965/brw_misc_state.c

@@ -248,7 +248,7 @@ static void emit_depthbuffer(struct brw_context *brw)
 
    if (intel->gen >= 6)
       len = 7;
-   else if (intel->is_g4x || intel->is_ironlake)
+   else if (intel->is_g4x || intel->gen == 5)
       len = 6;
    else
       len = 5;
@@ -262,7 +262,7 @@ static void emit_depthbuffer(struct brw_context *brw)
       OUT_BATCH(0);
       OUT_BATCH(0);
 
-      if (intel->is_g4x || intel->is_ironlake || intel->gen >= 6)
+      if (intel->is_g4x || intel->gen >= 5)
          OUT_BATCH(0);
 
       if (intel->gen >= 6)
@@ -306,7 +306,7 @@ static void emit_depthbuffer(struct brw_context *brw)
 		((region->height - 1) << 19));
       OUT_BATCH(0);
 
-      if (intel->is_g4x || intel->is_ironlake || intel->gen >= 6)
+      if (intel->is_g4x || intel->gen >= 5)
          OUT_BATCH(0);
 
       if (intel->gen >= 6)
@@ -608,7 +608,7 @@ static void upload_state_base_address( struct brw_context *brw )
        OUT_BATCH(1); /* Indirect object upper bound */
        OUT_BATCH(1); /* Instruction access upper bound */
        ADVANCE_BATCH();
-   } else if (intel->is_ironlake) {
+   } else if (intel->gen == 5) {
        BEGIN_BATCH(8);
        OUT_BATCH(CMD_STATE_BASE_ADDRESS << 16 | (8 - 2));
        OUT_BATCH(1); /* General state base address */

src/mesa/drivers/dri/i965/brw_program.c

@@ -34,6 +34,7 @@
 #include "shader/prog_parameter.h"
 #include "shader/program.h"
 #include "shader/programopt.h"
+#include "shader/shader_api.h"
 #include "tnl/tnl.h"
 
 #include "brw_context.h"
@@ -119,12 +120,28 @@ static GLboolean brwIsProgramNative( GLcontext *ctx,
    return GL_TRUE;
 }
 
+static void
+shader_error(GLcontext *ctx, struct gl_program *prog, const char *msg)
+{
+   struct gl_shader_program *shader;
+
+   shader = _mesa_lookup_shader_program(ctx, prog->Id);
+
+   if (shader) {
+      if (shader->InfoLog) {
+	 free(shader->InfoLog);
+      }
+      shader->InfoLog = _mesa_strdup(msg);
+      shader->LinkStatus = GL_FALSE;
+   }
+}
 
 static GLboolean brwProgramStringNotify( GLcontext *ctx,
                                          GLenum target,
                                          struct gl_program *prog )
 {
    struct brw_context *brw = brw_context(ctx);
+   int i;
 
    if (target == GL_FRAGMENT_PROGRAM_ARB) {
       struct gl_fragment_program *fprog = (struct gl_fragment_program *) prog;
@@ -160,7 +177,22 @@ static GLboolean brwProgramStringNotify( GLcontext *ctx,
       _tnl_program_string(ctx, target, prog);
    }
 
-   /* XXX check if program is legal, within limits */
+   /* Reject programs with subroutines, which are totally broken at the moment
+    * (all program flows return when any program flow returns, and
+    * the VS also hangs if a function call calls a function.
+    *
+    * See piglit glsl-{vs,fs}-functions-[23] tests.
+    */
+   for (i = 0; i < prog->NumInstructions; i++) {
+      if (prog->Instructions[i].Opcode == OPCODE_CAL) {
+	 shader_error(ctx, prog,
+		      "i965 driver doesn't yet support uninlined function "
+		      "calls.  Move to using a single return statement at "
+		      "the end of the function to work around it.");
+	 return GL_FALSE;
+      }
+   }
+
    return GL_TRUE;
 }
 

src/mesa/drivers/dri/i965/brw_sf_emit.c

@@ -162,7 +162,7 @@ static void do_flatshade_triangle( struct brw_sf_compile *c )
    if (c->key.primitive == SF_UNFILLED_TRIS)
       return;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        jmpi = 2;
 
    brw_push_insn_state(p);
@@ -201,7 +201,7 @@ static void do_flatshade_line( struct brw_sf_compile *c )
    if (c->key.primitive == SF_UNFILLED_TRIS)
       return;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        jmpi = 2;
 
    brw_push_insn_state(p);

src/mesa/drivers/dri/i965/brw_sf_state.c

@@ -76,7 +76,20 @@ static void upload_sf_vp(struct brw_context *brw)
     * Note that the hardware's coordinates are inclusive, while Mesa's min is
     * inclusive but max is exclusive.
     */
-   if (render_to_fbo) {
+
+   if (ctx->DrawBuffer->_Xmin == ctx->DrawBuffer->_Xmax ||
+       ctx->DrawBuffer->_Ymin == ctx->DrawBuffer->_Ymax) {
+      /* If the scissor was out of bounds and got clamped to 0
+       * width/height at the bounds, the subtraction of 1 from
+       * maximums could produce a negative number and thus not clip
+       * anything.  Instead, just provide a min > max scissor inside
+       * the bounds, which produces the expected no rendering.
+       */
+      sfv.scissor.xmin = 1;
+      sfv.scissor.xmax = 0;
+      sfv.scissor.ymin = 1;
+      sfv.scissor.ymax = 0;
+   } else if (render_to_fbo) {
       /* texmemory: Y=0=bottom */
       sfv.scissor.xmin = ctx->DrawBuffer->_Xmin;
       sfv.scissor.xmax = ctx->DrawBuffer->_Xmax - 1;
@@ -177,7 +190,7 @@ sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
 
    sf.thread3.dispatch_grf_start_reg = 3;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
        sf.thread3.urb_entry_read_offset = 3;
    else
        sf.thread3.urb_entry_read_offset = 1;
@@ -190,7 +203,7 @@ sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
    /* Each SF thread produces 1 PUE, and there can be up to 24 (Pre-Ironlake) or
     * 48 (Ironlake) threads.
     */
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       chipset_max_threads = 48;
    else
       chipset_max_threads = 24;

src/mesa/drivers/dri/i965/brw_structs.h

@@ -925,7 +925,7 @@ struct brw_gs_unit_state
    struct
    {
       GLuint pad0:8;
-      GLuint rendering_enable:1; /* for IGDNG */
+      GLuint rendering_enable:1; /* for Ironlake */
       GLuint pad4:1;
       GLuint stats_enable:1; 
       GLuint nr_urb_entries:7; 
@@ -1035,7 +1035,7 @@ struct brw_wm_unit_state
    GLfloat global_depth_offset_constant;  
    GLfloat global_depth_offset_scale;   
    
-   /* for IGDNG only */
+   /* for Ironlake only */
    struct {
       GLuint pad0:1;
       GLuint grf_reg_count_1:3; 
@@ -1448,7 +1448,7 @@ struct brw_instruction
            GLuint end_of_thread:1;
            GLuint pad1:1;
            GLuint sfid:4;
-       } send_igdng;  /* for IGDNG only */
+       } send_gen5;  /* for Ironlake only */
 
    } bits2;
 
@@ -1549,7 +1549,7 @@ struct brw_instruction
 	 GLuint msg_length:4;
 	 GLuint pad1:2;
 	 GLuint end_of_thread:1;
-      } math_igdng;
+      } math_gen5;
 
       struct {
 	 GLuint binding_table_index:8;
@@ -1585,7 +1585,7 @@ struct brw_instruction
 	 GLuint msg_length:4;
 	 GLuint pad1:2;
 	 GLuint end_of_thread:1;
-      } sampler_igdng;
+      } sampler_gen5;
 
       struct brw_urb_immediate urb;
 
@@ -1603,7 +1603,7 @@ struct brw_instruction
 	 GLuint msg_length:4;
 	 GLuint pad1:2;
 	 GLuint end_of_thread:1;
-      } urb_igdng;
+      } urb_gen5;
 
       struct {
 	 GLuint binding_table_index:8;
@@ -1628,7 +1628,7 @@ struct brw_instruction
 	 GLuint msg_length:4;
 	 GLuint pad1:2;
 	 GLuint end_of_thread:1;
-      } dp_read_igdng;
+      } dp_read_gen5;
 
       struct {
 	 GLuint binding_table_index:8;
@@ -1655,7 +1655,7 @@ struct brw_instruction
 	 GLuint msg_length:4;
 	 GLuint pad1:2;
 	 GLuint end_of_thread:1;
-      } dp_write_igdng;
+      } dp_write_gen5;
 
       struct {
 	 GLuint pad:16;
@@ -1673,7 +1673,7 @@ struct brw_instruction
 	 GLuint msg_length:4;
 	 GLuint pad1:2;
 	 GLuint end_of_thread:1;
-      } generic_igdng;
+      } generic_gen5;
 
       GLint d;
       GLuint ud;

src/mesa/drivers/dri/i965/brw_tex_layout.c

@@ -48,77 +48,31 @@ GLboolean brw_miptree_layout(struct intel_context *intel,
 
    switch (mt->target) {
    case GL_TEXTURE_CUBE_MAP:
-      if (intel->is_ironlake) {
-          GLuint align_h = 2, align_w = 4;
+      if (intel->gen == 5) {
+          GLuint align_h = 2;
           GLuint level;
-          GLuint x = 0;
-          GLuint y = 0;
-          GLuint width = mt->width0;
-          GLuint height = mt->height0;
           GLuint qpitch = 0;
-          GLuint y_pitch = 0;
+	  int h0, h1, q;
 
-          mt->pitch = mt->width0;
-          intel_get_texture_alignment_unit(mt->internal_format, &align_w, &align_h);
-          y_pitch = ALIGN(height, align_h);
+	  /* On Ironlake, cube maps are finally represented as just a series
+	   * of MIPLAYOUT_BELOW 2D textures (like 2D texture arrays), separated
+	   * by a pitch of qpitch rows, where qpitch is defined by the equation
+	   * given in Volume 1 of the BSpec.
+	   */
+	  h0 = ALIGN(mt->height0, align_h);
+	  h1 = ALIGN(minify(h0), align_h);
+	  qpitch = (h0 + h1 + 11 * align_h);
+          if (mt->compressed)
+	     qpitch /= 4;
 
-          if (mt->compressed) {
-              mt->pitch = ALIGN(mt->width0, align_w);
-          }
-
-          if (mt->first_level != mt->last_level) {
-              GLuint mip1_width;
-
-              if (mt->compressed) {
-                  mip1_width = ALIGN(minify(mt->width0), align_w)
-                      + ALIGN(minify(minify(mt->width0)), align_w);
-              } else {
-                  mip1_width = ALIGN(minify(mt->width0), align_w)
-                      + minify(minify(mt->width0));
-              }
-
-              if (mip1_width > mt->pitch) {
-                  mt->pitch = mip1_width;
-              }
-          }
-
-          mt->pitch = intel_miptree_pitch_align(intel, mt, tiling, mt->pitch);
-
-          if (mt->compressed) {
-              qpitch = (y_pitch + ALIGN(minify(y_pitch), align_h) + 11 * align_h) / 4;
-              mt->total_height = (y_pitch + ALIGN(minify(y_pitch), align_h) + 11 * align_h) / 4 * 6;
-          } else {
-              qpitch = (y_pitch + ALIGN(minify(y_pitch), align_h) + 11 * align_h);
-              mt->total_height = (y_pitch + ALIGN(minify(y_pitch), align_h) + 11 * align_h) * 6;
-          }
+	  i945_miptree_layout_2d(intel, mt, tiling, 6);
 
           for (level = mt->first_level; level <= mt->last_level; level++) {
-              GLuint img_height;
-              GLuint nr_images = 6;
-              GLuint q = 0;
-
-              intel_miptree_set_level_info(mt, level, nr_images, x, y, width, 
-                                           height, 1);
-
-              for (q = 0; q < nr_images; q++)
-                  intel_miptree_set_image_offset(mt, level, q,
-						 x, y + q * qpitch);
-
-              if (mt->compressed)
-                  img_height = MAX2(1, height/4);
-              else
-                  img_height = ALIGN(height, align_h);
-
-              if (level == mt->first_level + 1) {
-                  x += ALIGN(width, align_w);
-              }
-              else {
-                  y += img_height;
-              }
-
-              width  = minify(width);
-              height = minify(height);
+	     for (q = 0; q < 6; q++) {
+		intel_miptree_set_image_offset(mt, level, q, 0, q * qpitch);
+	     }
           }
+	  mt->total_height = qpitch * 6;
 
           break;
       }
@@ -208,7 +162,7 @@ GLboolean brw_miptree_layout(struct intel_context *intel,
    }
 
    default:
-      i945_miptree_layout_2d(intel, mt, tiling);
+      i945_miptree_layout_2d(intel, mt, tiling, 1);
       break;
    }
    DBG("%s: %dx%dx%d - sz 0x%x\n", __FUNCTION__,

src/mesa/drivers/dri/i965/brw_urb.c

@@ -148,7 +148,7 @@ static void recalculate_urb_fence( struct brw_context *brw )
 
       brw->urb.constrained = 0;
 
-      if (intel->is_ironlake) {
+      if (intel->gen == 5) {
          brw->urb.nr_vs_entries = 128;
          brw->urb.nr_sf_entries = 48;
          if (check_urb_layout(brw)) {

src/mesa/drivers/dri/i965/brw_vs_emit.c

@@ -182,7 +182,7 @@ static void brw_vs_alloc_regs( struct brw_vs_compile *c )
 
    if (intel->gen >= 6)
       mrf = 6;
-   else if (intel->is_ironlake)
+   else if (intel->gen == 5)
       mrf = 8;
    else
       mrf = 4;
@@ -283,7 +283,7 @@ static void brw_vs_alloc_regs( struct brw_vs_compile *c )
 
    if (intel->gen >= 6)
       c->prog_data.urb_entry_size = (attributes_in_vue + 4 + 7) / 8;
-   else if (intel->is_ironlake)
+   else if (intel->gen == 5)
       c->prog_data.urb_entry_size = (attributes_in_vue + 6 + 3) / 4;
    else
       c->prog_data.urb_entry_size = (attributes_in_vue + 2 + 3) / 4;
@@ -1288,7 +1288,7 @@ static void emit_vertex_write( struct brw_vs_compile *c)
       brw_MOV(p, offset(m0, 2), pos);
       brw_MOV(p, offset(m0, 5), pos);
       len_vertex_header = 4;
-   } else if (intel->is_ironlake) {
+   } else if (intel->gen == 5) {
       /* There are 20 DWs (D0-D19) in VUE header on Ironlake:
        * dword 0-3 (m1) of the header is indices, point width, clip flags.
        * dword 4-7 (m2) is the ndc position (set above)
@@ -1710,7 +1710,7 @@ void brw_vs_emit(struct brw_vs_compile *c )
 
             loop_depth--;
 
-	    if (intel->is_ironlake)
+	    if (intel->gen == 5)
 	       br = 2;
 
             inst0 = inst1 = brw_WHILE(p, loop_inst[loop_depth]);

src/mesa/drivers/dri/i965/brw_vs_state.c

@@ -98,7 +98,7 @@ vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
     */
    vs.thread1.single_program_flow = 0;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       vs.thread1.binding_table_entry_count = 0; /* hardware requirement */
    else
       vs.thread1.binding_table_entry_count = key->nr_surfaces;
@@ -109,7 +109,7 @@ vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
    vs.thread3.urb_entry_read_offset = 0;
    vs.thread3.const_urb_entry_read_offset = key->curbe_offset * 2;
 
-   if (intel->is_ironlake) {
+   if (intel->gen == 5) {
       switch (key->nr_urb_entries) {
       case 8:
       case 12:
@@ -150,7 +150,7 @@ vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
 
    /* No samplers for ARB_vp programs:
     */
-   /* It has to be set to 0 for IGDNG
+   /* It has to be set to 0 for Ironlake
     */
    vs.vs5.sampler_count = 0;
 

src/mesa/drivers/dri/i965/brw_wm.h

@@ -67,7 +67,7 @@ struct brw_wm_prog_key {
    GLuint flat_shade:1;
    GLuint linear_color:1;  /**< linear interpolation vs perspective interp */
    GLuint runtime_check_aads_emit:1;
-   GLuint nr_color_regions:2;
+   GLuint nr_color_regions:5;
    
    GLbitfield proj_attrib_mask; /**< one bit per fragment program attribute */
    GLuint shadowtex_mask:16;

src/mesa/drivers/dri/i965/brw_wm_emit.c

@@ -898,7 +898,7 @@ void emit_tex(struct brw_wm_compile *c,
    }
 
    /* Pre-Ironlake, the 8-wide sampler always took u,v,r. */
-   if (!intel->is_ironlake && c->dispatch_width == 8)
+   if (intel->gen < 5 && c->dispatch_width == 8)
       nr_texcoords = 3;
 
    /* For shadow comparisons, we have to supply u,v,r. */
@@ -916,7 +916,7 @@ void emit_tex(struct brw_wm_compile *c,
 
    /* Fill in the shadow comparison reference value. */
    if (shadow) {
-      if (intel->is_ironlake) {
+      if (intel->gen == 5) {
 	 /* Fill in the cube map array index value. */
 	 brw_MOV(p, brw_message_reg(cur_mrf), brw_imm_f(0));
 	 cur_mrf += mrf_per_channel;
@@ -929,11 +929,11 @@ void emit_tex(struct brw_wm_compile *c,
       cur_mrf += mrf_per_channel;
    }
 
-   if (intel->is_ironlake) {
+   if (intel->gen == 5) {
       if (shadow)
-	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_IGDNG;
+	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_GEN5;
       else
-	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_IGDNG;
+	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_GEN5;
    } else {
       /* Note that G45 and older determines shadow compare and dispatch width
        * from message length for most messages.
@@ -981,16 +981,16 @@ void emit_txb(struct brw_wm_compile *c,
     * undefined, and trust the execution mask to keep the undefined pixels
     * from mattering.
     */
-   if (c->dispatch_width == 16 || !intel->is_ironlake) {
-      if (intel->is_ironlake)
-	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_IGDNG;
+   if (c->dispatch_width == 16 || intel->gen < 5) {
+      if (intel->gen == 5)
+	 msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_GEN5;
       else
 	 msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS;
       mrf_per_channel = 2;
       dst_retyped = retype(vec16(dst[0]), BRW_REGISTER_TYPE_UW);
       response_length = 8;
    } else {
-      msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_IGDNG;
+      msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_GEN5;
       mrf_per_channel = 1;
       dst_retyped = retype(vec8(dst[0]), BRW_REGISTER_TYPE_UW);
       response_length = 4;

src/mesa/drivers/dri/i965/brw_wm_glsl.c

@@ -2006,7 +2006,7 @@ static void brw_wm_emit_glsl(struct brw_context *brw, struct brw_wm_compile *c)
                   struct brw_instruction *inst0, *inst1;
                   GLuint br = 1;
 
-                  if (intel->is_ironlake)
+                  if (intel->gen == 5)
                      br = 2;
 
 		  assert(loop_depth > 0);

src/mesa/drivers/dri/i965/brw_wm_state.c

@@ -137,7 +137,7 @@ wm_unit_create_from_key(struct brw_context *brw, struct brw_wm_unit_key *key,
    wm.thread1.depth_coef_urb_read_offset = 1;
    wm.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       wm.thread1.binding_table_entry_count = 0; /* hardware requirement */
    else
       wm.thread1.binding_table_entry_count = key->nr_surfaces;
@@ -157,7 +157,7 @@ wm_unit_create_from_key(struct brw_context *brw, struct brw_wm_unit_key *key,
    wm.thread3.const_urb_entry_read_length = key->curb_entry_read_length;
    wm.thread3.const_urb_entry_read_offset = key->curbe_offset * 2;
 
-   if (intel->is_ironlake)
+   if (intel->gen == 5)
       wm.wm4.sampler_count = 0; /* hardware requirement */
    else
       wm.wm4.sampler_count = (key->sampler_count + 1) / 4;

src/mesa/drivers/dri/i965/brw_wm_surface_state.c

@@ -588,7 +588,7 @@ brw_update_renderbuffer_surface(struct brw_context *brw,
 	    tile_base = ((key.draw_y / 32) * (32 * pitch));
 	    tile_base += (key.draw_x - tile_x) / (128 / key.cpp) * 4096;
 	 }
-	 assert(intel->is_g4x || (tile_x == 0 && tile_y == 0));
+	 assert(brw->has_surface_tile_offset || (tile_x == 0 && tile_y == 0));
 	 assert(tile_x % 4 == 0);
 	 assert(tile_y % 2 == 0);
 	 /* Note that the low bits of these fields are missing, so

src/mesa/drivers/dri/intel/intel_chipset.h

@@ -93,7 +93,7 @@
 
 #define IS_ILD(devid)           (devid == PCI_CHIP_ILD_G)
 #define IS_ILM(devid)           (devid == PCI_CHIP_ILM_G)
-#define IS_IGDNG(devid)           (IS_ILD(devid) || IS_ILM(devid))
+#define IS_GEN5(devid)          (IS_ILD(devid) || IS_ILM(devid))
 
 #define IS_915(devid)		(devid == PCI_CHIP_I915_G || \
 				 devid == PCI_CHIP_E7221_G || \
@@ -118,7 +118,7 @@
 
 #define IS_965(devid)		(IS_GEN4(devid) || \
 				 IS_G4X(devid) || \
-				 IS_IGDNG(devid) || \
+				 IS_GEN5(devid) || \
 				 IS_GEN6(devid))
 
 #define IS_9XX(devid)		(IS_915(devid) || \

src/mesa/drivers/dri/intel/intel_context.c

@@ -159,10 +159,10 @@ intelGetString(GLcontext * ctx, GLenum name)
          chipset = "Intel(R) B43";
          break;
       case PCI_CHIP_ILD_G:
-         chipset = "Intel(R) IGDNG_D";
+         chipset = "Intel(R) Ironlake Desktop";
          break;
       case PCI_CHIP_ILM_G:
-         chipset = "Intel(R) IGDNG_M";
+         chipset = "Intel(R) Ironlake Mobile";
          break;
       default:
          chipset = "Unknown Intel Chipset";
@@ -362,7 +362,7 @@ intel_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable)
        if (buffers[i].attachment == __DRI_BUFFER_DEPTH)
 	  depth_region = region;
 
-       intel_renderbuffer_set_region(rb, region);
+       intel_renderbuffer_set_region(intel, rb, region);
        intel_region_release(&region);
 
        if (buffers[i].attachment == __DRI_BUFFER_DEPTH_STENCIL) {
@@ -374,7 +374,7 @@ intel_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable)
 		   continue;
 
 	     intel_region_reference(&stencil_region, region);
-	     intel_renderbuffer_set_region(rb, stencil_region);
+	     intel_renderbuffer_set_region(intel, rb, stencil_region);
 	     intel_region_release(&stencil_region);
 	  }
        }
@@ -610,12 +610,21 @@ intelInitContext(struct intel_context *intel,
    intel->driContext = driContextPriv;
    intel->driFd = sPriv->fd;
 
+   intel->has_xrgb_textures = GL_TRUE;
    if (IS_GEN6(intel->intelScreen->deviceID)) {
       intel->gen = 6;
       intel->needs_ff_sync = GL_TRUE;
       intel->has_luminance_srgb = GL_TRUE;
+   } else if (IS_GEN5(intel->intelScreen->deviceID)) {
+      intel->gen = 5;
+      intel->needs_ff_sync = GL_TRUE;
+      intel->has_luminance_srgb = GL_TRUE;
    } else if (IS_965(intel->intelScreen->deviceID)) {
       intel->gen = 4;
+      if (IS_G4X(intel->intelScreen->deviceID)) {
+	  intel->has_luminance_srgb = GL_TRUE;
+	  intel->is_g4x = GL_TRUE;
+      }
    } else if (IS_9XX(intel->intelScreen->deviceID)) {
       intel->gen = 3;
       if (IS_945(intel->intelScreen->deviceID)) {
@@ -623,15 +632,10 @@ intelInitContext(struct intel_context *intel,
       }
    } else {
       intel->gen = 2;
-   }
-
-   if (IS_IGDNG(intel->intelScreen->deviceID)) {
-      intel->is_ironlake = GL_TRUE;
-      intel->needs_ff_sync = GL_TRUE;
-      intel->has_luminance_srgb = GL_TRUE;
-   } else if (IS_G4X(intel->intelScreen->deviceID)) {
-      intel->has_luminance_srgb = GL_TRUE;
-      intel->is_g4x = GL_TRUE;
+      if (intel->intelScreen->deviceID == PCI_CHIP_I830_M ||
+	  intel->intelScreen->deviceID == PCI_CHIP_845_G) {
+	 intel->has_xrgb_textures = GL_FALSE;
+      }
    }
 
    driParseConfigFiles(&intel->optionCache, &intelScreen->optionCache,

src/mesa/drivers/dri/intel/intel_context.h

@@ -142,10 +142,10 @@ struct intel_context
     */
    int gen;
    GLboolean needs_ff_sync;
-   GLboolean is_ironlake;
    GLboolean is_g4x;
    GLboolean is_945;
    GLboolean has_luminance_srgb;
+   GLboolean has_xrgb_textures;
 
    int urb_size;
 

src/mesa/drivers/dri/intel/intel_fbo.c

@@ -42,7 +42,9 @@
 #include "intel_fbo.h"
 #include "intel_mipmap_tree.h"
 #include "intel_regions.h"
-
+#ifndef I915
+#include "brw_state.h"
+#endif
 
 #define FILE_DEBUG_FLAG DEBUG_FBO
 
@@ -280,7 +282,8 @@ intel_nop_alloc_storage(GLcontext * ctx, struct gl_renderbuffer *rb,
 
 
 void
-intel_renderbuffer_set_region(struct intel_renderbuffer *rb,
+intel_renderbuffer_set_region(struct intel_context *intel,
+			      struct intel_renderbuffer *rb,
 			      struct intel_region *region)
 {
    struct intel_region *old;
@@ -288,6 +291,12 @@ intel_renderbuffer_set_region(struct intel_renderbuffer *rb,
    old = rb->region;
    rb->region = NULL;
    intel_region_reference(&rb->region, region);
+#ifndef I915
+   if (old) {
+      brw_state_cache_bo_delete(&brw_context(&intel->ctx)->surface_cache,
+				old->buffer);
+   }
+#endif
    intel_region_release(&old);
 }
 

src/mesa/drivers/dri/intel/intel_fbo.h

@@ -82,7 +82,8 @@ intel_get_renderbuffer(struct gl_framebuffer *fb, int attIndex)
 
 
 extern void
-intel_renderbuffer_set_region(struct intel_renderbuffer *irb,
+intel_renderbuffer_set_region(struct intel_context *intel,
+			      struct intel_renderbuffer *irb,
 			      struct intel_region *region);
 
 

src/mesa/drivers/dri/intel/intel_mipmap_tree.c

@@ -117,17 +117,16 @@ intel_miptree_create(struct intel_context *intel,
 		     GLboolean expect_accelerated_upload)
 {
    struct intel_mipmap_tree *mt;
-   uint32_t tiling;
+   uint32_t tiling = I915_TILING_NONE;
 
    if (intel->use_texture_tiling && compress_byte == 0) {
       if (intel->gen >= 4 &&
 	  (base_format == GL_DEPTH_COMPONENT ||
 	   base_format == GL_DEPTH_STENCIL_EXT))
 	 tiling = I915_TILING_Y;
-      else
+      else if (width0 >= 64)
 	 tiling = I915_TILING_X;
-   } else
-      tiling = I915_TILING_NONE;
+   }
 
    mt = intel_miptree_create_internal(intel, target, internal_format,
 				      first_level, last_level, width0,

src/mesa/drivers/dri/intel/intel_pixel_draw.c

@@ -167,7 +167,7 @@ intel_stencil_drawpixels(GLcontext * ctx,
    irb = intel_create_renderbuffer(MESA_FORMAT_ARGB8888);
    irb->Base.Width = depth_irb->Base.Width;
    irb->Base.Height = depth_irb->Base.Height;
-   intel_renderbuffer_set_region(irb, depth_irb->region);
+   intel_renderbuffer_set_region(intel, irb, depth_irb->region);
 
    /* Create a name for our renderbuffer, which lets us use other mesa
     * rb functions for convenience.

src/mesa/drivers/dri/intel/intel_tex_format.c

@@ -49,7 +49,14 @@ intelChooseTextureFormat(GLcontext * ctx, GLint internalFormat,
       if (format == GL_RGB && type == GL_UNSIGNED_SHORT_5_6_5) {
          return MESA_FORMAT_RGB565;
       }
-      return do32bpt ? MESA_FORMAT_XRGB8888 : MESA_FORMAT_RGB565;
+      if (do32bpt) {
+	 if (intel->has_xrgb_textures)
+	    return MESA_FORMAT_XRGB8888;
+	 else
+	    return MESA_FORMAT_ARGB8888;
+      } else {
+	 return MESA_FORMAT_RGB565;
+      }
 
    case GL_RGBA8:
    case GL_RGB10_A2:
@@ -68,7 +75,10 @@ intelChooseTextureFormat(GLcontext * ctx, GLint internalFormat,
    case GL_RGB10:
    case GL_RGB12:
    case GL_RGB16:
-      return MESA_FORMAT_XRGB8888;
+      if (intel->has_xrgb_textures)
+	 return MESA_FORMAT_XRGB8888;
+      else
+	 return MESA_FORMAT_ARGB8888;
 
    case GL_RGB5:
    case GL_RGB4:

src/mesa/drivers/dri/intel/intel_tex_layout.c

@@ -63,9 +63,9 @@ void intel_get_texture_alignment_unit(GLenum internalFormat, GLuint *w, GLuint *
     }
 }
 
-void i945_miptree_layout_2d( struct intel_context *intel,
-			     struct intel_mipmap_tree *mt,
-			     uint32_t tiling )
+void i945_miptree_layout_2d(struct intel_context *intel,
+			    struct intel_mipmap_tree *mt,
+			    uint32_t tiling, int nr_images)
 {
    GLuint align_h = 2, align_w = 4;
    GLuint level;
@@ -111,7 +111,7 @@ void i945_miptree_layout_2d( struct intel_context *intel,
    for ( level = mt->first_level ; level <= mt->last_level ; level++ ) {
       GLuint img_height;
 
-      intel_miptree_set_level_info(mt, level, 1, x, y, width, 
+      intel_miptree_set_level_info(mt, level, nr_images, x, y, width,
 				   height, 1);
 
       if (mt->compressed)

src/mesa/drivers/dri/intel/intel_tex_layout.h

@@ -40,5 +40,5 @@ static INLINE GLuint minify( GLuint d )
 
 extern void i945_miptree_layout_2d(struct intel_context *intel,
 				   struct intel_mipmap_tree *mt,
-				   uint32_t tiling);
+				   uint32_t tiling, int nr_images);
 extern void intel_get_texture_alignment_unit(GLenum, GLuint *, GLuint *);

src/mesa/drivers/dri/r300/compiler/radeon_pair_schedule.c

@@ -141,12 +141,28 @@ static void add_inst_to_list(struct schedule_instruction ** list, struct schedul
 	*list = inst;
 }
 
+static void add_inst_to_list_end(struct schedule_instruction ** list,
+					struct schedule_instruction * inst)
+{
+	if(!*list){
+		*list = inst;
+	}else{
+		struct schedule_instruction * temp = *list;
+		while(temp->NextReady){
+			temp = temp->NextReady;
+		}
+		temp->NextReady = inst;
+	}
+}
+
 static void instruction_ready(struct schedule_state * s, struct schedule_instruction * sinst)
 {
 	DBG("%i is now ready\n", sinst->Instruction->IP);
 
+	/* Adding Ready TEX instructions to the end of the "Ready List" helps
+	 * us emit TEX instructions in blocks without losing our place. */
 	if (sinst->Instruction->Type == RC_INSTRUCTION_NORMAL)
-		add_inst_to_list(&s->ReadyTEX, sinst);
+		add_inst_to_list_end(&s->ReadyTEX, sinst);
 	else if (sinst->Instruction->U.P.Alpha.Opcode == RC_OPCODE_NOP)
 		add_inst_to_list(&s->ReadyRGB, sinst);
 	else if (sinst->Instruction->U.P.RGB.Opcode == RC_OPCODE_NOP)
@@ -163,11 +179,14 @@ static void decrease_dependencies(struct schedule_state * s, struct schedule_ins
 		instruction_ready(s, sinst);
 }
 
-static void commit_instruction(struct schedule_state * s, struct schedule_instruction * sinst)
-{
-	DBG("%i: commit\n", sinst->Instruction->IP);
-
-	for(unsigned int i = 0; i < sinst->NumReadValues; ++i) {
+/**
+ * This function decreases the dependencies of the next instruction that
+ * wants to write to each of sinst's read values.
+ */
+static void commit_update_reads(struct schedule_state * s,
+					struct schedule_instruction * sinst){
+	unsigned int i;
+	for(i = 0; i < sinst->NumReadValues; ++i) {
 		struct reg_value * v = sinst->ReadValues[i];
 		assert(v->NumReaders > 0);
 		v->NumReaders--;
@@ -176,8 +195,12 @@ static void commit_instruction(struct schedule_state * s, struct schedule_instru
 				decrease_dependencies(s, v->Next->Writer);
 		}
 	}
+}
 
-	for(unsigned int i = 0; i < sinst->NumWriteValues; ++i) {
+static void commit_update_writes(struct schedule_state * s,
+					struct schedule_instruction * sinst){
+	unsigned int i;
+	for(i = 0; i < sinst->NumWriteValues; ++i) {
 		struct reg_value * v = sinst->WriteValues[i];
 		if (v->NumReaders) {
 			for(struct reg_value_reader * r = v->Readers; r; r = r->Next) {
@@ -196,6 +219,15 @@ static void commit_instruction(struct schedule_state * s, struct schedule_instru
 	}
 }
 
+static void commit_alu_instruction(struct schedule_state * s, struct schedule_instruction * sinst)
+{
+	DBG("%i: commit\n", sinst->Instruction->IP);
+
+	commit_update_reads(s, sinst);
+
+	commit_update_writes(s, sinst);
+}
+
 /**
  * Emit all ready texture instructions in a single block.
  *
@@ -208,21 +240,37 @@ static void emit_all_tex(struct schedule_state * s, struct rc_instruction * befo
 
 	assert(s->ReadyTEX);
 
-	/* Don't let the ready list change under us! */
-	readytex = s->ReadyTEX;
-	s->ReadyTEX = 0;
-
 	/* Node marker for R300 */
 	struct rc_instruction * inst_begin = rc_insert_new_instruction(s->C, before->Prev);
 	inst_begin->U.I.Opcode = RC_OPCODE_BEGIN_TEX;
 
 	/* Link texture instructions back in */
+	readytex = s->ReadyTEX;
 	while(readytex) {
-		struct schedule_instruction * tex = readytex;
-		readytex = readytex->NextReady;
+		rc_insert_instruction(before->Prev, readytex->Instruction);
+		DBG("%i: commit TEX reads\n", readytex->Instruction->IP);
 
-		rc_insert_instruction(before->Prev, tex->Instruction);
-		commit_instruction(s, tex);
+		/* All of the TEX instructions in the same TEX block have
+		 * their source registers read from before any of the
+		 * instructions in that block write to their destination
+		 * registers.  This means that when we commit a TEX
+		 * instruction, any other TEX instruction that wants to write
+		 * to one of the committed instruction's source register can be
+		 * marked as ready and should be emitted in the same TEX
+		 * block. This prevents the following sequence from being
+		 * emitted in two different TEX blocks:
+		 * 0: TEX temp[0].xyz, temp[1].xy__, 2D[0];
+		 * 1: TEX temp[1].xyz, temp[2].xy__, 2D[0];
+		 */
+		commit_update_reads(s, readytex);
+		readytex = readytex->NextReady;
+	}
+	readytex = s->ReadyTEX;
+	s->ReadyTEX = 0;
+	while(readytex){
+		DBG("%i: commit TEX writes\n", readytex->Instruction->IP);
+		commit_update_writes(s, readytex);
+		readytex = readytex->NextReady;
 	}
 }
 
@@ -328,7 +376,7 @@ static void emit_one_alu(struct schedule_state *s, struct rc_instruction * befor
 		}
 
 		rc_insert_instruction(before->Prev, sinst->Instruction);
-		commit_instruction(s, sinst);
+		commit_alu_instruction(s, sinst);
 	} else {
 		struct schedule_instruction **prgb;
 		struct schedule_instruction **palpha;
@@ -346,8 +394,8 @@ static void emit_one_alu(struct schedule_state *s, struct rc_instruction * befor
 				*prgb = (*prgb)->NextReady;
 				*palpha = (*palpha)->NextReady;
 				rc_insert_instruction(before->Prev, psirgb->Instruction);
-				commit_instruction(s, psirgb);
-				commit_instruction(s, psialpha);
+				commit_alu_instruction(s, psirgb);
+				commit_alu_instruction(s, psialpha);
 				goto success;
 			}
 		}
@@ -357,7 +405,7 @@ static void emit_one_alu(struct schedule_state *s, struct rc_instruction * befor
 		s->ReadyRGB = s->ReadyRGB->NextReady;
 
 		rc_insert_instruction(before->Prev, sinst->Instruction);
-		commit_instruction(s, sinst);
+		commit_alu_instruction(s, sinst);
 	success: ;
 	}
 }

src/mesa/drivers/dri/r300/r300_context.c

@@ -456,7 +456,7 @@ static void r300InitGLExtensions(GLcontext *ctx)
 	if (!r300->radeon.radeonScreen->drmSupportsOcclusionQueries) {
 		_mesa_disable_extension(ctx, "GL_ARB_occlusion_query");
 	}
-	if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV350)
+        if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_R420)
   		_mesa_enable_extension(ctx, "GL_ARB_half_float_vertex");
 
 	if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV515)

src/mesa/drivers/dri/r300/r300_draw.c

@@ -523,8 +523,7 @@ static void r300AllocDmaRegions(GLcontext *ctx, const struct gl_client_array *in
 			r300ConvertAttrib(ctx, count, input[i], &vbuf->attribs[index]);
 		} else {
 			if (input[i]->BufferObj->Name) {
-				if (stride % 4 != 0) {
-					assert(((intptr_t) input[i]->Ptr) % input[i]->StrideB == 0);
+				if (stride % 4 != 0 || (intptr_t)input[i]->Ptr % 4 != 0) {
 					r300AlignDataToDword(ctx, input[i], count, &vbuf->attribs[index]);
 					vbuf->attribs[index].is_named_bo = GL_FALSE;
 				} else {

src/mesa/drivers/dri/r600/r700_assembler.c

@@ -293,7 +293,9 @@ GLuint GetSurfaceFormat(GLenum eType, GLuint nChannels, GLuint * pClient_size)
                 case 2:
                     format = FMT_16_16; break;
                 case 3:
-                    format = FMT_16_16_16; break;
+                    /* 3 comp GL_SHORT vertex format doesnt work on r700
+                       4 somehow works, test - sauerbraten  */
+                    format = FMT_16_16_16_16; break;
                 case 4:
                     format = FMT_16_16_16_16; break;
                 default:
@@ -1051,6 +1053,67 @@ void checkop_init(r700_AssemblerBase* pAsm)
     pAsm->aArgSubst[3] = -1;
 }
 
+static GLboolean next_ins(r700_AssemblerBase *pAsm)
+{
+    struct prog_instruction *pILInst = &(pAsm->pILInst[pAsm->uiCurInst]);
+
+    if (GL_TRUE == pAsm->is_tex)
+    {
+        if (pILInst->TexSrcTarget == TEXTURE_RECT_INDEX)
+        {
+            if (GL_FALSE == assemble_tex_instruction(pAsm, GL_FALSE))
+            {
+                radeon_error("Error assembling TEX instruction\n");
+                return GL_FALSE;
+            }
+        }
+        else
+        {
+            if (GL_FALSE == assemble_tex_instruction(pAsm, GL_TRUE))
+            {
+                radeon_error("Error assembling TEX instruction\n");
+                return GL_FALSE;
+            }
+        }
+    }
+    else
+    {   //ALU
+        if (GL_FALSE == assemble_alu_instruction(pAsm))
+        {
+            radeon_error("Error assembling ALU instruction\n");
+            return GL_FALSE;
+        }
+    }
+
+    if (pAsm->D.dst.rtype == DST_REG_OUT)
+    {
+        assert(pAsm->D.dst.reg >= pAsm->starting_export_register_number);
+
+        if (pAsm->D.dst.op3)
+        {
+            // There is no mask for OP3 instructions, so all channels are written
+            pAsm->pucOutMask[pAsm->D.dst.reg - pAsm->starting_export_register_number] = 0xF;
+        }
+        else
+        {
+            pAsm->pucOutMask[pAsm->D.dst.reg - pAsm->starting_export_register_number]
+               |= (unsigned char)pAsm->pILInst[pAsm->uiCurInst].DstReg.WriteMask;
+        }
+    }
+
+    //reset for next inst.
+    pAsm->D.bits    = 0;
+    pAsm->D2.bits   = 0;
+    pAsm->S[0].bits = 0;
+    pAsm->S[1].bits = 0;
+    pAsm->S[2].bits = 0;
+    pAsm->is_tex = GL_FALSE;
+    pAsm->need_tex_barrier = GL_FALSE;
+    pAsm->D2.bits = 0;
+    pAsm->C[0].bits = pAsm->C[1].bits = pAsm->C[2].bits = pAsm->C[3].bits = 0;
+    return GL_TRUE;
+}
+
 GLboolean mov_temp(r700_AssemblerBase* pAsm, int src)
 {
     GLuint tmp = gethelpr(pAsm);
@@ -1201,7 +1264,7 @@ GLboolean checkop3(r700_AssemblerBase* pAsm)
 	    {
             if( GL_FALSE == mov_temp(pAsm, 1) )
             {
-                return 1;
+                return GL_FALSE;
             }
         }
 
@@ -2578,62 +2641,6 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
     return GL_TRUE;
 }
 
-GLboolean next_ins(r700_AssemblerBase *pAsm)
-{
-    struct prog_instruction *pILInst = &(pAsm->pILInst[pAsm->uiCurInst]);
-
-    if( GL_TRUE == pAsm->is_tex )
-    {
-	    if (pILInst->TexSrcTarget == TEXTURE_RECT_INDEX) {
-		    if( GL_FALSE == assemble_tex_instruction(pAsm, GL_FALSE) ) 
-		    {
-			    radeon_error("Error assembling TEX instruction\n");
-			    return GL_FALSE;
-		    }
-	    } else {
-		    if( GL_FALSE == assemble_tex_instruction(pAsm, GL_TRUE) ) 
-		    {
-			    radeon_error("Error assembling TEX instruction\n");
-			    return GL_FALSE;
-		    }
-	    }
-    }
-    else 
-    {   //ALU      
-        if( GL_FALSE == assemble_alu_instruction(pAsm) ) 
-        {
-            radeon_error("Error assembling ALU instruction\n");
-            return GL_FALSE;
-        }
-    } 
-      
-    if(pAsm->D.dst.rtype == DST_REG_OUT) 
-    {
-        if(pAsm->D.dst.op3) 
-        {        
-            // There is no mask for OP3 instructions, so all channels are written        
-            pAsm->pucOutMask[pAsm->D.dst.reg - pAsm->starting_export_register_number] = 0xF;
-        }
-        else 
-        {
-            pAsm->pucOutMask[pAsm->D.dst.reg - pAsm->starting_export_register_number] 
-               |= (unsigned char)pAsm->pILInst[pAsm->uiCurInst].DstReg.WriteMask;
-        }
-    }
-    
-    //reset for next inst.
-    pAsm->D.bits    = 0;
-    pAsm->D2.bits   = 0;
-    pAsm->S[0].bits = 0;
-    pAsm->S[1].bits = 0;
-    pAsm->S[2].bits = 0;
-    pAsm->is_tex = GL_FALSE;
-    pAsm->need_tex_barrier = GL_FALSE;
-    pAsm->D2.bits = 0;
-    pAsm->C[0].bits = pAsm->C[1].bits = pAsm->C[2].bits = pAsm->C[3].bits = 0;
-    return GL_TRUE;
-}
-
 GLboolean assemble_math_function(r700_AssemblerBase* pAsm, BITS opcode)
 {
     BITS tmp;
@@ -2865,25 +2872,92 @@ GLboolean assemble_CMP(r700_AssemblerBase *pAsm)
 
 GLboolean assemble_TRIG(r700_AssemblerBase *pAsm, BITS opcode)
 {
+    /* 
+     * r600 - trunc to -PI..PI range
+     * r700 - normalize by dividing by 2PI
+     * see fdo bug 27901
+     */
+  
     int tmp;
     checkop1(pAsm);
 
     tmp = gethelpr(pAsm);
 
-    pAsm->D.dst.opcode = SQ_OP2_INST_MUL;
+    pAsm->D.dst.opcode = SQ_OP3_INST_MULADD;
+    pAsm->D.dst.op3    = 1;
+
     setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
     pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
     pAsm->D.dst.reg    = tmp;
-    pAsm->D.dst.writex = 1;
 
     assemble_src(pAsm, 0, -1);
 
     pAsm->S[1].src.rtype = SRC_REC_LITERAL;
     setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_X);
+    
+    pAsm->S[2].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[2].src), SQ_SEL_Y);
+
     pAsm->D2.dst2.literal_slots = 1;
     pAsm->C[0].f = 1/(3.1415926535 * 2);
-    pAsm->C[1].f = 0.0F;
-    next_ins(pAsm);
+    pAsm->C[1].f = 0.5f;
+    
+    if ( GL_FALSE == next_ins(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    pAsm->D.dst.opcode = SQ_OP2_INST_FRACT;
+
+    setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+    pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg    = tmp;
+    pAsm->D.dst.writex = 1;
+
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = tmp;
+    setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_X);
+
+    if(( GL_FALSE == next_ins(pAsm) ))
+    {
+        return GL_FALSE;
+    }
+    pAsm->D.dst.opcode = SQ_OP3_INST_MULADD;
+    pAsm->D.dst.op3    = 1;
+
+    setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+    pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg    = tmp;
+
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = tmp;
+    setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_X);
+
+    pAsm->S[1].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_X);
+
+    pAsm->S[2].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[2].src), SQ_SEL_Y);
+
+    pAsm->D2.dst2.literal_slots = 1;
+
+    if (pAsm->bR6xx)
+    {
+       pAsm->C[0].f = 3.1415926535897f * 2.0f;
+       pAsm->C[1].f = -3.1415926535897f;
+    }
+    else 
+    {
+       pAsm->C[0].f = 1.0f;
+       pAsm->C[1].f = -0.5f;
+    }
+
+    if(( GL_FALSE == next_ins(pAsm) ))
+    {
+        return GL_FALSE;
+    }
 
     pAsm->D.dst.opcode = opcode;
     pAsm->D.dst.math = 1;
@@ -4023,22 +4097,79 @@ GLboolean assemble_SCS(r700_AssemblerBase *pAsm)
     checkop1(pAsm);
 
     tmp = gethelpr(pAsm);
-    /* tmp.x = src /2*PI */
-    pAsm->D.dst.opcode = SQ_OP2_INST_MUL;
+
+    pAsm->D.dst.opcode = SQ_OP3_INST_MULADD;
+    pAsm->D.dst.op3    = 1;
+
     setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
     pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
     pAsm->D.dst.reg    = tmp;
-    pAsm->D.dst.writex = 1;
 
     assemble_src(pAsm, 0, -1);
 
     pAsm->S[1].src.rtype = SRC_REC_LITERAL;
     setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_X);
+
+    pAsm->S[2].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[2].src), SQ_SEL_Y);
+
     pAsm->D2.dst2.literal_slots = 1;
     pAsm->C[0].f = 1/(3.1415926535 * 2);
-    pAsm->C[1].f = 0.0F;
+    pAsm->C[1].f = 0.5F;
 
-    next_ins(pAsm);
+    if ( GL_FALSE == next_ins(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    pAsm->D.dst.opcode = SQ_OP2_INST_FRACT;
+
+    setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+    pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg    = tmp;
+    pAsm->D.dst.writex = 1;
+
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = tmp;
+    setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_X);
+
+    if(( GL_FALSE == next_ins(pAsm) ))
+    {
+        return GL_FALSE;
+    }
+    pAsm->D.dst.opcode = SQ_OP3_INST_MULADD;
+    pAsm->D.dst.op3    = 1;
+
+    setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+    pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg    = tmp;
+
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = tmp;
+    setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_X);
+
+    pAsm->S[1].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_X);
+
+    pAsm->S[2].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[2].src), SQ_SEL_Y);
+
+    pAsm->D2.dst2.literal_slots = 1;
+
+    if(pAsm->bR6xx) {
+       pAsm->C[0].f = 3.1415926535897f * 2.0f;
+       pAsm->C[1].f = -3.1415926535897f;
+    } else {
+       pAsm->C[0].f = 1.0f;
+       pAsm->C[1].f = -0.5f;
+    }
+
+    if(( GL_FALSE == next_ins(pAsm) ))
+    {
+        return GL_FALSE;
+    }
 
     // COS dst.x,    a.x
     pAsm->D.dst.opcode = SQ_OP2_INST_COS;
@@ -6154,7 +6285,7 @@ GLboolean callPreSub(r700_AssemblerBase* pAsm,
     }
     if(uNumValidSrc > 0)
     {
-        prelude_cf_ptr     = pAsm->cf_current_alu_clause_ptr;
+        prelude_cf_ptr     = (R700ControlFlowGenericClause*) pAsm->cf_current_alu_clause_ptr;
         pAsm->alu_x_opcode = SQ_CF_INST_ALU;
     }
 
@@ -6274,7 +6405,7 @@ GLboolean callPreSub(r700_AssemblerBase* pAsm,
 
         next_ins(pAsm);        
 
-        pAsm->callers[pAsm->unCallerArrayPointer - 1].finale_cf_ptr  = pAsm->cf_current_alu_clause_ptr;
+        pAsm->callers[pAsm->unCallerArrayPointer - 1].finale_cf_ptr  = (R700ControlFlowGenericClause*) pAsm->cf_current_alu_clause_ptr;
         pAsm->callers[pAsm->unCallerArrayPointer - 1].prelude_cf_ptr = prelude_cf_ptr;
         pAsm->alu_x_opcode = SQ_CF_INST_ALU;
     }
@@ -6347,6 +6478,8 @@ GLboolean Process_Export(r700_AssemblerBase* pAsm,
 
     if (export_count == 1) 
     {
+        assert(starting_register_number >= pAsm->starting_export_register_number);
+
         ucWriteMask = pAsm->pucOutMask[starting_register_number - pAsm->starting_export_register_number];
 	/* exports Z as a float into Red channel */
 	if (GL_TRUE == is_depth_export)
@@ -6437,6 +6570,7 @@ GLboolean Process_Fragment_Exports(r700_AssemblerBase *pR700AsmCode,
 { 
     unsigned int unBit;
     GLuint export_count = 0;
+    unsigned int i;
 
     if(pR700AsmCode->depth_export_register_number >= 0) 
     {
@@ -6446,39 +6580,24 @@ GLboolean Process_Fragment_Exports(r700_AssemblerBase *pR700AsmCode,
 		}
     }
 
-    unBit = 1 << FRAG_RESULT_COLOR;
-	if(OutputsWritten & unBit)
-	{
-		if( GL_FALSE == Process_Export(pR700AsmCode,
-                                       SQ_EXPORT_PIXEL, 
-                                       0, 
-                                       1, 
-                                       pR700AsmCode->uiFP_OutputMap[FRAG_RESULT_COLOR], 
-                                       GL_FALSE) ) 
+    for (i = 0; i < FRAG_RESULT_MAX; ++i)
+    {
+        unBit = 1 << i;
+
+        if (OutputsWritten & unBit)
         {
-            return GL_FALSE;
+            GLboolean is_depth = i == FRAG_RESULT_DEPTH ? GL_TRUE : GL_FALSE;
+            if (!Process_Export(pR700AsmCode, SQ_EXPORT_PIXEL, 0, 1, pR700AsmCode->uiFP_OutputMap[i], is_depth))
+                return GL_FALSE;
+            ++export_count;
         }
-        export_count++;
-	}
-	unBit = 1 << FRAG_RESULT_DEPTH;
-	if(OutputsWritten & unBit)
-	{
-        if( GL_FALSE == Process_Export(pR700AsmCode,
-                                       SQ_EXPORT_PIXEL, 
-                                       0, 
-                                       1, 
-                                       pR700AsmCode->uiFP_OutputMap[FRAG_RESULT_DEPTH], 
-                                       GL_TRUE)) 
-        {
-            return GL_FALSE;
-        }
-        export_count++;
-	}
+    }
+
     /* Need to export something, otherwise we'll hang
      * results are undefined anyway */
     if(export_count == 0)
     {
-        Process_Export(pR700AsmCode, SQ_EXPORT_PIXEL, 0, 1, 0, GL_FALSE);
+        Process_Export(pR700AsmCode, SQ_EXPORT_PIXEL, 0, 1, pR700AsmCode->starting_export_register_number, GL_FALSE);
     }
     
     if(pR700AsmCode->cf_last_export_ptr != NULL) 
@@ -6511,13 +6630,30 @@ GLboolean Process_Vertex_Exports(r700_AssemblerBase *pR700AsmCode,
         {
             return GL_FALSE;
         }
-
+        export_starting_index++;
         export_count--;
+        }
+
+    unBit = 1 << VERT_RESULT_PSIZ;
+    if(OutputsWritten & unBit)
+    {
+        if( GL_FALSE == Process_Export(pR700AsmCode,
+                                       SQ_EXPORT_POS,
+                                       export_starting_index,
+                                       1,
+                                       pR700AsmCode->ucVP_OutputMap[VERT_RESULT_PSIZ],
+                                       GL_FALSE) )
+        {
+            return GL_FALSE;
+        }
+        export_count--;
+    }
+
+    pR700AsmCode->cf_last_export_ptr->m_Word1.f.cf_inst = SQ_CF_INST_EXPORT_DONE;
 
-        pR700AsmCode->cf_last_export_ptr->m_Word1.f.cf_inst = SQ_CF_INST_EXPORT_DONE;
-	}
 
     pR700AsmCode->number_of_exports = export_count;
+    export_starting_index = 0;
 
 	unBit = 1 << VERT_RESULT_COL0;
 	if(OutputsWritten & unBit)

src/mesa/drivers/dri/r600/r700_assembler.h

@@ -582,7 +582,6 @@ GLboolean check_scalar(r700_AssemblerBase* pAsm,
 GLboolean check_vector(r700_AssemblerBase* pAsm,
                        R700ALUInstruction* alu_instruction_ptr);
 GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm);
-GLboolean next_ins(r700_AssemblerBase *pAsm);
 
 GLboolean pops(r700_AssemblerBase *pAsm, GLuint pops);
 GLboolean jumpToOffest(r700_AssemblerBase *pAsm, GLuint pops, GLint offset);

src/mesa/drivers/dri/r600/r700_fragprog.c

@@ -226,22 +226,23 @@ void Map_Fragment_Program(r700_AssemblerBase         *pAsm,
 	pAsm->number_of_exports = 0;
 	pAsm->number_of_colorandz_exports = 0; /* don't include stencil and mask out. */
 	pAsm->starting_export_register_number = pAsm->number_used_registers;
-	unBit = 1 << FRAG_RESULT_COLOR;
-	if(mesa_fp->Base.OutputsWritten & unBit)
-	{
-		pAsm->uiFP_OutputMap[FRAG_RESULT_COLOR] = pAsm->number_used_registers++;
-		pAsm->number_of_exports++;
-		pAsm->number_of_colorandz_exports++;
-	}
-	unBit = 1 << FRAG_RESULT_DEPTH;
-	if(mesa_fp->Base.OutputsWritten & unBit)
-	{
-        pAsm->depth_export_register_number = pAsm->number_used_registers;
-		pAsm->uiFP_OutputMap[FRAG_RESULT_DEPTH] = pAsm->number_used_registers++;
-		pAsm->number_of_exports++;
-		pAsm->number_of_colorandz_exports++;
-		pAsm->pR700Shader->depthIsExported = 1;
-	}
+
+    for (i = 0; i < FRAG_RESULT_MAX; ++i)
+    {
+        unBit = 1 << i;
+        if (mesa_fp->Base.OutputsWritten & unBit)
+        {
+            if (i == FRAG_RESULT_DEPTH)
+            {
+                pAsm->depth_export_register_number = pAsm->number_used_registers;
+                pAsm->pR700Shader->depthIsExported = 1;
+            }
+
+            pAsm->uiFP_OutputMap[i] = pAsm->number_used_registers++;
+            ++pAsm->number_of_exports;
+            ++pAsm->number_of_colorandz_exports;
+        }
+    }
 
     pAsm->pucOutMask = (unsigned char*) MALLOC(pAsm->number_of_exports);
     for(ui=0; ui<pAsm->number_of_exports; ui++)
@@ -560,27 +561,34 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
         CLEARbit(r700->SPI_PS_IN_CONTROL_1.u32All, FRONT_FACE_ENA_bit);
     }
 
-    /* see if we need any point_sprite replacements */
-    for (i = VERT_RESULT_TEX0; i<= VERT_RESULT_TEX7; i++)
+    /* see if we need any point_sprite replacements, also increase num_interp
+     * as there's no vp output for them */
+    if (ctx->Point.PointSprite)
     {
-        if(ctx->Point.CoordReplace[i - VERT_RESULT_TEX0] == GL_TRUE)
-            point_sprite = GL_TRUE;
+        for (i = FRAG_ATTRIB_TEX0; i<= FRAG_ATTRIB_TEX7; i++)
+        {
+            if (ctx->Point.CoordReplace[i - FRAG_ATTRIB_TEX0] == GL_TRUE)
+            {
+                ui++;
+                point_sprite = GL_TRUE;
+            }
+        }
     }
 
+    if( mesa_fp->Base.InputsRead & (1 << FRAG_ATTRIB_PNTC))
+        ui++;
+
     if ((mesa_fp->Base.InputsRead & (1 << FRAG_ATTRIB_PNTC)) || point_sprite)
     {
-        /* for FRAG_ATTRIB_PNTC we need to increase num_interp */
-        if(mesa_fp->Base.InputsRead & (1 << FRAG_ATTRIB_PNTC))
-        {
-            ui++;
-            SETfield(r700->SPI_PS_IN_CONTROL_0.u32All, ui, NUM_INTERP_shift, NUM_INTERP_mask);
-        }
+        SETfield(r700->SPI_PS_IN_CONTROL_0.u32All, ui, NUM_INTERP_shift, NUM_INTERP_mask);
         SETbit(r700->SPI_INTERP_CONTROL_0.u32All, PNT_SPRITE_ENA_bit);
         SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_S, PNT_SPRITE_OVRD_X_shift, PNT_SPRITE_OVRD_X_mask);
         SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_T, PNT_SPRITE_OVRD_Y_shift, PNT_SPRITE_OVRD_Y_mask);
         SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_0, PNT_SPRITE_OVRD_Z_shift, PNT_SPRITE_OVRD_Z_mask);
         SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_1, PNT_SPRITE_OVRD_W_shift, PNT_SPRITE_OVRD_W_mask);
-        if(ctx->Point.SpriteOrigin == GL_LOWER_LEFT)
+        /* Like e.g. viewport and winding, point sprite coordinates are
+         * inverted when rendering to FBO. */
+        if ((ctx->Point.SpriteOrigin == GL_LOWER_LEFT) == !ctx->DrawBuffer->Name)
             SETbit(r700->SPI_INTERP_CONTROL_0.u32All, PNT_SPRITE_TOP_1_bit);
         else
             CLEARbit(r700->SPI_INTERP_CONTROL_0.u32All, PNT_SPRITE_TOP_1_bit);
@@ -668,8 +676,9 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
 
     for(i=0; i<8; i++)
     {
+	    GLboolean coord_replace = ctx->Point.PointSprite && ctx->Point.CoordReplace[i];
 	    unBit = 1 << (VERT_RESULT_TEX0 + i);
-	    if(OutputsWritten & unBit)
+	    if ((OutputsWritten & unBit) || coord_replace)
 	    {
 		    ui = pAsm->uiFP_AttributeMap[FRAG_ATTRIB_TEX0 + i];
 		    SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
@@ -677,7 +686,7 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
 			     SEMANTIC_shift, SEMANTIC_mask);
 		    CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
 		    /* ARB_point_sprite */
-		    if(ctx->Point.CoordReplace[i] == GL_TRUE)
+		    if (coord_replace)
 		    {
 			     SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, PT_SPRITE_TEX_bit);
 		    }

src/mesa/drivers/dri/r600/r700_vertprog.c

@@ -628,6 +628,16 @@ GLboolean r700SetupVertexProgram(GLcontext * ctx)
 
     R600_STATECHANGE(context, spi);
 
+    if(vp->mesa_program->Base.OutputsWritten & (1 << VERT_RESULT_PSIZ)) {
+        R600_STATECHANGE(context, cl);
+        SETbit(r700->PA_CL_VS_OUT_CNTL.u32All, USE_VTX_POINT_SIZE_bit);
+        SETbit(r700->PA_CL_VS_OUT_CNTL.u32All, VS_OUT_MISC_VEC_ENA_bit);
+    } else if (r700->PA_CL_VS_OUT_CNTL.u32All != 0) {
+        R600_STATECHANGE(context, cl);
+        CLEARbit(r700->PA_CL_VS_OUT_CNTL.u32All, USE_VTX_POINT_SIZE_bit);
+        CLEARbit(r700->PA_CL_VS_OUT_CNTL.u32All, VS_OUT_MISC_VEC_ENA_bit);
+    }
+
     SETfield(r700->SPI_VS_OUT_CONFIG.u32All,
 	     vp->r700Shader.nParamExports ? (vp->r700Shader.nParamExports - 1) : 0,
              VS_EXPORT_COUNT_shift, VS_EXPORT_COUNT_mask);

src/mesa/drivers/dri/radeon/radeon_chipset.h

@@ -413,9 +413,9 @@ enum {
    CHIP_FAMILY_R350,
    CHIP_FAMILY_RV350,
    CHIP_FAMILY_RV380,
+   CHIP_FAMILY_RS400,
    CHIP_FAMILY_R420,
    CHIP_FAMILY_RV410,
-   CHIP_FAMILY_RS400,
    CHIP_FAMILY_RS600,
    CHIP_FAMILY_RS690,
    CHIP_FAMILY_RS740,

src/mesa/drivers/dri/radeon/radeon_common_context.c

@@ -300,10 +300,10 @@ void radeonDestroyContext(__DRIcontext *driContextPriv )
 	_mesa_meta_free(radeon->glCtx);
 
 	if (radeon == current) {
-		radeon_firevertices(radeon);
 		_mesa_make_current(NULL, NULL, NULL);
 	}
 
+	radeon_firevertices(radeon);
 	if (!is_empty_list(&radeon->dma.reserved)) {
 		rcommonFlushCmdBuf( radeon, __FUNCTION__ );
 	}

src/mesa/drivers/dri/radeon/radeon_mipmap_tree.c

@@ -588,17 +588,17 @@ int radeon_validate_texture_miptree(GLcontext * ctx, struct gl_texture_object *t
 			__FUNCTION__, texObj ,t->minLod, t->maxLod);
 
 	radeon_mipmap_tree *dst_miptree;
-	dst_miptree = get_biggest_matching_miptree(t, t->minLod, t->maxLod);
+	dst_miptree = get_biggest_matching_miptree(t, t->base.BaseLevel, t->base.MaxLevel);
 
+	radeon_miptree_unreference(&t->mt);
 	if (!dst_miptree) {
-		radeon_miptree_unreference(&t->mt);
 		radeon_try_alloc_miptree(rmesa, t);
-		dst_miptree = t->mt;
 		radeon_print(RADEON_TEXTURE, RADEON_NORMAL,
 			"%s: No matching miptree found, allocated new one %p\n",
 			__FUNCTION__, t->mt);
 
 	} else {
+		radeon_miptree_reference(dst_miptree, &t->mt);
 		radeon_print(RADEON_TEXTURE, RADEON_NORMAL,
 			"%s: Using miptree %p\n", __FUNCTION__, t->mt);
 	}
@@ -615,7 +615,7 @@ int radeon_validate_texture_miptree(GLcontext * ctx, struct gl_texture_object *t
 				"Checking image level %d, face %d, mt %p ... ",
 				level, face, img->mt);
 			
-			if (img->mt != dst_miptree) {
+			if (img->mt != t->mt) {
 				radeon_print(RADEON_TEXTURE, RADEON_TRACE,
 					"MIGRATING\n");
 
@@ -623,7 +623,7 @@ int radeon_validate_texture_miptree(GLcontext * ctx, struct gl_texture_object *t
 				if (src_bo && radeon_bo_is_referenced_by_cs(src_bo, rmesa->cmdbuf.cs)) {
 					radeon_firevertices(rmesa);
 				}
-				migrate_image_to_miptree(dst_miptree, img, face, level);
+				migrate_image_to_miptree(t->mt, img, face, level);
 			} else
 				radeon_print(RADEON_TEXTURE, RADEON_TRACE, "OK\n");
 		}

src/mesa/drivers/dri/radeon/radeon_tex_copy.c

@@ -52,22 +52,34 @@ do_copy_texsubimage(GLcontext *ctx,
     gl_format dst_mesaformat;
     unsigned src_width;
     unsigned dst_width;
+    unsigned flip_y;
 
     if (!radeon->vtbl.blit) {
         return GL_FALSE;
     }
 
     if (_mesa_get_format_bits(timg->base.TexFormat, GL_DEPTH_BITS) > 0) {
-        rrb = radeon_get_depthbuffer(radeon);
+        if (ctx->ReadBuffer->_DepthBuffer && ctx->ReadBuffer->_DepthBuffer->Wrapped) {
+            rrb = radeon_renderbuffer(ctx->ReadBuffer->_DepthBuffer->Wrapped);
+        } else {
+            rrb = radeon_renderbuffer(ctx->ReadBuffer->_DepthBuffer);
+        }
+        flip_y = ctx->ReadBuffer->Attachment[BUFFER_DEPTH].Type == GL_NONE;
     } else {
-        rrb = radeon_get_colorbuffer(radeon);
+        rrb = radeon_renderbuffer(ctx->ReadBuffer->_ColorReadBuffer);
+        flip_y = ctx->ReadBuffer->Attachment[BUFFER_COLOR0].Type == GL_NONE;
+    }
+
+    // This is software renderbuffer, fallback to swrast
+    if (!rrb) {
+        return GL_FALSE;
     }
 
     if (!timg->mt) {
         radeon_validate_texture_miptree(ctx, &tobj->base);
     }
 
-    assert(rrb && rrb->bo);
+    assert(rrb->bo);
     assert(timg->mt);
     assert(timg->mt->bo);
     assert(timg->base.Width >= dstx + width);
@@ -124,7 +136,7 @@ do_copy_texsubimage(GLcontext *ctx,
                              timg->mt->bo, dst_offset, dst_mesaformat,
                              timg->mt->levels[level].rowstride / dst_bpp,
                              dst_width, timg->base.Height,
-                             dstx, dsty, width, height, 1);
+                             dstx, dsty, width, height, flip_y);
 }
 
 void

src/mesa/drivers/dri/radeon/radeon_texture.c

@@ -1016,7 +1016,15 @@ radeon_get_tex_image(GLcontext * ctx, GLenum target, GLint level,
 			__func__, ctx, texObj, image, compressed);
 
 	if (image->mt) {
+		radeonContextPtr rmesa = RADEON_CONTEXT(ctx);
 		/* Map the texture image read-only */
+		if (radeon_bo_is_referenced_by_cs(image->mt->bo, rmesa->cmdbuf.cs)) {
+				radeon_print(RADEON_TEXTURE, RADEON_VERBOSE,
+				    "%s: called for texture that is queued for GPU processing\n",
+				    __func__);
+				radeon_firevertices(rmesa);
+		}
+
 		radeon_teximage_map(image, GL_FALSE);
 	} else {
 		/* Image hasn't been uploaded to a miptree yet */

src/mesa/drivers/osmesa/Makefile

@@ -19,17 +19,11 @@ INCLUDE_DIRS = \
 	-I$(TOP)/src/mesa \
 	-I$(TOP)/src/mesa/main
 
-# Standalone osmesa needs to be linked with core Mesa APIs
-ifeq ($(DRIVER_DIRS), osmesa)
 CORE_MESA = \
 	$(TOP)/src/mesa/libmesa.a \
 	$(TOP)/src/mesa/libglapi.a \
 	$(TOP)/src/glsl/cl/libglslcl.a \
 	$(TOP)/src/glsl/pp/libglslpp.a
-else
-CORE_MESA =
-endif
-
 
 .c.o:
 	$(CC) -c $(INCLUDE_DIRS) $(CFLAGS) $< -o $@

src/mesa/drivers/osmesa/osmesa.c

@@ -1004,7 +1004,20 @@ new_osmesa_renderbuffer(GLcontext *ctx, GLenum format, GLenum type)
       rb->AllocStorage = osmesa_renderbuffer_storage;
 
       rb->InternalFormat = GL_RGBA;
-      rb->Format = MESA_FORMAT_RGBA8888;
+      switch (type) {
+      case GL_UNSIGNED_BYTE:
+         rb->Format = MESA_FORMAT_RGBA8888;
+         break;
+      case GL_UNSIGNED_SHORT:
+         rb->Format = MESA_FORMAT_SIGNED_RGBA_16;
+         break;
+      case GL_FLOAT:
+         rb->Format = MESA_FORMAT_RGBA_FLOAT32;
+         break;
+      default:
+         assert(0 && "Unexpected type in new_osmesa_renderbuffer()");
+         rb->Format = MESA_FORMAT_RGBA8888;
+      }
       rb->_BaseFormat = GL_RGBA;
       rb->DataType = type;
    }
@@ -1048,7 +1061,6 @@ OSMesaCreateContextExt( GLenum format, GLint depthBits, GLint stencilBits,
    struct dd_function_table functions;
    GLint rind, gind, bind, aind;
    GLint redBits = 0, greenBits = 0, blueBits = 0, alphaBits =0;
-   GLenum type = CHAN_TYPE;
 
    rind = gind = bind = aind = 0;
    if (format==OSMESA_RGBA) {
@@ -1167,11 +1179,9 @@ OSMesaCreateContextExt( GLenum format, GLint depthBits, GLint stencilBits,
          return NULL;
       }
 
-      /* create front color buffer in user-provided memory (no back buffer) */
-      osmesa->rb = new_osmesa_renderbuffer(&osmesa->mesa, format, type);
-      _mesa_add_renderbuffer(osmesa->gl_buffer, BUFFER_FRONT_LEFT, osmesa->rb);
-      assert(osmesa->rb->RefCount == 2);
-                        
+      /* Create depth/stencil/accum buffers.  We'll create the color
+       * buffer later in OSMesaMakeCurrent().
+       */
       _mesa_add_soft_renderbuffers(osmesa->gl_buffer,
                                    GL_FALSE, /* color */
                                    osmesa->gl_visual->haveDepthBuffer,
@@ -1308,11 +1318,24 @@ OSMesaMakeCurrent( OSMesaContext osmesa, void *buffer, GLenum type,
     */
    _glapi_check_multithread();
 
+
+   /* Create a front/left color buffer which wraps the user-provided buffer.
+    * There is no back color buffer.
+    * If the user tries to use a 8, 16 or 32-bit/channel buffer that
+    * doesn't match what Mesa was compiled for (CHAN_BITS) the
+    * _mesa_add_renderbuffer() function will create a "wrapper" renderbuffer
+    * that converts rendering from CHAN_BITS to the user-requested channel
+    * size.
+    */
+   osmesa->rb = new_osmesa_renderbuffer(&osmesa->mesa, osmesa->format, type);
+   _mesa_remove_renderbuffer(osmesa->gl_buffer, BUFFER_FRONT_LEFT);
+   _mesa_add_renderbuffer(osmesa->gl_buffer, BUFFER_FRONT_LEFT, osmesa->rb);
+   assert(osmesa->rb->RefCount == 2);
+
    /* Set renderbuffer fields.  Set width/height = 0 to force 
     * osmesa_renderbuffer_storage() being called by _mesa_resize_framebuffer()
     */
    osmesa->rb->Data = buffer;
-   osmesa->rb->DataType = type;
    osmesa->rb->Width = osmesa->rb->Height = 0;
 
    /* Set the framebuffer's size.  This causes the

src/mesa/main/bufferobj.c

@@ -1880,7 +1880,7 @@ _mesa_BufferObjectUnpurgeable(GLcontext *ctx, GLuint name, GLenum option)
 
    bufObj->Purgeable = GL_FALSE;
 
-   retval = GL_RETAINED_APPLE;
+   retval = option;
    if (ctx->Driver.BufferObjectUnpurgeable)
       retval = ctx->Driver.BufferObjectUnpurgeable(ctx, bufObj, option);
 
@@ -1910,11 +1910,11 @@ _mesa_RenderObjectUnpurgeable(GLcontext *ctx, GLuint name, GLenum option)
 
    bufObj->Purgeable = GL_FALSE;
 
-   retval = GL_RETAINED_APPLE;
+   retval = option;
    if (ctx->Driver.RenderObjectUnpurgeable)
       retval = ctx->Driver.RenderObjectUnpurgeable(ctx, bufObj, option);
 
-   return option;
+   return retval;
 }
 
 
@@ -1940,7 +1940,7 @@ _mesa_TextureObjectUnpurgeable(GLcontext *ctx, GLuint name, GLenum option)
 
    bufObj->Purgeable = GL_FALSE;
 
-   retval = GL_RETAINED_APPLE;
+   retval = option;
    if (ctx->Driver.TextureObjectUnpurgeable)
       retval = ctx->Driver.TextureObjectUnpurgeable(ctx, bufObj, option);
 

src/mesa/main/context.c

@@ -1358,6 +1358,8 @@ _mesa_make_current( GLcontext *newCtx, GLframebuffer *drawBuffer,
       if (newCtx->FirstTimeCurrent) {
          _mesa_compute_version(newCtx);
 
+         newCtx->Extensions.String = _mesa_make_extension_string(newCtx);
+
          check_context_limits(newCtx);
 
          /* We can use this to help debug user's problems.  Tell them to set

src/mesa/main/fbobject.c

@@ -1766,10 +1766,10 @@ _mesa_FramebufferRenderbufferEXT(GLenum target, GLenum attachment,
       rb = NULL;
    }
 
-   if (attachment == GL_DEPTH_STENCIL_ATTACHMENT) {
+   if (attachment == GL_DEPTH_STENCIL_ATTACHMENT &&
+       rb && rb->Format != MESA_FORMAT_NONE) {
       /* make sure the renderbuffer is a depth/stencil format */
-      const GLenum baseFormat =
-         _mesa_get_format_base_format(att->Renderbuffer->Format);
+      const GLenum baseFormat = _mesa_get_format_base_format(rb->Format);
       if (baseFormat != GL_DEPTH_STENCIL) {
          _mesa_error(ctx, GL_INVALID_OPERATION,
                      "glFramebufferRenderbufferEXT(renderbuffer"

src/mesa/main/getstring.c

@@ -71,8 +71,6 @@ _mesa_GetString( GLenum name )
       case GL_VERSION:
          return (const GLubyte *) ctx->VersionString;
       case GL_EXTENSIONS:
-         if (!ctx->Extensions.String)
-            ctx->Extensions.String = _mesa_make_extension_string(ctx);
          return (const GLubyte *) ctx->Extensions.String;
 #if FEATURE_ARB_shading_language_100
       case GL_SHADING_LANGUAGE_VERSION_ARB:

src/mesa/main/state.c

@@ -537,7 +537,7 @@ _mesa_update_state_locked( GLcontext *ctx )
 
    /* Determine which state flags effect vertex/fragment program state */
    if (ctx->FragmentProgram._MaintainTexEnvProgram) {
-      prog_flags |= (_NEW_TEXTURE | _NEW_FOG |
+      prog_flags |= (_NEW_BUFFERS | _NEW_TEXTURE | _NEW_FOG |
 		     _NEW_ARRAY | _NEW_LIGHT | _NEW_POINT | _NEW_RENDERMODE |
 		     _NEW_PROGRAM);
    }

src/mesa/main/texenvprogram.c

@@ -98,6 +98,7 @@ struct state_key {
    GLuint fog_enabled:1;
    GLuint fog_mode:2;          /**< FOG_x */
    GLuint inputs_available:12;
+   GLuint num_draw_buffers:4;
 
    /* NOTE: This array of structs must be last! (see "keySize" below) */
    struct {
@@ -485,6 +486,9 @@ static GLuint make_state_key( GLcontext *ctx,  struct state_key *key )
       inputs_referenced |= FRAG_BIT_FOGC; /* maybe */
    }
 
+   /* _NEW_BUFFERS */
+   key->num_draw_buffers = ctx->DrawBuffer->_NumColorDrawBuffers;
+
    key->inputs_available = (inputs_available & inputs_referenced);
 
    /* compute size of state key, ignoring unused texture units */
@@ -1199,11 +1203,14 @@ emit_texenv(struct texenv_fragment_program *p, GLuint unit)
    else
       alpha_saturate = GL_FALSE;
 
-   /* If this is the very last calculation, emit direct to output reg:
+   /* If this is the very last calculation (and various other conditions
+    * are met), emit directly to the color output register.  Otherwise,
+    * emit to a temporary register.
     */
    if (key->separate_specular ||
        unit != p->last_tex_stage ||
        alpha_shift ||
+       key->num_draw_buffers != 1 ||
        rgb_shift)
       dest = get_temp( p );
    else
@@ -1438,10 +1445,10 @@ create_new_program(GLcontext *ctx, struct state_key *key,
    p.program->Base.Parameters = _mesa_new_parameter_list();
    p.program->Base.InputsRead = 0x0;
 
-   if (ctx->DrawBuffer->_NumColorDrawBuffers == 1)
+   if (key->num_draw_buffers == 1)
       p.program->Base.OutputsWritten = 1 << FRAG_RESULT_COLOR;
    else {
-      for (i = 0; i < ctx->DrawBuffer->_NumColorDrawBuffers; i++)
+      for (i = 0; i < key->num_draw_buffers; i++)
 	 p.program->Base.OutputsWritten |= (1 << (FRAG_RESULT_DATA0 + i));
    }
 
@@ -1493,8 +1500,8 @@ create_new_program(GLcontext *ctx, struct state_key *key,
 
    cf = get_source( &p, SRC_PREVIOUS, 0 );
 
-   for (i = 0; i < ctx->DrawBuffer->_NumColorDrawBuffers; i++) {
-      if (ctx->DrawBuffer->_NumColorDrawBuffers == 1)
+   for (i = 0; i < key->num_draw_buffers; i++) {
+      if (key->num_draw_buffers == 1)
 	 out = make_ureg( PROGRAM_OUTPUT, FRAG_RESULT_COLOR );
       else {
 	 out = make_ureg( PROGRAM_OUTPUT, FRAG_RESULT_DATA0 + i );

src/mesa/main/version.h

@@ -34,8 +34,8 @@
 /* Mesa version */
 #define MESA_MAJOR 7
 #define MESA_MINOR 8
-#define MESA_PATCH 1
-#define MESA_VERSION_STRING "7.8.1"
+#define MESA_PATCH 3
+#define MESA_VERSION_STRING "7.8.3"
 
 /* To make version comparison easy */
 #define MESA_VERSION(a,b,c) (((a) << 16) + ((b) << 8) + (c))

src/mesa/shader/shader_api.c

@@ -1259,6 +1259,54 @@ lookup_uniform_parameter(GLcontext *ctx, GLuint program, GLint location,
 }
 
 
+/**
+ * GLGL uniform arrays and structs require special handling.
+ *
+ * The GL_ARB_shader_objects spec says that if you use
+ * glGetUniformLocation to get the location of an array, you CANNOT
+ * access other elements of the array by adding an offset to the
+ * returned location.  For example, you must call
+ * glGetUniformLocation("foo[16]") if you want to set the 16th element
+ * of the array with glUniform().
+ *
+ * HOWEVER, some other OpenGL drivers allow accessing array elements
+ * by adding an offset to the returned array location.  And some apps
+ * seem to depend on that behaviour.
+ *
+ * Mesa's gl_uniform_list doesn't directly support this since each
+ * entry in the list describes one uniform variable, not one uniform
+ * element.  We could insert dummy entries in the list for each array
+ * element after [0] but that causes complications elsewhere.
+ *
+ * We solve this problem by encoding two values in the location that's
+ * returned by glGetUniformLocation():
+ *  a) index into gl_uniform_list::Uniforms[] for the uniform
+ *  b) an array/field offset (0 for simple types)
+ *
+ * These two values are encoded in the high and low halves of a GLint.
+ * By putting the uniform number in the high part and the offset in the
+ * low part, we can support the unofficial ability to index into arrays
+ * by adding offsets to the location value.
+ */
+static void
+merge_location_offset(GLint *location, GLint offset)
+{
+   *location = (*location << 16) | offset;
+}
+
+
+/**
+ * Seperate the uniform location and parameter offset.  See above.
+ */
+static void
+split_location_offset(GLint *location, GLint *offset)
+{
+   *offset = *location & 0xffff;
+   *location = *location >> 16;
+}
+
+
+
 /**
  * Called via ctx->Driver.GetUniformfv().
  */
@@ -1268,6 +1316,9 @@ _mesa_get_uniformfv(GLcontext *ctx, GLuint program, GLint location,
 {
    struct gl_program *prog;
    GLint paramPos;
+   GLint offset;
+
+   split_location_offset(&location, &offset);
 
    lookup_uniform_parameter(ctx, program, location, &prog, &paramPos);
 
@@ -1298,7 +1349,10 @@ _mesa_get_uniformiv(GLcontext *ctx, GLuint program, GLint location,
 {
    struct gl_program *prog;
    GLint paramPos;
+   GLint offset;
 
+   split_location_offset(&location, &offset);
+ 
    lookup_uniform_parameter(ctx, program, location, &prog, &paramPos);
 
    if (prog) {
@@ -1318,31 +1372,6 @@ _mesa_get_uniformiv(GLcontext *ctx, GLuint program, GLint location,
 }
 
 
-/**
- * The value returned by GetUniformLocation actually encodes two things:
- * 1. the index into the prog->Uniforms[] array for the uniform
- * 2. an offset in the prog->ParameterValues[] array for specifying array
- *    elements or structure fields.
- * This function merges those two values.
- */
-static void
-merge_location_offset(GLint *location, GLint offset)
-{
-   *location = *location | (offset << 16);
-}
-
-
-/**
- * Seperate the uniform location and parameter offset.  See above.
- */
-static void
-split_location_offset(GLint *location, GLint *offset)
-{
-   *offset = (*location >> 16);
-   *location = *location & 0xffff;
-}
-
-
 /**
  * Called via ctx->Driver.GetUniformLocation().
  *

src/mesa/shader/slang/slang_builtin.c

@@ -399,7 +399,7 @@ lookup_statevar(const char *var, GLint index1, GLint index2, const char *field,
    }
 
    if (isMatrix) {
-      /* load all four columns of matrix */
+      /* load all four rows (or columns) of matrix */
       GLint pos[4];
       GLuint j;
       for (j = 0; j < 4; j++) {
@@ -489,9 +489,26 @@ emit_statevars(const char *name, int array_len,
          tokens[0] = STATE_TEXGEN;
          tokens[2] = STATE_TEXGEN_OBJECT_Q;
       }
+      else if (strcmp(name, "gl_TextureMatrix") == 0) {
+         tokens[0] = STATE_TEXTURE_MATRIX;
+         tokens[4] = STATE_MATRIX_TRANSPOSE;
+      }
+      else if (strcmp(name, "gl_TextureMatrixInverse") == 0) {
+         tokens[0] = STATE_TEXTURE_MATRIX;
+         tokens[4] = STATE_MATRIX_INVTRANS;
+      }
+      else if (strcmp(name, "gl_TextureMatrixTranspose") == 0) {
+         tokens[0] = STATE_TEXTURE_MATRIX;
+         tokens[4] = 0;
+      }
+      else if (strcmp(name, "gl_TextureMatrixInverseTranspose") == 0) {
+         tokens[0] = STATE_TEXTURE_MATRIX;
+         tokens[4] = STATE_MATRIX_INVERSE;
+      }
       else {
          return -1; /* invalid array name */
       }
+      /* emit state vars for each array element */
       for (i = 0; i < array_len; i++) {
          GLint p;
          tokens[1] = i;
@@ -513,6 +530,19 @@ emit_statevars(const char *name, int array_len,
       }
       return pos;
    }
+   else if (type->type == SLANG_SPEC_MAT4) {
+      /* unroll/emit 4 array rows (or columns) */
+      slang_type_specifier vec4;
+      GLint i, p, pos = -1;
+      vec4.type = SLANG_SPEC_VEC4;
+      for (i = 0; i < 4; i++) {
+         tokens[2] = tokens[3] = i; /* row[i] (or column[i]) of matrix */
+         p = emit_statevars(NULL, 0, &vec4, tokens, paramList);
+         if (pos == -1)
+            pos = p;
+      }
+      return pos;
+   }
    else {
       GLint pos;
       assert(type->type == SLANG_SPEC_VEC4 ||

src/mesa/shader/slang/slang_codegen.c

@@ -4187,11 +4187,10 @@ _slang_gen_variable(slang_assemble_ctx * A, slang_operation *oper)
    slang_atom name = oper->var ? oper->var->a_name : oper->a_id;
    slang_variable *var = _slang_variable_locate(oper->locals, name, GL_TRUE);
    slang_ir_node *n;
-   if (!var) {
+   if (!var || !var->declared) {
       slang_info_log_error(A->log, "undefined variable '%s'", (char *) name);
       return NULL;
    }
-   assert(var->declared);
    n = new_var(A, var);
    return n;
 }

src/mesa/shader/slang/slang_compile_variable.h

@@ -41,7 +41,7 @@ typedef struct slang_variable_
    GLuint size;                     /**< Variable's size in bytes */
    GLboolean is_global;
    GLboolean isTemp;                /**< a named temporary (__resultTmp) */
-   GLboolean declared;              /**< for debug */
+   GLboolean declared;              /**< has the var been declared? */
    struct slang_ir_storage_ *store; /**< Storage for this var */
 } slang_variable;
 

src/mesa/shader/slang/slang_emit.c

@@ -2362,7 +2362,10 @@ emit(slang_emit_info *emitInfo, slang_ir_node *n)
 #if 0
       assert(!n->Store);
 #endif
-      n->Store = n->Children[1]->Store;
+      if (n->Children[1]->Store)
+         n->Store = n->Children[1]->Store;
+      else
+         n->Store = n->Children[0]->Store;
       return inst;
 
    case IR_SCOPE:
@@ -2370,6 +2373,7 @@ emit(slang_emit_info *emitInfo, slang_ir_node *n)
       _slang_push_var_table(emitInfo->vt);
       inst = emit(emitInfo, n->Children[0]);
       _slang_pop_var_table(emitInfo->vt);
+      n->Store = n->Children[0]->Store;
       return inst;
 
    case IR_VAR_DECL:

src/mesa/shader/slang/slang_link.c

@@ -921,7 +921,10 @@ _slang_link(GLcontext *ctx,
 
    if (!vertNotify || !fragNotify) {
       /* driver rejected one/both of the vertex/fragment programs */
-      link_error(shProg, "Vertex and/or fragment program rejected by driver\n");
+      if (!shProg->InfoLog) {
+	 link_error(shProg,
+		    "Vertex and/or fragment program rejected by driver\n");
+      }
    }
    else {
       shProg->LinkStatus = (shProg->VertexProgram || shProg->FragmentProgram);

src/mesa/state_tracker/st_cb_drawpixels.c

@@ -1135,7 +1135,7 @@ st_destroy_drawpix(struct st_context *st)
    st_reference_fragprog(st, &st->drawpix.z_shader, NULL);
    st_reference_fragprog(st, &st->pixel_xfer.combined_prog, NULL);
    if (st->drawpix.vert_shaders[0])
-      free(st->drawpix.vert_shaders[0]);
+      ureg_free_tokens(st->drawpix.vert_shaders[0]);
    if (st->drawpix.vert_shaders[1])
-      free(st->drawpix.vert_shaders[1]);
+      ureg_free_tokens(st->drawpix.vert_shaders[1]);
 }

src/mesa/state_tracker/st_cb_fbo.c

@@ -460,26 +460,38 @@ static void
 st_validate_framebuffer(GLcontext *ctx, struct gl_framebuffer *fb)
 {
    struct pipe_screen *screen = ctx->st->pipe->screen;
-   const struct gl_renderbuffer *depthRb =
-      fb->Attachment[BUFFER_DEPTH].Renderbuffer;
-   const struct gl_renderbuffer *stencilRb =
-      fb->Attachment[BUFFER_STENCIL].Renderbuffer;
+   const struct gl_renderbuffer_attachment *depth =
+         &fb->Attachment[BUFFER_DEPTH];
+   const struct gl_renderbuffer_attachment *stencil =
+         &fb->Attachment[BUFFER_STENCIL];
    GLuint i;
 
-   if (stencilRb && depthRb && stencilRb != depthRb) {
+   if (depth->Type && stencil->Type && depth->Type != stencil->Type) {
+      fb->_Status = GL_FRAMEBUFFER_UNSUPPORTED_EXT;
+      return;
+   }
+   if (depth->Type == GL_RENDERBUFFER_EXT &&
+       stencil->Type == GL_RENDERBUFFER_EXT &&
+       depth->Renderbuffer != stencil->Renderbuffer) {
+      fb->_Status = GL_FRAMEBUFFER_UNSUPPORTED_EXT;
+      return;
+   }
+   if (depth->Type == GL_TEXTURE &&
+       stencil->Type == GL_TEXTURE &&
+       depth->Texture != stencil->Texture) {
       fb->_Status = GL_FRAMEBUFFER_UNSUPPORTED_EXT;
       return;
    }
 
    if (!st_validate_attachment(screen,
-			       &fb->Attachment[BUFFER_DEPTH],
-			       PIPE_TEXTURE_USAGE_DEPTH_STENCIL)) {
+                               depth,
+                               PIPE_TEXTURE_USAGE_DEPTH_STENCIL)) {
       fb->_Status = GL_FRAMEBUFFER_UNSUPPORTED_EXT;
       return;
    }
    if (!st_validate_attachment(screen,
-			       &fb->Attachment[BUFFER_STENCIL],
-			       PIPE_TEXTURE_USAGE_DEPTH_STENCIL)) {
+                               stencil,
+                               PIPE_TEXTURE_USAGE_DEPTH_STENCIL)) {
       fb->_Status = GL_FRAMEBUFFER_UNSUPPORTED_EXT;
       return;
    }

src/mesa/state_tracker/st_cb_readpixels.c

@@ -69,6 +69,10 @@ st_read_stencil_pixels(GLcontext *ctx, GLint x, GLint y,
    ubyte *stmap;
    GLint j;
 
+   if (strb->Base.Wrapped) {
+      strb = st_renderbuffer(strb->Base.Wrapped);
+   }
+
    if (st_fb_orientation(ctx->DrawBuffer) == Y_0_TOP) {
       y = ctx->DrawBuffer->Height - y - height;
    }
@@ -366,6 +370,9 @@ st_readpixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
    }
    else if (format == GL_DEPTH_COMPONENT) {
       strb = st_renderbuffer(ctx->ReadBuffer->_DepthBuffer);
+      if (strb->Base.Wrapped) {
+         strb = st_renderbuffer(strb->Base.Wrapped);
+      }
    }
    else {
       /* Read color buffer */