docs: Add release notes for 10.0.3 release.

Just before making the actual release.

Makefile.am

@@ -36,7 +36,6 @@ PACKAGE_NAME = MesaLib-$(PACKAGE_VERSION)
 EXTRA_FILES = \
 	aclocal.m4					\
 	configure					\
-	VERSION					\
 	bin/ar-lib					\
 	bin/compile					\
 	bin/config.sub					\
@@ -46,6 +45,7 @@ EXTRA_FILES = \
 	bin/ltmain.sh					\
 	bin/missing					\
 	bin/ylwrap					\
+        bin/test-driver					\
 	src/glsl/glsl_parser.cpp			\
 	src/glsl/glsl_parser.h				\
 	src/glsl/glsl_lexer.cpp				\

VERSION

@@ -1 +1 @@
-10.0.0-rc1
+10.0.3

bin/.cherry-ignore

@@ -1,3 +1,30 @@
 # Since we've disabled DRI3 completely in 10.0, this commit is no longer
 # necessary.
 f0f202e6b764be803470e27cba9102f14361ae22 glx: conditionaly build dri3 and present loader (v3)
+
+# This patch makes bug #71870 worse, so it won't be cherry picked until that
+# issue can be resolved.  See
+# http://lists.freedesktop.org/archives/mesa-dev/2013-November/048899.html
+068a073c1d4853b5c8f33efdeb481026f42e23a5 meta: fix meta clear of layered framebuffers
+
+# This patch isn't actually necessary because that bug that it fixes isn't in
+# the 10.0 branch.  See
+# http://lists.freedesktop.org/archives/mesa-stable/2013-December/000500.html
+a057b837ddd1c725a7504eedc53c6df05a012773 egl: add HAVE_LIBDRM define, fix EGL X11 platform
+
+# Author requested skipping due to regressions
+# Picking it would require at least also picking:
+# 73c3c7e3, 3e0e9e3b, c59a605c
+b2d1c579bb84a88179072a6a783f8827e218db55 glcpp: Set extension defines after resolving the GLSL version.
+
+# These patches depend on other code not in stable branch.
+# (at least 3b22146dc714b6090f7423abbc4df53d7d1fdaa9)
+e190709119d8eb85c67bfbad5be699d39ad0118e mesa: Ensure that transform feedback refers to the correct program.
+43e77215b13b2f86e461cd8a62b542fc6854dd1c i965/gen7: Use to the correct program when uploading transform feedback state.
+
+# Author requested to ignore these four (since they depend on commits not in
+# stable).
+3313cc269bd428ca96a132d86da5fddc0f27386a i965: Add an option to ignore sample qualifier
+a92e5f7cf63d496ad7830b5cea4bbab287c25b8e i965: Use sample barycentric coordinates with per sample shading
+f5cfb4ae21df8eebfc6b86c0ce858b1c0a9160dd i965: Ignore 'centroid' interpolation qualifier in case of persample shading
+dc2f94bc786768329973403248820a2e5249f102 i965: Ignore 'centroid' interpolation qualifier in case of persample shading

common.py

@@ -100,4 +100,4 @@ def AddOptions(opts):
 	opts.Add(BoolOption('quiet', 'DEPRECATED: profile build', 'yes'))
 	opts.Add(BoolOption('texture_float', 'enable floating-point textures and renderbuffers', 'no'))
 	if host_platform == 'windows':
-		opts.Add(EnumOption('MSVC_VERSION', 'MS Visual C++ version', None, allowed_values=('7.1', '8.0', '9.0', '10.0', '11.0')))
+		opts.Add(EnumOption('MSVC_VERSION', 'MS Visual C++ version', None, allowed_values=('7.1', '8.0', '9.0', '10.0', '11.0', '12.0')))

configure.ac

@@ -29,7 +29,7 @@ AC_SUBST([OSMESA_VERSION])
 dnl Versions for external dependencies
 LIBDRM_REQUIRED=2.4.24
 LIBDRM_RADEON_REQUIRED=2.4.46
-LIBDRM_INTEL_REQUIRED=2.4.38
+LIBDRM_INTEL_REQUIRED=2.4.49
 LIBDRM_NVVIEUX_REQUIRED=2.4.33
 LIBDRM_NOUVEAU_REQUIRED="2.4.33 libdrm >= 2.4.41"
 LIBDRM_FREEDRENO_REQUIRED=2.4.39
@@ -227,6 +227,12 @@ esac
 AC_SUBST([VISIBILITY_CFLAGS])
 AC_SUBST([VISIBILITY_CXXFLAGS])
 
+dnl
+dnl Optional flags, check for compiler support
+dnl
+AX_CHECK_COMPILE_FLAG([-msse4.1], [SSE41_SUPPORTED=1], [SSE41_SUPPORTED=0])
+AM_CONDITIONAL([SSE41_SUPPORTED], [test x$SSE41_SUPPORTED = x1])
+
 dnl
 dnl Hacks to enable 32 or 64 bit build
 dnl
@@ -1527,9 +1533,9 @@ if test "x$enable_gallium_llvm" = xauto; then
 fi
 if test "x$enable_gallium_llvm" = xyes; then
     if test "x$llvm_prefix" != x; then
-        AC_PATH_PROG([LLVM_CONFIG], [llvm-config], [no], ["$llvm_prefix/bin"])
+        AC_PATH_TOOL([LLVM_CONFIG], [llvm-config], [no], ["$llvm_prefix/bin"])
     else
-        AC_PATH_PROG([LLVM_CONFIG], [llvm-config], [no])
+        AC_PATH_TOOL([LLVM_CONFIG], [llvm-config], [no])
     fi
 
     if test "x$LLVM_CONFIG" != xno; then

docs/relnotes/10.0.1.html

@@ -0,0 +1,150 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
+<html lang="en">
+<head>
+  <meta http-equiv="content-type" content="text/html; charset=utf-8">
+  <title>Mesa Release Notes</title>
+  <link rel="stylesheet" type="text/css" href="../mesa.css">
+</head>
+<body>
+
+<div class="header">
+  <h1>The Mesa 3D Graphics Library</h1>
+</div>
+
+<iframe src="../contents.html"></iframe>
+<div class="content">
+
+<h1>Mesa 10.0.1 Release Notes / (December 12, 2013)</h1>
+
+<p>
+Mesa 10.0.1 is a bug fix release which fixes bugs found since the 10.0 release.
+</p>
+<p>
+Mesa 10.0.1 implements the OpenGL 3.3 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 3.3.  OpenGL
+3.3 is <strong>only</strong> available if requested at context creation
+because compatibility contexts not supported.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+0a72ca5b36046a658bf6038326ff32ed  MesaLib-10.0.1.tar.bz2
+01bde35c912e504ba62caf1ef9f7022c  MesaLib-10.0.1.tar.gz
+59a174a11a89e6b1b8ee9c3f7e3c388c  MesaLib-10.0.1.zip
+</pre>
+
+
+<h2>New features</h2>
+<p>None</p>
+
+<h2>Bug fixes</h2>
+
+<p>This list is likely incomplete.</p>
+
+<ul>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=64323">Bug 64323</a> - Severe misrendering in Left 4 Dead 2</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=68838">Bug 68838</a> - GLSL: struct declarations produce a &quot;empty declaration warning&quot; in 9.2</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=69155">Bug 69155</a> - [NV50 gallium] [piglit] bin/varying-packing-simple triggers memory corruption/failures</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70250">Bug 70250</a> - weston-terminal rendering corrupted with output transform 90 and 270</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70601">Bug 70601</a> - [SNB Bisected]Piglit spec/ARB_texture_float/multisample-formats 2 GL_ARB_texture_float fails</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72230">Bug 72230</a> - Unable to extract MesaLib-10.0.0.tar.{gz,bz2} with bsdtar</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72325">Bug 72325</a> - [swrast] piglit glean fbo regression</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72327">Bug 72327</a> - [swrast] piglit glean pointSprite regression</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>The full set of changes can be viewed by using the following git command:</p>
+
+<pre>
+  git log mesa-10.0..mesa-10.0.1
+</pre>
+
+<p>Axel Davy (2):</p>
+<ul>
+  <li>egl/wayland: Flush the wl_display at the end of SwapBuffers</li>
+  <li>Enable throttling in SwapBuffers</li>
+</ul>
+
+<p>Chad Versace (2):</p>
+<ul>
+  <li>i965/hsw: Apply non-msrt fast color clear w/a to all HSW GTs</li>
+  <li>i965: Add extra-alignment for non-msrt fast color clear for all hw (v2)</li>
+</ul>
+
+<p>Dave Airlie (1):</p>
+<ul>
+  <li>swrast: fix readback regression since inversion fix</li>
+</ul>
+
+<p>Emil Velikov (1):</p>
+<ul>
+  <li>automake: include only one copy VERSION in tarball</li>
+</ul>
+
+<p>Ian Romanick (3):</p>
+<ul>
+  <li>docs: Add 10.0 release md5sums</li>
+  <li>Remove a057b83 from the pick list</li>
+  <li>glsl: Don't emit empty declaration warning for a struct specifier</li>
+</ul>
+
+<p>Ilia Mirkin (8):</p>
+<ul>
+  <li>mesa: don't leak performance monitors on context destroy</li>
+  <li>nv50: Fix GPU_READING/WRITING bit removal</li>
+  <li>nouveau: avoid leaking fences while waiting</li>
+  <li>nv50: wait on the buf's fence before sticking it into pushbuf</li>
+  <li>nv50: enable h264 and mpeg4 for nv98+ (vp3, vp4.0)</li>
+  <li>nouveau/video: update h264 picparm field names based on usage</li>
+  <li>nouveau/video: update a few more h264 picparm field names</li>
+  <li>nv50: report 15 max inputs for fragment programs</li>
+</ul>
+
+<p>Jordan Justen (1):</p>
+<ul>
+  <li>dri megadriver_stub: add compatibility for older DRI loaders</li>
+</ul>
+
+<p>Kristian Høgsberg (2):</p>
+<ul>
+  <li>egl/wayland: Damage INT32_MAX x INT32_MAX region for eglSwapBuffers</li>
+  <li>egl/wayland: Send commit after flushing the driver context</li>
+</ul>
+
+<p>Maarten Lankhorst (1):</p>
+<ul>
+  <li>nouveau: Fix compiler warning regression</li>
+</ul>
+
+<p>Paul Berry (1):</p>
+<ul>
+  <li>i965/gen6: Fix multisample resolve blits for luminance/intensity 32F formats.</li>
+</ul>
+
+<p>Thomas Hellstrom (1):</p>
+<ul>
+  <li>st/xa: Bump major version number to 2</li>
+</ul>
+
+<p>Tom Stellard (2):</p>
+<ul>
+  <li>r300/compiler/tests: Fix segfault</li>
+  <li>r300/compiler/tests: Fix line length check in test parser</li>
+</ul>
+
+</div>
+</body>
+</html>

docs/relnotes/10.0.2.html

@@ -0,0 +1,161 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
+<html lang="en">
+<head>
+  <meta http-equiv="content-type" content="text/html; charset=utf-8">
+  <title>Mesa Release Notes</title>
+  <link rel="stylesheet" type="text/css" href="../mesa.css">
+</head>
+<body>
+
+<div class="header">
+  <h1>The Mesa 3D Graphics Library</h1>
+</div>
+
+<iframe src="../contents.html"></iframe>
+<div class="content">
+
+<h1>Mesa 10.0.2 Release Notes / (January 9, 2014)</h1>
+
+<p>
+Mesa 10.0.2 is a bug fix release which fixes bugs found since the 10.0.1 release.
+</p>
+<p>
+Mesa 10.0.2 implements the OpenGL 3.3 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 3.3.  OpenGL
+3.3 is <strong>only</strong> available if requested at context creation
+because compatibility contexts not supported.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+de7d14baf0101b697c140d2f47ef27e9  MesaLib-10.0.2.tar.gz
+8544c0ab3e438a08b5103421ea15b6d2  MesaLib-10.0.2.tar.bz2
+181b0d6c1afca38e98a930d0e564ed90  MesaLib-10.0.2.zip
+</pre>
+
+
+<h2>New features</h2>
+<p>None</p>
+
+<h2>Bug fixes</h2>
+
+<p>This list is likely incomplete.</p>
+
+<ul>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70740">Bug 70740</a> - HiZ on SNB causes GPU hang with WebGL web app</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72026">Bug 72026</a> - SIGSEGV in fs_visitor::visit(ir_dereference_variable*)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72264">Bug 72264</a> - GLSL error reporting</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72369">Bug 72369</a> - glitches in serious sam 3 with the sb shader backend</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>The full set of changes can be viewed by using the following git command:</p>
+
+<pre>
+  git log mesa-10.0.1..mesa-10.0.2
+</pre>
+
+<p>Aaron Watry (8):</p>
+<ul>
+  <li>clover: Remove unused variable</li>
+  <li>pipe_loader/sw: close dev-&gt;lib when initialization fails</li>
+  <li>radeon/compute: Stop leaking LLVMContexts in radeon_llvm_parse_bitcode</li>
+  <li>r600/compute: Free compiled kernels when deleting compute state</li>
+  <li>r600/compute: Use the correct FREE macro when deleting compute state</li>
+  <li>radeon/llvm: Free target data at end of optimization</li>
+  <li>st/vdpau: Destroy context when initialization fails</li>
+  <li>r600/pipe: Stop leaking context-&gt;start_compute_cs_cmd.buf on EG/CM</li>
+</ul>
+
+<p>Alex Deucher (1):</p>
+<ul>
+  <li>r600g: fix SUMO2 pci id</li>
+</ul>
+
+<p>Alexander von Gluck IV (1):</p>
+<ul>
+  <li>Haiku: Add in public GL kit headers</li>
+</ul>
+
+<p>Anuj Phogat (1):</p>
+<ul>
+  <li>mesa: Fix error code generation in glBeginConditionalRender()</li>
+</ul>
+
+<p>Carl Worth (2):</p>
+<ul>
+  <li>docs: Add md5sums for the 10.0.1 release.</li>
+  <li>Update version to 10.0.2</li>
+</ul>
+
+<p>Chad Versace (1):</p>
+<ul>
+  <li>i965/gen6: Fix HiZ hang in WebGL Google Maps</li>
+</ul>
+
+<p>Erik Faye-Lund (1):</p>
+<ul>
+  <li>glcpp: error on multiple #else/#elif directives</li>
+</ul>
+
+<p>Henri Verbeet (1):</p>
+<ul>
+  <li>i915: Add support for gl_FragData[0] reads.</li>
+</ul>
+
+<p>Ilia Mirkin (1):</p>
+<ul>
+  <li>nv50: fix a small leak on context destroy</li>
+</ul>
+
+<p>Jonathan Liu (2):</p>
+<ul>
+  <li>st/mesa: use pipe_sampler_view_release()</li>
+  <li>llvmpipe: use pipe_sampler_view_release() to avoid segfault</li>
+</ul>
+
+<p>Kenneth Graunke (2):</p>
+<ul>
+  <li>i965: Fix 3DSTATE_PUSH_CONSTANT_ALLOC_PS packet creation.</li>
+  <li>Revert "mesa: Remove GLXContextID typedef from glx.h."</li>
+</ul>
+
+<p>Kevin Rogovin (1):</p>
+<ul>
+  <li>Use line number information from entire function expression</li>
+</ul>
+
+<p>Kristian Høgsberg (1):</p>
+<ul>
+  <li>dri_util: Don't assume __DRIcontext-&gt;driverPrivate is a gl_context</li>
+</ul>
+
+<p>Marek Olšák (2):</p>
+<ul>
+  <li>mesa: fix interpretation of glClearBuffer(drawbuffer)</li>
+  <li>st/mesa: fix glClear with multiple colorbuffers and different formats</li>
+</ul>
+
+<p>Paul Berry (2):</p>
+<ul>
+  <li>glsl: Teach ir_variable_refcount about ir_loop::counter variables.</li>
+  <li>glsl: Fix inconsistent assumptions about ir_loop::counter.</li>
+</ul>
+
+<p>Vadim Girlin (1):</p>
+<ul>
+  <li>r600g/sb: fix stack size computation on evergreen</li>
+</ul>
+
+</div>
+</body>
+</html>

docs/relnotes/10.0.3.html

@@ -0,0 +1,203 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
+<html lang="en">
+<head>
+  <meta http-equiv="content-type" content="text/html; charset=utf-8">
+  <title>Mesa Release Notes</title>
+  <link rel="stylesheet" type="text/css" href="../mesa.css">
+</head>
+<body>
+
+<div class="header">
+  <h1>The Mesa 3D Graphics Library</h1>
+</div>
+
+<iframe src="../contents.html"></iframe>
+<div class="content">
+
+<h1>Mesa 10.0.3 Release Notes / (February 3, 2014)</h1>
+
+<p>
+Mesa 10.0.3 is a bug fix release which fixes bugs found since the 10.0.2 release.
+</p>
+<p>
+Mesa 10.0.3 implements the OpenGL 3.3 API, but the version reported by
+glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) /
+glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 3.3.  OpenGL
+3.3 is <strong>only</strong> available if requested at context creation
+because compatibility contexts not supported.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+</pre>
+
+
+<h2>New features</h2>
+<p>None</p>
+
+<h2>Bug fixes</h2>
+
+<p>This list is likely incomplete.</p>
+
+<ul>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72708">Bug 72708</a> - Master fails to build with older gcc due to -msse4.1</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72926">Bug 72926</a> - [REGRESSION,swrast] Memory-related crash with anti-aliasing enabled</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=73096">Bug 73096</a> - Query GL_RGBA_SIGNED_COMPONENTS_EXT missing</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=73100">Bug 73100</a> - Please use AC_PATH_TOOL instead of AC_PATH_PROG for llvm-config</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=73418">Bug 73418</a> - OpenCL hangs graphics on CAYMAN</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=73473">Bug 73473</a> - Potential crash bug in src/gallium/auxiliary/rtasm/rtasm_execmem.c</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=73915">Bug 73915</a> - sample shading + centroid broken since f5cfb4a</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=73956">Bug 73956</a> - SIGSEGV when passing GL_NONE to glReadBuffer</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=74026">Bug 74026</a> - Compiler rejects chained assignments involving array dereferences</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>The full set of changes can be viewed by using the following git command:</p>
+
+<pre>
+  git log mesa-10.0.2..mesa-10.0.3
+</pre>
+
+<p>Aaron Watry (2):</p>
+<ul>
+  <li>radeon: Move gfx/dma cs cleanup to r600_common_context_cleanup</li>
+  <li>st/dri: prevent leak of dri option default values</li>
+</ul>
+
+<p>Andreas Fänger (1):</p>
+<ul>
+  <li>swrast: fix delayed texel buffer allocation regression for OpenMP</li>
+</ul>
+
+<p>Anuj Phogat (3):</p>
+<ul>
+  <li>glsl: Disable ARB_texture_rectangle in shader version 100.</li>
+  <li>i965: Use sample barycentric coordinates with per sample shading</li>
+  <li>i965: Ignore 'centroid' interpolation qualifier in case of persample shading</li>
+</ul>
+
+<p>Brian Paul (3):</p>
+<ul>
+  <li>mesa: implement missing glGet(GL_RGBA_SIGNED_COMPONENTS_EXT) query</li>
+  <li>st/mesa: fix glReadBuffer(GL_NONE) segfault</li>
+  <li>draw: fix incorrect vertex size computation in LLVM drawing code</li>
+</ul>
+
+<p>Carl Worth (5):</p>
+<ul>
+  <li>Add md5sums for 10.0.2. release.</li>
+  <li>cherry-ignore: Ignore several patches not yet ready for the stable branch</li>
+  <li>Drop another couple of patches.</li>
+  <li>cherry-ignore: Ignore 4 patches at teh request of the author, (Anuj).</li>
+  <li>Update version to 10.0.3</li>
+</ul>
+
+<p>Chad Versace (1):</p>
+<ul>
+  <li>i965/gen6/blorp: Emit more flushes to workaround hangs</li>
+</ul>
+
+<p>Chris Forbes (1):</p>
+<ul>
+  <li>i965: fold offset into coord for textureOffset(gsampler2DRect)</li>
+</ul>
+
+<p>Emil Velikov (5):</p>
+<ul>
+  <li>mesa: use signed temporary variable to store _ColorDrawBufferIndexes</li>
+  <li>st/mesa: use signed temporary variable to store _ColorDrawBufferIndexes</li>
+  <li>nv50: access only the available amount of textures</li>
+  <li>nv50: access only the available amount of constbuf</li>
+  <li>gallium/rtasm: handle mmap failures appropriately</li>
+</ul>
+
+<p>Eric Anholt (2):</p>
+<ul>
+  <li>i965: Fix handling of MESA_pack_invert in blit (PBO) readpixels.</li>
+  <li>i965: Don't do the temporary-and-blit-copy for INVALIDATE_RANGE maps.</li>
+</ul>
+
+<p>Ian Romanick (2):</p>
+<ul>
+  <li>mesa: Add COMPRESSED_RGBA_S3TC_DXT1_EXT to COMPRESSED_TEXTURE_FORMATS for GLES</li>
+  <li>radeon / r200: Pass the API into _mesa_initialize_context</li>
+</ul>
+
+<p>Ilia Mirkin (2):</p>
+<ul>
+  <li>mesa: fix GL_COLOR_SUM enum for drivers without ARB_vertex_program</li>
+  <li>st/vdpau: don't return a device if the screen doesn't support NPOT</li>
+</ul>
+
+<p>José Fonseca (1):</p>
+<ul>
+  <li>mesa: Use IROUND instead of roundf.</li>
+</ul>
+
+<p>Kenneth Graunke (2):</p>
+<ul>
+  <li>glsl: Rename "expr" to "lhs_expr" in vector_extract munging code.</li>
+  <li>glsl: Fix chained assignments of vector channels.</li>
+</ul>
+
+<p>Lauri Kasanen (1):</p>
+<ul>
+  <li>mesa: Fix build to properly check for supported compiler flags</li>
+</ul>
+
+<p>Marek Olšák (2):</p>
+<ul>
+  <li>st/mesa: use sRGB formats for MSAA resolving if destination is sRGB</li>
+  <li>gallium/util: util_format_srgb should not return FORMAT_NONE for sRGB formats</li>
+</ul>
+
+<p>Matt Turner (2):</p>
+<ul>
+  <li>glcpp: Define GL_EXT_shader_integer_mix in both GL and ES.</li>
+  <li>glx: Update glxext.h to revision 24777.</li>
+</ul>
+
+<p>Michał Górny (1):</p>
+<ul>
+  <li>Use AC_PATH_TOOL instead of AC_PATH_PROG for llvm-config.</li>
+</ul>
+
+<p>Paul Berry (1):</p>
+<ul>
+  <li>i965: Ensure that all necessary state is re-emitted if we run out of aperture.</li>
+</ul>
+
+<p>Paul Seidler (1):</p>
+<ul>
+  <li>build: move ARCH_LIBS definition outside of ASM definition</li>
+</ul>
+
+<p>Thomas Sondergaard (4):</p>
+<ul>
+  <li>mesa: Preliminary support for MSVC_VERSION=12.0</li>
+  <li>mesa: Fix compile error with MSVC 2013</li>
+  <li>mesa: Work around internal compiler error</li>
+  <li>mesa: Namespace qualify fma to override ambiguity with fma from math.h</li>
+</ul>
+
+<p>Tom Stellard (1):</p>
+<ul>
+  <li>r600g/compute: Emit DEALLOC_STATE on cayman after dispatching a compute shader.</li>
+</ul>
+
+</div>
+</body>
+</html>

docs/relnotes/10.0.html

@@ -14,7 +14,7 @@
 <iframe src="../contents.html"></iframe>
 <div class="content">
 
-<h1>Mesa 10.0 Release Notes / TBD</h1>
+<h1>Mesa 10.0 Release Notes / (November 30th, 2013)</h1>
 
 <p>
 Mesa 10.0 is a new development release.
@@ -33,7 +33,9 @@ because compatibility contexts are not supported.
 
 <h2>MD5 checksums</h2>
 <pre>
-TBD.
+b38626b96c664db67a534d7859682436  MesaLib-10.0.0.tar.gz
+f3fe55d9735bea158bbe97ed9a0da819  MesaLib-10.0.0.tar.bz2
+c6ee1ce51e3bf35947d2978b872daf51  MesaLib-10.0.0.zip
 </pre>
 
 
@@ -55,16 +57,89 @@ Note: some of the new features are only available with certain drivers.
 <li>GL_ARB_vertex_attrib_binding</li>
 <li>GL_ARB_vertex_type_10f_11f_11f_rev on i965 and r600g</li>
 <li>GL_KHR_debug</li>
+<li>GLX_MESA_query_renderer</li>
 </ul>
 
 
 <h2>Bug fixes</h2>
 
-TBD.
+<p>Attempts have been made to <b>not</b> include bugs fixed in previous 9.2
+releases or bugs that were regressions during 10.0 development. This list is
+likely incomplete.</p>
+
+<ul>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=47755">Bug 47755</a> - [glsl-compiler] no error checking when Interpolation qualifier for built-in variable is different in vertex and fragment shader</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=52171">Bug 52171</a> - [gallium/r600/clover] Simple benchmarks failed to run</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=53077">Bug 53077</a> - [IVB] Output error with msaa when both of framebuffer and source color's alpha are not 1</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=54867">Bug 54867</a> - bug in r300 compiler</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=60929">Bug 60929</a> - [r600-llvm] mono games with opengl are blocking on start</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=62142">Bug 62142</a> - Mesa/demo mipmap_limits upside down with running by SOFTWARE</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=62698">Bug 62698</a> - [bisected] WebGL demo &quot;Consumed&quot;: texstate.c:628: update_texture_state: Assertion „__builtin_popcount(enabledTargets) == 1“ failed.</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=64225">Bug 64225</a> - bfgminer --scyte generates Segmentation Fault on Northern Island</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=64226">Bug 64226</a> - python-opencl package generate segmentation fault at pipe_r600.so</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=64261">Bug 64261</a> - [SNB Bisected]Ogles3conform GL3Tests_color_buffer_float_color_buffer_float_clamp_fixed.test fail</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=66213">Bug 66213</a> - Certain Mesa Demos Rendering Inverted (vertically)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=66806">Bug 66806</a> - [softpipe] glxgears floating point exception</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=67921">Bug 67921</a> - [bisected commit 883987] crosscompiling fails with util/u_cpu_detect.c:247:4: error: 'asm' undeclared (first use in this function)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=68162">Bug 68162</a> - [radeonsi] texture rendering is broken in Source-Engine games</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=68451">Bug 68451</a> - Texture flicker in native Dota2 in mesa 9.2.0rc1</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=68503">Bug 68503</a> - Graphical glitches in Serious Sam 3 when SB is enabled</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=68792">Bug 68792</a> - Problems during playback of h264 files using UVD and VLC on AMD E-350 CPU</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=68845">Bug 68845</a> - VDPAU/UVD regression</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=69078">Bug 69078</a> - Modern Warfare (1, 2 and 3) broken in Wine on SNB</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=69321">Bug 69321</a> - starting openCL crashes/boots system</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70042">Bug 70042</a> - Major texture flickering in Dota 2 (r600g on HD 6950)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70088">Bug 70088</a> - Glamor on r600g crashes Xserver</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70123">Bug 70123</a> - Freeze caused by 'winsys/radeon: remove cs_queue_empty' commit</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70327">Bug 70327</a> - Casting floating point variable to integer not working properly while constant gets converted properly</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70891">Bug 70891</a> - CL_INVALID_BUILD_OPTIONS results in CL_INVALID_DEVICE when asking for build log</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70913">Bug 70913</a> - [PIGLIT,radeonsi] crash in &quot;spec/EXT_framebuffer_multisample/sample-alpha-to-coverage 4 depth&quot; (buffer overflow)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=71022">Bug 71022</a> - configure: error: Expat required for DRI.</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=71110">Bug 71110</a> - xorg_driver.c:1030:2: error: too many arguments to function ‘DamageUnregister’</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=71172">Bug 71172</a> - Segfault when running glxinfo. NV25GL [Quadro4 900 XGL]</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=71512">Bug 71512</a> - dlopen.h:54: undefined reference to `dlopen'</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=71870">Bug 71870</a> - Metro: Last Light rendering issues</li>
+
+</ul>
 
 <h2>Changes</h2>
 
-TBD.
+<ul>
+<li>Removed X.Org state tracker (unmaintained and broken)</li>
+<li>Removed the video-accel r300 targets</li>
+<li>Removed the video-accel softpipe targets</li>
+</ul>
 
 </div>
 </body>

include/GL/glx.h

@@ -168,6 +168,7 @@ typedef XID GLXDrawable;
 /* GLX 1.3 and later */
 typedef struct __GLXFBConfigRec *GLXFBConfig;
 typedef XID GLXFBConfigID;
+typedef XID GLXContextID;
 typedef XID GLXWindow;
 typedef XID GLXPbuffer;
 

include/GL/glxext.h

@@ -33,10 +33,10 @@ extern "C" {
 ** used to make the header, and the header can be found at
 **   http://www.opengl.org/registry/
 **
-** Khronos $Revision: 23422 $ on $Date: 2013-10-08 15:40:45 -0700 (Tue, 08 Oct 2013) $
+** Khronos $Revision: 24777 $ on $Date: 2014-01-14 14:02:32 -0800 (Tue, 14 Jan 2014) $
 */
 
-#define GLX_GLXEXT_VERSION 20131008
+#define GLX_GLXEXT_VERSION 20140114
 
 /* Generated C header for:
  * API: glx
@@ -49,6 +49,7 @@ extern "C" {
 
 #ifndef GLX_VERSION_1_3
 #define GLX_VERSION_1_3 1
+typedef XID GLXContextID;
 typedef struct __GLXFBConfigRec *GLXFBConfig;
 typedef XID GLXWindow;
 typedef XID GLXPbuffer;
@@ -272,7 +273,6 @@ __GLXextFuncPtr glXGetProcAddressARB (const GLubyte *procName);
 
 #ifndef GLX_EXT_import_context
 #define GLX_EXT_import_context 1
-typedef XID GLXContextID;
 #define GLX_SHARE_CONTEXT_EXT             0x800A
 #define GLX_VISUAL_ID_EXT                 0x800B
 #define GLX_SCREEN_EXT                    0x800C
@@ -433,6 +433,14 @@ void glXCopyImageSubDataNV (Display *dpy, GLXContext srcCtx, GLuint srcName, GLe
 #endif
 #endif /* GLX_NV_copy_image */
 
+#ifndef GLX_NV_delay_before_swap
+#define GLX_NV_delay_before_swap 1
+typedef Bool ( *PFNGLXDELAYBEFORESWAPNVPROC) (Display *dpy, GLXDrawable drawable, GLfloat seconds);
+#ifdef GLX_GLXEXT_PROTOTYPES
+Bool glXDelayBeforeSwapNV (Display *dpy, GLXDrawable drawable, GLfloat seconds);
+#endif
+#endif /* GLX_NV_delay_before_swap */
+
 #ifndef GLX_NV_float_buffer
 #define GLX_NV_float_buffer 1
 #define GLX_FLOAT_COMPONENTS_NV           0x20B0

include/HaikuGL/GLRenderer.h

@@ -0,0 +1,76 @@
+/*
+ * Copyright 2006, Philippe Houdoin. All rights reserved.
+ * Distributed under the terms of the MIT License.
+
+ * This header defines BGLRenderer, the base class making up
+ * the Haiku GL renderer add-ons (essentially selfcontained C++
+ * shared libraries that do the actual rendering such as
+ * libswpipe.so and libswrast.so)
+ */
+#ifndef GLRENDERER_H
+#define GLRENDERER_H
+
+
+#include <BeBuild.h>
+#include <GLView.h>
+
+
+class BGLDispatcher;
+class GLRendererRoster;
+
+typedef unsigned long renderer_id;
+
+class BGLRenderer
+{
+							// Private unimplemented copy constructors
+							BGLRenderer(const BGLRenderer &);
+							BGLRenderer & operator=(const BGLRenderer &);
+	
+public:
+							BGLRenderer(BGLView *view, ulong bgl_options,
+								BGLDispatcher *dispatcher);
+	virtual					~BGLRenderer();
+
+	void 					Acquire();
+	void					Release();
+
+	virtual void			LockGL();
+	virtual void 			UnlockGL();
+	
+	virtual	void 			SwapBuffers(bool VSync = false);
+	virtual	void			Draw(BRect updateRect);
+	virtual status_t		CopyPixelsOut(BPoint source, BBitmap *dest);
+	virtual status_t    	CopyPixelsIn(BBitmap *source, BPoint dest);
+
+ 	virtual void			FrameResized(float width, float height);
+	
+	virtual void			DirectConnected(direct_buffer_info *info);
+	virtual void			EnableDirectMode(bool enabled);
+
+	inline	int32			ReferenceCount() const { return fRefCount; };
+	inline	ulong			Options() const { return fOptions; };
+	inline	BGLView*		GLView() { return fView; };
+	inline	BGLDispatcher*	GLDispatcher() { return fDispatcher; };
+
+private:
+	friend class GLRendererRoster;
+
+	virtual status_t		_Reserved_Renderer_0(int32, void *);
+	virtual status_t		_Reserved_Renderer_1(int32, void *);
+	virtual status_t		_Reserved_Renderer_2(int32, void *);
+	virtual status_t		_Reserved_Renderer_3(int32, void *);
+	virtual status_t		_Reserved_Renderer_4(int32, void *);
+
+	volatile int32			fRefCount;	// How much we're still usefull?
+	BGLView*				fView;		// Never forget who is the boss!
+	ulong					fOptions;	// Keep that tune in memory
+	BGLDispatcher*			fDispatcher;// Our personal GL API call dispatcher
+
+	GLRendererRoster*		fOwningRoster;
+	renderer_id				fID;
+};
+
+extern "C" _EXPORT BGLRenderer* instantiate_gl_renderer(BGLView *view, ulong options, BGLDispatcher *dispatcher);
+
+
+#endif	// GLRENDERER_H

include/HaikuGL/GLView.h

@@ -0,0 +1,193 @@
+/*
+ * Copyright 2008-2013, Haiku, Inc. All Rights Reserved.
+ * Distributed under the terms of the MIT License.
+ *
+ * This header defines BGLView, the base class making up
+ * the Haiku GL Kit.
+ *
+ */
+#ifndef BGLVIEW_H
+#define BGLVIEW_H
+
+
+#include <GL/gl.h>
+
+#define BGL_RGB			0
+#define BGL_INDEX		1
+#define BGL_SINGLE		0
+#define BGL_DOUBLE		2
+#define BGL_DIRECT		0
+#define BGL_INDIRECT	4
+#define BGL_ACCUM		8
+#define BGL_ALPHA		16
+#define BGL_DEPTH		32
+#define BGL_OVERLAY		64
+#define BGL_UNDERLAY	128
+#define BGL_STENCIL		512
+
+#ifdef __cplusplus
+
+#include <AppKit.h>
+#include <Bitmap.h>
+#include <DirectWindow.h>
+#include <View.h>
+#include <Window.h>
+#include <WindowScreen.h>
+
+
+struct glview_direct_info;
+class BGLRenderer;
+class GLRendererRoster;
+
+class BGLView : public BView {
+public:
+							BGLView(BRect rect, const char* name,
+								ulong resizingMode, ulong mode,
+								ulong options);
+	virtual					~BGLView();
+
+			void			LockGL();
+			void			UnlockGL();
+			void			SwapBuffers();
+			void			SwapBuffers(bool vSync);
+
+			BView*			EmbeddedView(); // deprecated, returns NULL
+			void*			GetGLProcAddress(const char* procName);
+
+			status_t		CopyPixelsOut(BPoint source, BBitmap *dest);
+			status_t		CopyPixelsIn(BBitmap *source, BPoint dest);
+
+	// Mesa's GLenum is uint where Be's ones was ulong!
+	virtual	void			ErrorCallback(unsigned long errorCode);
+
+	virtual void			Draw(BRect updateRect);
+	virtual void			AttachedToWindow();
+	virtual void			AllAttached();
+	virtual void			DetachedFromWindow();
+	virtual void			AllDetached();
+
+	virtual void			FrameResized(float newWidth, float newHeight);
+	virtual status_t		Perform(perform_code d, void *arg);
+
+	virtual status_t		Archive(BMessage *data, bool deep = true) const;
+
+	virtual void			MessageReceived(BMessage *message);
+	virtual void			SetResizingMode(uint32 mode);
+
+	virtual void			Show();
+	virtual void			Hide();
+
+	virtual BHandler*		ResolveSpecifier(BMessage *msg, int32 index,
+								BMessage *specifier, int32 form,
+								const char *property);
+	virtual status_t		GetSupportedSuites(BMessage *data);
+
+			void			DirectConnected(direct_buffer_info *info);
+			void			EnableDirectMode(bool enabled);
+
+			void*			getGC()	{ return fGc; } // ???
+
+	virtual void			GetPreferredSize(float* width, float* height);
+
+private:
+
+	virtual void			_ReservedGLView1();
+	virtual void			_ReservedGLView2();
+	virtual void			_ReservedGLView3();
+	virtual void			_ReservedGLView4();
+	virtual void			_ReservedGLView5();
+	virtual void			_ReservedGLView6();
+	virtual void			_ReservedGLView7();
+	virtual void			_ReservedGLView8();
+
+							BGLView(const BGLView &);
+							BGLView &operator=(const BGLView &);
+
+			void			_DitherFront();
+			bool			_ConfirmDither();
+			void			_Draw(BRect rect);
+			void			_CallDirectConnected();
+
+			void*			fGc;
+			uint32			fOptions;
+			uint32			fDitherCount;
+			BLocker			fDrawLock;
+			BLocker			fDisplayLock;
+			glview_direct_info* fClipInfo;
+
+			BGLRenderer*	fRenderer;
+			GLRendererRoster* fRoster;
+
+			BBitmap*		fDitherMap;
+			BRect			fBounds;
+			int16*			fErrorBuffer[2];
+			uint64			_reserved[8];
+
+			void			_LockDraw();
+			void			_UnlockDraw();
+
+// BeOS compatibility
+private:
+							BGLView(BRect rect, char* name,
+								ulong resizingMode, ulong mode,
+								ulong options);
+};
+
+
+class BGLScreen : public BWindowScreen {
+public:
+							BGLScreen(char* name,
+								ulong screenMode, ulong options,
+								status_t *error, bool debug=false);
+							~BGLScreen();
+
+			void			LockGL();
+			void			UnlockGL();
+			void			SwapBuffers();
+							// Mesa's GLenum is uint where Be's ones was ulong!
+	virtual	void			ErrorCallback(unsigned long errorCode);
+
+	virtual void			ScreenConnected(bool connected);
+	virtual void			FrameResized(float width, float height);
+	virtual status_t		Perform(perform_code code, void *arg);
+
+	virtual status_t		Archive(BMessage *data, bool deep = true) const;
+	virtual void			MessageReceived(BMessage *message);
+
+	virtual void			Show();
+	virtual void			Hide();
+
+	virtual BHandler*		ResolveSpecifier(BMessage *message,
+								int32 index,
+								BMessage *specifier,
+								int32 form,
+								const char *property);
+	virtual status_t		GetSupportedSuites(BMessage *data);
+
+private:
+
+	virtual void			_ReservedGLScreen1();
+	virtual void			_ReservedGLScreen2();
+	virtual void			_ReservedGLScreen3();
+	virtual void			_ReservedGLScreen4();
+	virtual void			_ReservedGLScreen5();
+	virtual void			_ReservedGLScreen6();
+	virtual void			_ReservedGLScreen7();
+	virtual void			_ReservedGLScreen8();
+
+							BGLScreen(const BGLScreen &);
+							BGLScreen &operator=(const BGLScreen &);
+
+			void*			fGc;
+			long			fOptions;
+			BLocker			fDrawLock;
+
+			int32			fColorSpace;
+			uint32			fScreenMode;
+
+			uint64			_reserved[7];
+};
+
+#endif // __cplusplus
+
+#endif // BGLVIEW_H

include/HaikuGL/OpenGLKit.h

@@ -0,0 +1,10 @@
+/*
+ * Master include file for the Haiku OpenGL Kit.
+ */
+
+#include <GL/gl.h>
+#include <GLView.h>
+
+// Projects needing GL/glu.h and GL/glut.h should now
+// include these headers independently as glu and glut
+// are no longe core parts of mesa

include/HaikuGL/README

@@ -0,0 +1,28 @@
+These headers make up the Haiku Op*nGL kit.
+
+Headers in this directory preserve some BeOS™ compatibility
+compatibility, so changes should be mentioned to the Haiku
+project mailing list.
+
+http://haiku-os.org
+
+Normal Haiku Op*enGL layout:
+
+  * headers/os/OpenGLKit.h
+  * headers/os/opengl/GLView.h
+  * headers/os/opengl/GLRenderer.h
+  * headers/os/opengl/GL/gl.h
+  * headers/os/opengl/GL/gl_mangle.h
+  * headers/os/opengl/GL/glext.h
+  * headers/os/opengl/GL/osmesa.h (needed?)
+
+Extras:
+
+  * headers/os/opengl/GL/glu.h
+  * headers/os/opengl/GL/glut.h
+
+OpenGL™ is a trademark of SGI. The usage of this trademark
+in the Haiku GL Kit is not a sign of any certification or
+endorsement by SGI or its affiliates. Usage is purely to
+allow legacy compatibility with the BeOS™ and its 3D GL
+rendering subsystem.

include/pci_ids/r600_pci_ids.h

@@ -208,7 +208,7 @@ CHIPSET(0x9644, SUMO2_9644, SUMO2)
 CHIPSET(0x9645, SUMO2_9645, SUMO2)
 CHIPSET(0x9647, SUMO_9647,  SUMO)
 CHIPSET(0x9648, SUMO_9648,  SUMO)
-CHIPSET(0x9649, SUMO_9649,  SUMO)
+CHIPSET(0x9649, SUMO2_9649, SUMO2)
 CHIPSET(0x964a, SUMO_964A,  SUMO)
 CHIPSET(0x964b, SUMO_964B,  SUMO)
 CHIPSET(0x964c, SUMO_964C,  SUMO)

m4/ax_check_compile_flag.m4

@@ -0,0 +1,72 @@
+# ===========================================================================
+#   http://www.gnu.org/software/autoconf-archive/ax_check_compile_flag.html
+# ===========================================================================
+#
+# SYNOPSIS
+#
+#   AX_CHECK_COMPILE_FLAG(FLAG, [ACTION-SUCCESS], [ACTION-FAILURE], [EXTRA-FLAGS])
+#
+# DESCRIPTION
+#
+#   Check whether the given FLAG works with the current language's compiler
+#   or gives an error.  (Warnings, however, are ignored)
+#
+#   ACTION-SUCCESS/ACTION-FAILURE are shell commands to execute on
+#   success/failure.
+#
+#   If EXTRA-FLAGS is defined, it is added to the current language's default
+#   flags (e.g. CFLAGS) when the check is done.  The check is thus made with
+#   the flags: "CFLAGS EXTRA-FLAGS FLAG".  This can for example be used to
+#   force the compiler to issue an error when a bad flag is given.
+#
+#   NOTE: Implementation based on AX_CFLAGS_GCC_OPTION. Please keep this
+#   macro in sync with AX_CHECK_{PREPROC,LINK}_FLAG.
+#
+# LICENSE
+#
+#   Copyright (c) 2008 Guido U. Draheim <guidod@gmx.de>
+#   Copyright (c) 2011 Maarten Bosmans <mkbosmans@gmail.com>
+#
+#   This program is free software: you can redistribute it and/or modify it
+#   under the terms of the GNU General Public License as published by the
+#   Free Software Foundation, either version 3 of the License, or (at your
+#   option) any later version.
+#
+#   This program is distributed in the hope that it will be useful, but
+#   WITHOUT ANY WARRANTY; without even the implied warranty of
+#   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
+#   Public License for more details.
+#
+#   You should have received a copy of the GNU General Public License along
+#   with this program. If not, see <http://www.gnu.org/licenses/>.
+#
+#   As a special exception, the respective Autoconf Macro's copyright owner
+#   gives unlimited permission to copy, distribute and modify the configure
+#   scripts that are the output of Autoconf when processing the Macro. You
+#   need not follow the terms of the GNU General Public License when using
+#   or distributing such scripts, even though portions of the text of the
+#   Macro appear in them. The GNU General Public License (GPL) does govern
+#   all other use of the material that constitutes the Autoconf Macro.
+#
+#   This special exception to the GPL applies to versions of the Autoconf
+#   Macro released by the Autoconf Archive. When you make and distribute a
+#   modified version of the Autoconf Macro, you may extend this special
+#   exception to the GPL to apply to your modified version as well.
+
+#serial 2
+
+AC_DEFUN([AX_CHECK_COMPILE_FLAG],
+[AC_PREREQ(2.59)dnl for _AC_LANG_PREFIX
+AS_VAR_PUSHDEF([CACHEVAR],[ax_cv_check_[]_AC_LANG_ABBREV[]flags_$4_$1])dnl
+AC_CACHE_CHECK([whether _AC_LANG compiler accepts $1], CACHEVAR, [
+  ax_check_save_flags=$[]_AC_LANG_PREFIX[]FLAGS
+  _AC_LANG_PREFIX[]FLAGS="$[]_AC_LANG_PREFIX[]FLAGS $4 $1"
+  AC_COMPILE_IFELSE([AC_LANG_PROGRAM()],
+    [AS_VAR_SET(CACHEVAR,[yes])],
+    [AS_VAR_SET(CACHEVAR,[no])])
+  _AC_LANG_PREFIX[]FLAGS=$ax_check_save_flags])
+AS_IF([test x"AS_VAR_GET(CACHEVAR)" = xyes],
+  [m4_default([$2], :)],
+  [m4_default([$3], :)])
+AS_VAR_POPDEF([CACHEVAR])dnl
+])dnl AX_CHECK_COMPILE_FLAGS

src/egl/drivers/dri2/platform_wayland.c

@@ -557,6 +557,8 @@ dri2_swap_buffers_with_damage(_EGLDriver *drv,
 {
    struct dri2_egl_display *dri2_dpy = dri2_egl_display(disp);
    struct dri2_egl_surface *dri2_surf = dri2_egl_surface(draw);
+   struct dri2_egl_context *dri2_ctx;
+   _EGLContext *ctx;
    int i, ret = 0;
 
    while (dri2_surf->frame_callback && ret != -1)
@@ -598,8 +600,8 @@ dri2_swap_buffers_with_damage(_EGLDriver *drv,
    dri2_surf->dy = 0;
 
    if (n_rects == 0) {
-      wl_surface_damage(dri2_surf->wl_win->surface, 0, 0,
-                        dri2_surf->base.Width, dri2_surf->base.Height);
+      wl_surface_damage(dri2_surf->wl_win->surface,
+                        0, 0, INT32_MAX, INT32_MAX);
    } else {
       for (i = 0; i < n_rects; i++) {
          const int *rect = &rects[i * 4];
@@ -610,11 +612,22 @@ dri2_swap_buffers_with_damage(_EGLDriver *drv,
       }
    }
 
-   wl_surface_commit(dri2_surf->wl_win->surface);
+   if (dri2_dpy->flush->base.version >= 4) {
+      ctx = _eglGetCurrentContext();
+      dri2_ctx = dri2_egl_context(ctx);
+      (*dri2_dpy->flush->flush_with_flags)(dri2_ctx->dri_context,
+                                           dri2_surf->dri_drawable,
+                                           __DRI2_FLUSH_DRAWABLE,
+                                           __DRI2_THROTTLE_SWAPBUFFER);
+   } else {
+      (*dri2_dpy->flush->flush)(dri2_surf->dri_drawable);
+   }
 
-   (*dri2_dpy->flush->flush)(dri2_surf->dri_drawable);
    (*dri2_dpy->flush->invalidate)(dri2_surf->dri_drawable);
 
+   wl_surface_commit(dri2_surf->wl_win->surface);
+   wl_display_flush(dri2_dpy->wl_dpy);
+
    return EGL_TRUE;
 }
 

src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c

@@ -72,8 +72,8 @@ static void fetch_pipeline_prepare( struct draw_pt_middle_end *middle,
 
    const unsigned gs_out_prim = (gs ? gs->output_primitive :
                                  u_assembled_prim(prim));
-   unsigned nr = MAX2( vs->info.num_inputs,
-		       draw_total_vs_outputs(draw) );
+   unsigned nr_vs_outputs = draw_total_vs_outputs(draw);
+   unsigned nr = MAX2(vs->info.num_inputs, nr_vs_outputs);
 
    if (gs) {
       nr = MAX2(nr, gs->info.num_outputs + 1);
@@ -129,6 +129,9 @@ static void fetch_pipeline_prepare( struct draw_pt_middle_end *middle,
    /* No need to prepare the shader.
     */
    vs->prepare(vs, draw);
+
+   /* Make sure that the vertex size didn't change at any point above */
+   assert(nr_vs_outputs == draw_total_vs_outputs(draw));
 }
 
 

src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline_llvm.c

@@ -141,19 +141,11 @@ llvm_middle_end_prepare( struct draw_pt_middle_end *middle,
    struct draw_geometry_shader *gs = draw->gs.geometry_shader;
    const unsigned out_prim = gs ? gs->output_primitive :
       u_assembled_prim(in_prim);
-   const unsigned nr = MAX2(vs->info.num_inputs,
-                            draw_total_vs_outputs(draw));
+   unsigned nr;
 
    fpme->input_prim = in_prim;
    fpme->opt = opt;
 
-   /* Always leave room for the vertex header whether we need it or
-    * not.  It's hard to get rid of it in particular because of the
-    * viewport code in draw_pt_post_vs.c.
-    */
-   fpme->vertex_size = sizeof(struct vertex_header) + nr * 4 * sizeof(float);
-
-
    draw_pt_post_vs_prepare( fpme->post_vs,
                             draw->clip_xy,
                             draw->clip_z,
@@ -177,6 +169,30 @@ llvm_middle_end_prepare( struct draw_pt_middle_end *middle,
       *max_vertices = 4096;
    }
 
+   /* Get the number of float[4] attributes per vertex.
+    * Note: this must be done after draw_pt_emit_prepare() since that
+    * can effect the vertex size.
+    */
+   nr = MAX2(vs->info.num_inputs, draw_total_vs_outputs(draw));
+
+   /* Always leave room for the vertex header whether we need it or
+    * not.  It's hard to get rid of it in particular because of the
+    * viewport code in draw_pt_post_vs.c.
+    */
+   fpme->vertex_size = sizeof(struct vertex_header) + nr * 4 * sizeof(float);
+
+   /* Get the number of float[4] attributes per vertex.
+    * Note: this must be done after draw_pt_emit_prepare() since that
+    * can effect the vertex size.
+    */
+   nr = MAX2(vs->info.num_inputs, draw_total_vs_outputs(draw));
+
+   /* Always leave room for the vertex header whether we need it or
+    * not.  It's hard to get rid of it in particular because of the
+    * viewport code in draw_pt_post_vs.c.
+    */
+   fpme->vertex_size = sizeof(struct vertex_header) + nr * 4 * sizeof(float);
+
    /* return even number */
    *max_vertices = *max_vertices & ~1;
 

src/gallium/auxiliary/pipe-loader/pipe_loader_sw.c

@@ -95,8 +95,11 @@ pipe_loader_sw_create_screen(struct pipe_loader_device *dev,
       return NULL;
 
    init = (void *)util_dl_get_proc_address(sdev->lib, "swrast_create_screen");
-   if (!init)
+   if (!init){
+      util_dl_close(sdev->lib);
+      sdev->lib = NULL;
       return NULL;
+   }
 
    return init(sdev->ws);
 }

src/gallium/auxiliary/rtasm/rtasm_execmem.c

@@ -69,7 +69,7 @@ static struct mem_block *exec_heap = NULL;
 static unsigned char *exec_mem = NULL;
 
 
-static void
+static int
 init_heap(void)
 {
    if (!exec_heap)
@@ -79,6 +79,8 @@ init_heap(void)
       exec_mem = (unsigned char *) mmap(0, EXEC_HEAP_SIZE, 
 					PROT_EXEC | PROT_READ | PROT_WRITE, 
 					MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+
+   return (exec_mem != MAP_FAILED);
 }
 
 
@@ -90,7 +92,8 @@ rtasm_exec_malloc(size_t size)
 
    pipe_mutex_lock(exec_mutex);
 
-   init_heap();
+   if (!init_heap())
+      goto bail;
 
    if (exec_heap) {
       size = (size + 31) & ~31;  /* next multiple of 32 bytes */
@@ -101,7 +104,8 @@ rtasm_exec_malloc(size_t size)
       addr = exec_mem + block->ofs;
    else 
       debug_printf("rtasm_exec_malloc failed\n");
-   
+
+bail:
    pipe_mutex_unlock(exec_mutex);
    
    return addr;

src/gallium/auxiliary/util/u_format.h

@@ -872,6 +872,9 @@ util_format_get_component_bits(enum pipe_format format,
 static INLINE enum pipe_format
 util_format_srgb(enum pipe_format format)
 {
+   if (util_format_is_srgb(format))
+      return format;
+
    switch (format) {
    case PIPE_FORMAT_L8_UNORM:
       return PIPE_FORMAT_L8_SRGB;

src/gallium/drivers/llvmpipe/lp_state_sampler.c

@@ -136,6 +136,12 @@ llvmpipe_set_sampler_views(struct pipe_context *pipe,
 
    /* set the new sampler views */
    for (i = 0; i < num; i++) {
+      /* Note: we're using pipe_sampler_view_release() here to work around
+       * a possible crash when the old view belongs to another context that
+       * was already destroyed.
+       */
+      pipe_sampler_view_release(pipe,
+                                &llvmpipe->sampler_views[shader][start + i]);
       pipe_sampler_view_reference(&llvmpipe->sampler_views[shader][start + i],
                                   views[i]);
    }

src/gallium/drivers/nouveau/nouveau_buffer.c

@@ -205,6 +205,9 @@ nouveau_transfer_write(struct nouveau_context *nv, struct nouveau_transfer *tx,
                   base, size / 4, (const uint32_t *)data);
    else
       nv->push_data(nv, buf->bo, buf->offset + base, buf->domain, size, data);
+
+   nouveau_fence_ref(nv->screen->fence.current, &buf->fence);
+   nouveau_fence_ref(nv->screen->fence.current, &buf->fence_wr);
 }
 
 

src/gallium/drivers/nouveau/nouveau_fence.c

@@ -189,16 +189,15 @@ nouveau_fence_wait(struct nouveau_fence *fence)
    /* wtf, someone is waiting on a fence in flush_notify handler? */
    assert(fence->state != NOUVEAU_FENCE_STATE_EMITTING);
 
-   if (fence->state < NOUVEAU_FENCE_STATE_EMITTED) {
+   if (fence->state < NOUVEAU_FENCE_STATE_EMITTED)
       nouveau_fence_emit(fence);
 
-      if (fence == screen->fence.current)
-         nouveau_fence_new(screen, &screen->fence.current, FALSE);
-   }
-   if (fence->state < NOUVEAU_FENCE_STATE_FLUSHED) {
+   if (fence->state < NOUVEAU_FENCE_STATE_FLUSHED)
       if (nouveau_pushbuf_kick(screen->pushbuf, screen->pushbuf->channel))
          return FALSE;
-   }
+
+   if (fence == screen->fence.current)
+      nouveau_fence_next(screen);
 
    do {
       nouveau_fence_update(screen, FALSE);

src/gallium/drivers/nouveau/nouveau_vp3_video.c

@@ -362,12 +362,7 @@ nouveau_vp3_screen_get_video_param(struct pipe_screen *pscreen,
    enum pipe_video_format codec = u_reduce_video_profile(profile);
    switch (param) {
    case PIPE_VIDEO_CAP_SUPPORTED:
-      /* For now, h264 and mpeg4 don't work on pre-nvc0. */
-      if (chipset < 0xc0)
-         return codec == PIPE_VIDEO_FORMAT_MPEG12 ||
-            codec == PIPE_VIDEO_FORMAT_VC1;
-      /* In the general case, this should work, once the pre-nvc0 problems are
-       * resolved. */
+      /* VP3 does not support MPEG4, VP4+ do. */
       return profile >= PIPE_VIDEO_PROFILE_MPEG1 && (
             !vp3 || codec != PIPE_VIDEO_FORMAT_MPEG4);
    case PIPE_VIDEO_CAP_NPOT_TEXTURES:

src/gallium/drivers/nouveau/nouveau_vp3_video_vp.c

@@ -114,21 +114,21 @@ struct h264_picparm_vp { // 700..a00
 	uint32_t stride1, stride2; // 04 08
 	uint32_t ofs[6]; // 0c..24 in-image offset
 
-	uint32_t u24; // nfi ac8 ?
+	uint32_t tmp_stride;
 	uint32_t bucket_size; // 28 bucket size
 	uint32_t inter_ring_data_size; // 2c
 
-	unsigned f0 : 1; // 0 0x01: into 640 shifted by 3, 540 shifted by 5, half size something?
-	unsigned f1 : 1; // 1 0x02: into vuc ofs 56
+	unsigned mb_adaptive_frame_field_flag : 1; // 0
+	unsigned direct_8x8_inference_flag : 1; // 1 0x02: into vuc ofs 56
 	unsigned weighted_pred_flag : 1; // 2 0x04
-	unsigned f3 : 1; // 3 0x08: into vuc ofs 68
+	unsigned constrained_intra_pred_flag : 1; // 3 0x08: into vuc ofs 68
 	unsigned is_reference : 1; // 4
 	unsigned interlace : 1; // 5 field_pic_flag
 	unsigned bottom_field_flag : 1; // 6
-	unsigned f7 : 1; // 7 0x80: nfi yet
+	unsigned second_field : 1; // 7 0x80: nfi yet
 
 	signed log2_max_frame_num_minus4 : 4; // 31 0..3
-	unsigned u31_45 : 2; // 31 4..5
+	unsigned chroma_format_idc : 2; // 31 4..5
 	unsigned pic_order_cnt_type : 2; // 31 6..7
 	signed pic_init_qp_minus26 : 6; // 32 0..5
 	signed chroma_qp_index_offset : 5; // 32 6..10
@@ -144,17 +144,15 @@ struct h264_picparm_vp { // 700..a00
 	uint32_t field_order_cnt[2]; // 38, 3c
 
 	struct { // 40
-		// 0x00223102
-		// nfi (needs: top_is_reference, bottom_is_reference, is_long_term, maybe some other state that was saved..
 		unsigned fifo_idx : 7; // 00 0..6
 		unsigned tmp_idx : 5; // 00 7..11
-		unsigned unk12 : 1; // 00 12 not seen yet, but set, maybe top_is_reference
-		unsigned unk13 : 1; // 00 13 not seen yet, but set, maybe bottom_is_reference?
-		unsigned unk14 : 1; // 00 14 skipped?
+		unsigned top_is_reference : 1; // 00 12
+		unsigned bottom_is_reference : 1; // 00 13
+		unsigned is_long_term : 1; // 00 14
 		unsigned notseenyet : 1; // 00 15 pad?
-		unsigned unk16 : 1; // 00 16
-		unsigned unk17 : 4; // 00 17..20
-		unsigned unk21 : 4; // 00 21..24
+		unsigned field_pic_flag : 1; // 00 16
+		unsigned top_field_marking : 4; // 00 17..20
+		unsigned bottom_field_marking : 4; // 00 21..24
 		unsigned pad : 7; // 00 d25..31
 
 		uint32_t field_order_cnt[2]; // 04,08
@@ -335,21 +333,21 @@ nouveau_vp3_fill_picparm_h264_vp(struct nouveau_vp3_decoder *dec,
    nouveau_vp3_ycbcr_offsets(dec, &h->ofs[1], &h->ofs[3], &h->ofs[4]);
    h->ofs[5] = h->ofs[3];
    h->ofs[0] = h->ofs[2] = 0;
-   h->u24 = dec->tmp_stride >> 8;
-   assert(h->u24);
+   h->tmp_stride = dec->tmp_stride >> 8;
+   assert(h->tmp_stride);
    nouveau_vp3_inter_sizes(dec, 1, &ring, &h->bucket_size, &h->inter_ring_data_size);
 
    h->u220 = 0;
-   h->f0 = d->pps->sps->mb_adaptive_frame_field_flag;
-   h->f1 = d->pps->sps->direct_8x8_inference_flag;
+   h->mb_adaptive_frame_field_flag = d->pps->sps->mb_adaptive_frame_field_flag;
+   h->direct_8x8_inference_flag = d->pps->sps->direct_8x8_inference_flag;
    h->weighted_pred_flag = d->pps->weighted_pred_flag;
-   h->f3 = d->pps->constrained_intra_pred_flag;
+   h->constrained_intra_pred_flag = d->pps->constrained_intra_pred_flag;
    h->is_reference = d->is_reference;
    h->interlace = d->field_pic_flag;
    h->bottom_field_flag = d->bottom_field_flag;
-   h->f7 = 0; // TODO: figure out when set..
+   h->second_field = 0; // TODO: figure out when set..
    h->log2_max_frame_num_minus4 = d->pps->sps->log2_max_frame_num_minus4;
-   h->u31_45 = 1;
+   h->chroma_format_idc = 1;
 
    h->pic_order_cnt_type = d->pps->sps->pic_order_cnt_type;
    h->pic_init_qp_minus26 = d->pps->pic_init_qp_minus26;
@@ -376,18 +374,19 @@ nouveau_vp3_fill_picparm_h264_vp(struct nouveau_vp3_decoder *dec,
       h->refs[j].field_order_cnt[1] = d->field_order_cnt_list[i][1];
       h->refs[j].frame_idx = d->frame_num_list[i];
       if (!dec->refs[refs[j]->valid_ref].field_pic_flag) {
-         h->refs[j].unk12 = d->top_is_reference[i];
-         h->refs[j].unk13 = d->bottom_is_reference[i];
+         h->refs[j].top_is_reference = d->top_is_reference[i];
+         h->refs[j].bottom_is_reference = d->bottom_is_reference[i];
       }
-      h->refs[j].unk14 = 0;
+      h->refs[j].is_long_term = d->is_long_term[i];
       h->refs[j].notseenyet = 0;
-      h->refs[j].unk16 = dec->refs[refs[j]->valid_ref].field_pic_flag;
-      h->refs[j].unk17 = dec->refs[refs[j]->valid_ref].decoded_top &&
-                         d->top_is_reference[i];
-      h->refs[j].unk21 = dec->refs[refs[j]->valid_ref].decoded_bottom &&
-                         d->bottom_is_reference[i];
+      h->refs[j].field_pic_flag = dec->refs[refs[j]->valid_ref].field_pic_flag;
+      h->refs[j].top_field_marking =
+         dec->refs[refs[j]->valid_ref].decoded_top && d->top_is_reference[i] ?
+         1 + d->is_long_term[i] : 0;
+      h->refs[j].bottom_field_marking =
+         dec->refs[refs[j]->valid_ref].decoded_bottom && d->bottom_is_reference[i] ?
+         1 + d->is_long_term[i] : 0;
       h->refs[j].pad = 0;
-      assert(!d->is_long_term[i]);
       j++;
    }
    for (; i < 16; ++i)
@@ -395,7 +394,7 @@ nouveau_vp3_fill_picparm_h264_vp(struct nouveau_vp3_decoder *dec,
    assert(d->num_ref_frames <= dec->base.max_references);
 
    for (; i < d->num_ref_frames; ++i)
-      h->refs[j].unk16 = d->field_pic_flag;
+      h->refs[j].field_pic_flag = d->field_pic_flag;
    *(struct h264_picparm_vp *)map = *h;
 
    return 0x1113;

src/gallium/drivers/nouveau/nv50/nv50_context.c

@@ -114,6 +114,8 @@ nv50_destroy(struct pipe_context *pipe)
    draw_destroy(nv50->draw);
 #endif
 
+   FREE(nv50->blit);
+
    nouveau_context_destroy(&nv50->base);
 }
 
@@ -163,7 +165,7 @@ nv50_invalidate_resource_storage(struct nouveau_context *ctx,
    }
 
    if (res->bind & PIPE_BIND_SAMPLER_VIEW) {
-      for (s = 0; s < 5; ++s) {
+      for (s = 0; s < 3; ++s) {
       for (i = 0; i < nv50->num_textures[s]; ++i) {
          if (nv50->textures[s][i] &&
              nv50->textures[s][i]->texture == res) {
@@ -177,7 +179,7 @@ nv50_invalidate_resource_storage(struct nouveau_context *ctx,
    }
 
    if (res->bind & PIPE_BIND_CONSTANT_BUFFER) {
-      for (s = 0; s < 5; ++s) {
+      for (s = 0; s < 3; ++s) {
       for (i = 0; i < nv50->num_vtxbufs; ++i) {
          if (!nv50->constbuf[s][i].user &&
              nv50->constbuf[s][i].u.buf == res) {

src/gallium/drivers/nouveau/nv50/nv50_screen.c

@@ -224,7 +224,7 @@ nv50_screen_get_shader_param(struct pipe_screen *pscreen, unsigned shader,
    case PIPE_SHADER_CAP_MAX_INPUTS:
       if (shader == PIPE_SHADER_VERTEX)
          return 32;
-      return 0x300 / 16;
+      return 15;
    case PIPE_SHADER_CAP_MAX_CONSTS:
       return 65536 / 16;
    case PIPE_SHADER_CAP_MAX_CONST_BUFFERS:

src/gallium/drivers/nouveau/nv50/nv50_state_validate.c

@@ -61,7 +61,7 @@ nv50_validate_fb(struct nv50_context *nv50)
       if (mt->base.status & NOUVEAU_BUFFER_STATUS_GPU_READING)
          nv50->state.rt_serialize = TRUE;
       mt->base.status |= NOUVEAU_BUFFER_STATUS_GPU_WRITING;
-      mt->base.status &= NOUVEAU_BUFFER_STATUS_GPU_READING;
+      mt->base.status &= ~NOUVEAU_BUFFER_STATUS_GPU_READING;
 
       /* only register for writing, otherwise we'd always serialize here */
       BCTX_REFN(nv50->bufctx_3d, FB, &mt->base, WR);
@@ -91,7 +91,7 @@ nv50_validate_fb(struct nv50_context *nv50)
       if (mt->base.status & NOUVEAU_BUFFER_STATUS_GPU_READING)
          nv50->state.rt_serialize = TRUE;
       mt->base.status |= NOUVEAU_BUFFER_STATUS_GPU_WRITING;
-      mt->base.status &= NOUVEAU_BUFFER_STATUS_GPU_READING;
+      mt->base.status &= ~NOUVEAU_BUFFER_STATUS_GPU_READING;
 
       BCTX_REFN(nv50->bufctx_3d, FB, &mt->base, WR);
    } else {

src/gallium/drivers/nouveau/nv50/nv50_tex.c

@@ -271,7 +271,7 @@ nv50_validate_tic(struct nv50_context *nv50, int s)
 
       nv50->screen->tic.lock[tic->id / 32] |= 1 << (tic->id % 32);
 
-      res->status &= NOUVEAU_BUFFER_STATUS_GPU_WRITING;
+      res->status &= ~NOUVEAU_BUFFER_STATUS_GPU_WRITING;
       res->status |= NOUVEAU_BUFFER_STATUS_GPU_READING;
 
       BCTX_REFN(nv50->bufctx_3d, TEXTURES, res, RD);

src/gallium/drivers/nouveau/nv50/nv50_vbo.c

@@ -597,6 +597,15 @@ nv50_draw_elements(struct nv50_context *nv50, boolean shorten,
 
       assert(nouveau_resource_mapped_by_gpu(nv50->idxbuf.buffer));
 
+      /* This shouldn't have to be here. The going theory is that the buffer
+       * is being filled in by PGRAPH, and it's not done yet by the time it
+       * gets submitted to PFIFO, which in turn starts immediately prefetching
+       * the not-yet-written data. Ideally this wait would only happen on
+       * pushbuf submit, but it's probably not a big performance difference.
+       */
+      if (buf->fence_wr && !nouveau_fence_signalled(buf->fence_wr))
+         nouveau_fence_wait(buf->fence_wr);
+
       while (instance_count--) {
          BEGIN_NV04(push, NV50_3D(VERTEX_BEGIN_GL), 1);
          PUSH_DATA (push, prim);

src/gallium/drivers/nouveau/nv50/nv98_video.c

@@ -64,10 +64,6 @@ nv98_create_decoder(struct pipe_context *context,
    struct nouveau_vp3_decoder *dec;
    struct nouveau_pushbuf **push;
    struct nv04_fifo nv04_data = {.vram = 0xbeef0201, .gart = 0xbeef0202};
-   union nouveau_bo_config cfg;
-
-   cfg.nv50.tile_mode = 0x20;
-   cfg.nv50.memtype = 0x70;
 
    int ret, i;
    uint32_t codec = 1, ppp_codec = 3;
@@ -200,7 +196,7 @@ nv98_create_decoder(struct pipe_context *context,
    dec->ref_stride = mb(templ->width)*16 * (mb_half(templ->height)*32 + nouveau_vp3_video_align(templ->height)/2);
    ret = nouveau_bo_new(screen->device, NOUVEAU_BO_VRAM, 0,
                         dec->ref_stride * (templ->max_references+2) + tmp_size,
-                        &cfg, &dec->ref_bo);
+                        NULL, &dec->ref_bo);
    if (ret)
       goto fail;
 

src/gallium/drivers/r300/compiler/tests/radeon_compiler_regalloc_tests.c

@@ -79,14 +79,13 @@ static void test_runner_rc_regalloc(
 
 static void tex_1d_swizzle(struct test_result *result)
 {
-	struct radeon_compiler c;
+	struct r300_fragment_program_compiler c;
 
-	init_compiler(&c, RC_FRAGMENT_PROGRAM, 0, 0);
-	struct r300_fragment_program_compiler *cc =
-		(struct r300_fragment_program_compiler*)&c;
-	cc->AllocateHwInputs = dummy_allocate_hw_inputs;
+	memset(&c, 0, sizeof(c));
+	init_compiler(&c.Base, RC_FRAGMENT_PROGRAM, 0, 0);
+	c.AllocateHwInputs = dummy_allocate_hw_inputs;
 
-	test_runner_rc_regalloc(result, &c, "regalloc_tex_1d_swizzle.test");
+	test_runner_rc_regalloc(result, &c.Base, "regalloc_tex_1d_swizzle.test");
 }
 
 unsigned radeon_compiler_regalloc_run_tests()

src/gallium/drivers/r300/compiler/tests/rc_test_helpers.c

@@ -542,6 +542,7 @@ unsigned load_program(
 	char **string_store;
 	unsigned i = 0;
 
+	memset(line, 0, sizeof(line));
 	snprintf(path, MAX_PATH_LENGTH, "compiler/tests/%s", filename);
 	file = fopen(path, "r");
 	if (!file) {
@@ -552,7 +553,8 @@ unsigned load_program(
 	count = &test->num_input_lines;
 
 	while (fgets(line, MAX_LINE_LENGTH, file)){
-		if (line[MAX_LINE_LENGTH - 2] == '\n') {
+		char last_char = line[MAX_LINE_LENGTH - 1];
+		if (last_char && last_char != '\n') {
 			fprintf(stderr, "Error line cannot be longer than 100 "
 				"characters:\n%s\n", line);
 			return 0;

src/gallium/drivers/r600/evergreen_compute.c

@@ -204,6 +204,8 @@ void *evergreen_create_compute_state(
 	const unsigned char * code;
 	unsigned i;
 
+	shader->llvm_ctx = LLVMContextCreate();
+
 	COMPUTE_DBG(ctx->screen, "*** evergreen_create_compute_state\n");
 
 	header = cso->prog;
@@ -216,13 +218,14 @@ void *evergreen_create_compute_state(
 	shader->input_size = cso->req_input_mem;
 
 #ifdef HAVE_OPENCL 
-	shader->num_kernels = radeon_llvm_get_num_kernels(code, header->num_bytes);
+	shader->num_kernels = radeon_llvm_get_num_kernels(shader->llvm_ctx, code,
+							header->num_bytes);
 	shader->kernels = CALLOC(sizeof(struct r600_kernel), shader->num_kernels);
 
 	for (i = 0; i < shader->num_kernels; i++) {
 		struct r600_kernel *kernel = &shader->kernels[i];
-		kernel->llvm_module = radeon_llvm_get_kernel_module(i, code,
-							header->num_bytes);
+		kernel->llvm_module = radeon_llvm_get_kernel_module(shader->llvm_ctx, i,
+							code, header->num_bytes);
 	}
 #endif
 	return shader;
@@ -232,7 +235,18 @@ void evergreen_delete_compute_state(struct pipe_context *ctx, void* state)
 {
 	struct r600_pipe_compute *shader = (struct r600_pipe_compute *)state;
 
-	free(shader);
+	if (!shader)
+		return;
+
+	FREE(shader->kernels);
+
+#ifdef HAVE_OPENCL
+	if (shader->llvm_ctx){
+		LLVMContextDispose(shader->llvm_ctx);
+	}
+#endif
+
+	FREE(shader);
 }
 
 static void evergreen_bind_compute_state(struct pipe_context *ctx_, void *state)
@@ -474,6 +488,17 @@ static void compute_emit_cs(struct r600_context *ctx, const uint *block_layout,
 	r600_flush_emit(ctx);
 	ctx->b.flags = 0;
 
+	if (ctx->b.chip_class >= CAYMAN) {
+		cs->buf[cs->cdw++] = PKT3(PKT3_EVENT_WRITE, 0, 0);
+		cs->buf[cs->cdw++] = EVENT_TYPE(EVENT_TYPE_CS_PARTIAL_FLUSH) | EVENT_INDEX(4);
+		/* DEALLOC_STATE prevents the GPU from hanging when a
+		 * SURFACE_SYNC packet is emitted some time after a DISPATCH_DIRECT
+		 * with any of the CB*_DEST_BASE_ENA or DB_DEST_BASE_ENA bits set.
+		 */
+		cs->buf[cs->cdw++] = PKT3C(PKT3_DEALLOC_STATE, 0, 0);
+		cs->buf[cs->cdw++] = 0;
+	}
+
 #if 0
 	COMPUTE_DBG(ctx->screen, "cdw: %i\n", cs->cdw);
 	for (i = 0; i < cs->cdw; i++) {

src/gallium/drivers/r600/evergreen_compute_internal.h

@@ -47,6 +47,10 @@ struct r600_pipe_compute {
 	unsigned private_size;
 	unsigned input_size;
 	struct r600_resource *kernel_param;
+
+#ifdef HAVE_OPENCL
+	LLVMContextRef llvm_ctx;
+#endif
 };
 
 struct r600_resource* r600_compute_buffer_alloc_vram(struct r600_screen *screen, unsigned size);

src/gallium/drivers/r600/evergreend.h

@@ -63,6 +63,7 @@
 #define R600_TEXEL_PITCH_ALIGNMENT_MASK        0x7
 
 #define PKT3_NOP                               0x10
+#define PKT3_DEALLOC_STATE                     0x14
 #define PKT3_DISPATCH_DIRECT                   0x15
 #define PKT3_DISPATCH_INDIRECT                 0x16
 #define PKT3_INDIRECT_BUFFER_END               0x17

src/gallium/drivers/r600/r600_blit.c

@@ -619,6 +619,36 @@ void r600_copy_buffer(struct pipe_context *ctx, struct pipe_resource *dst, unsig
 	}
 }
 
+/**
+ * Global buffers are not really resources, they are are actually offsets
+ * into a single global resource (r600_screen::global_pool).  The means
+ * they don't have their own cs_buf handle, so they cannot be passed
+ * to r600_copy_buffer() and must be handled separately.
+ *
+ * XXX: It should be possible to implement this function using
+ * r600_copy_buffer() by passing the memory_pool resource as both src
+ * and dst and updating dstx and src_box to point to the correct offsets.
+ * This would likely perform better than the current implementation.
+ */
+static void r600_copy_global_buffer(struct pipe_context *ctx,
+				    struct pipe_resource *dst, unsigned
+				    dstx, struct pipe_resource *src,
+				    const struct pipe_box *src_box)
+{
+	struct pipe_box dst_box; struct pipe_transfer *src_pxfer,
+	*dst_pxfer;
+
+	u_box_1d(dstx, src_box->width, &dst_box);
+	void *src_ptr = ctx->transfer_map(ctx, src, 0, PIPE_TRANSFER_READ,
+					  src_box, &src_pxfer);
+	void *dst_ptr = ctx->transfer_map(ctx, dst, 0, PIPE_TRANSFER_WRITE,
+					  &dst_box, &dst_pxfer);
+	memcpy(dst_ptr, src_ptr, src_box->width);
+
+	ctx->transfer_unmap(ctx, src_pxfer);
+	ctx->transfer_unmap(ctx, dst_pxfer);
+}
+
 static void r600_clear_buffer(struct pipe_context *ctx, struct pipe_resource *dst,
 			      unsigned offset, unsigned size, unsigned value)
 {
@@ -671,7 +701,12 @@ static void r600_resource_copy_region(struct pipe_context *ctx,
 
 	/* Handle buffers first. */
 	if (dst->target == PIPE_BUFFER && src->target == PIPE_BUFFER) {
-		r600_copy_buffer(ctx, dst, dstx, src, src_box);
+		if ((src->bind & PIPE_BIND_GLOBAL) ||
+					(dst->bind & PIPE_BIND_GLOBAL)) {
+			r600_copy_global_buffer(ctx, dst, dstx, src, src_box);
+		} else {
+			r600_copy_buffer(ctx, dst, dstx, src, src_box);
+		}
 		return;
 	}
 

src/gallium/drivers/r600/r600_pipe.c

@@ -192,12 +192,7 @@ static void r600_destroy_context(struct pipe_context *context)
 
 	r600_release_command_buffer(&rctx->start_cs_cmd);
 
-	if (rctx->b.rings.gfx.cs) {
-		rctx->b.ws->cs_destroy(rctx->b.rings.gfx.cs);
-	}
-	if (rctx->b.rings.dma.cs) {
-		rctx->b.ws->cs_destroy(rctx->b.rings.dma.cs);
-	}
+	FREE(rctx->start_compute_cs_cmd.buf);
 
 	r600_common_context_cleanup(&rctx->b);
 	FREE(rctx);

src/gallium/drivers/r600/sb/sb_bc.h

@@ -614,6 +614,10 @@ public:
 	unsigned num_slots;
 	bool uses_mova_gpr;
 
+	bool stack_workaround_8xx;
+	bool stack_workaround_9xx;
+
+	unsigned wavefront_size;
 	unsigned stack_entry_size;
 
 	static unsigned dump_pass;
@@ -638,6 +642,23 @@ public:
 	bool is_cayman() {return hw_class == HW_CLASS_CAYMAN;}
 	bool is_egcm() {return hw_class >= HW_CLASS_EVERGREEN;}
 
+	bool needs_8xx_stack_workaround() {
+		if (!is_evergreen())
+			return false;
+
+		switch (hw_chip) {
+		case HW_CHIP_CYPRESS:
+		case HW_CHIP_JUNIPER:
+			return false;
+		default:
+			return true;
+		}
+	}
+
+	bool needs_9xx_stack_workaround() {
+		return is_cayman();
+	}
+
 	sb_hw_class_bits hw_class_bit() {
 		switch (hw_class) {
 		case HW_CLASS_R600:return HB_R6;

src/gallium/drivers/r600/sb/sb_bc_finalize.cpp

@@ -40,8 +40,9 @@ namespace r600_sb {
 
 int bc_finalizer::run() {
 
-	regions_vec &rv = sh.get_regions();
+	run_on(sh.root);
 
+	regions_vec &rv = sh.get_regions();
 	for (regions_vec::reverse_iterator I = rv.rbegin(), E = rv.rend(); I != E;
 			++I) {
 		region_node *r = *I;
@@ -58,8 +59,6 @@ int bc_finalizer::run() {
 		r->expand();
 	}
 
-	run_on(sh.root);
-
 	cf_peephole();
 
 	// workaround for some problems on r6xx/7xx
@@ -213,18 +212,36 @@ void bc_finalizer::run_on(container_node* c) {
 		if (n->is_alu_group()) {
 			finalize_alu_group(static_cast<alu_group_node*>(n));
 		} else {
-			if (n->is_fetch_inst()) {
+			if (n->is_alu_clause()) {
+				cf_node *c = static_cast<cf_node*>(n);
+
+				if (c->bc.op == CF_OP_ALU_PUSH_BEFORE && ctx.is_egcm()) {
+					if (ctx.stack_workaround_8xx) {
+						region_node *r = c->get_parent_region();
+						if (r) {
+							unsigned ifs, loops;
+							unsigned elems = get_stack_depth(r, loops, ifs);
+							unsigned dmod1 = elems % ctx.stack_entry_size;
+							unsigned dmod2 = (elems + 1) % ctx.stack_entry_size;
+
+							if (elems && (!dmod1 || !dmod2))
+								c->flags |= NF_ALU_STACK_WORKAROUND;
+						}
+					} else if (ctx.stack_workaround_9xx) {
+						region_node *r = c->get_parent_region();
+						if (r) {
+							unsigned ifs, loops;
+							get_stack_depth(r, loops, ifs);
+							if (loops >= 2)
+								c->flags |= NF_ALU_STACK_WORKAROUND;
+						}
+					}
+				}
+			} else if (n->is_fetch_inst()) {
 				finalize_fetch(static_cast<fetch_node*>(n));
 			} else if (n->is_cf_inst()) {
 				finalize_cf(static_cast<cf_node*>(n));
-			} else if (n->is_alu_clause()) {
-
-			} else if (n->is_fetch_clause()) {
-
-			} else {
-				assert(!"unexpected node");
 			}
-
 			if (n->is_container())
 				run_on(static_cast<container_node*>(n));
 		}
@@ -578,10 +595,6 @@ void bc_finalizer::finalize_cf(cf_node* c) {
 
 	unsigned flags = c->bc.op_ptr->flags;
 
-	if (flags & CF_CALL) {
-		update_nstack(c->get_parent_region(), ctx.is_cayman() ? 1 : 2);
-	}
-
 	c->bc.end_of_program = 0;
 	last_cf = c;
 
@@ -715,17 +728,8 @@ void bc_finalizer::finalize_cf(cf_node* c) {
 
 			c->bc.index_gpr = reg >= 0 ? reg : 0;
 		}
-
-
-
-	} else {
-
-#if 0
-		if ((flags & (CF_BRANCH | CF_LOOP)) && !sh.uses_gradients) {
-			c->bc.valid_pixel_mode = 1;
-		}
-#endif
-
+	} else if (flags & CF_CALL) {
+		update_nstack(c->get_parent_region(), ctx.wavefront_size == 16 ? 2 : 1);
 	}
 }
 
@@ -763,37 +767,86 @@ void bc_finalizer::update_ngpr(unsigned gpr) {
 		ngpr = gpr + 1;
 }
 
+unsigned bc_finalizer::get_stack_depth(node *n, unsigned &loops,
+                                           unsigned &ifs, unsigned add) {
+	unsigned stack_elements = add;
+	bool has_non_wqm_push = (add != 0);
+	region_node *r = n->is_region() ?
+			static_cast<region_node*>(n) : n->get_parent_region();
+
+	loops = 0;
+	ifs = 0;
+
+	while (r) {
+		if (r->is_loop()) {
+			++loops;
+		} else {
+			++ifs;
+			has_non_wqm_push = true;
+		}
+		r = r->get_parent_region();
+	}
+	stack_elements += (loops * ctx.stack_entry_size) + ifs;
+
+	// reserve additional elements in some cases
+	switch (ctx.hw_class) {
+	case HW_CLASS_R600:
+	case HW_CLASS_R700:
+		// If any non-WQM push is invoked, 2 elements should be reserved.
+		if (has_non_wqm_push)
+			stack_elements += 2;
+		break;
+	case HW_CLASS_CAYMAN:
+		// If any stack operation is invoked, 2 elements should be reserved
+		if (stack_elements)
+			stack_elements += 2;
+		break;
+	case HW_CLASS_EVERGREEN:
+		// According to the docs we need to reserve 1 element for each of the
+		// following cases:
+		//   1) non-WQM push is used with WQM/LOOP frames on stack
+		//   2) ALU_ELSE_AFTER is used at the point of max stack usage
+		// NOTE:
+		// It was found that the conditions above are not sufficient, there are
+		// other cases where we also need to reserve stack space, that's why
+		// we always reserve 1 stack element if we have non-WQM push on stack.
+		// Condition 2 is ignored for now because we don't use this instruction.
+		if (has_non_wqm_push)
+			++stack_elements;
+		break;
+	}
+	return stack_elements;
+}
+
 void bc_finalizer::update_nstack(region_node* r, unsigned add) {
 	unsigned loops = 0;
 	unsigned ifs = 0;
+	unsigned elems = r ? get_stack_depth(r, loops, ifs, add) : add;
 
-	while (r) {
-		if (r->is_loop())
-			++loops;
-		else
-			++ifs;
-
-		r = r->get_parent_region();
-	}
-
-	unsigned stack_elements = (loops * ctx.stack_entry_size) + ifs + add;
-
-	// FIXME calculate more precisely
-	if (ctx.is_evergreen()) {
-		++stack_elements;
-	} else {
-		stack_elements += 2;
-		if (ctx.is_cayman())
-			++stack_elements;
-	}
-
-	unsigned stack_entries = (stack_elements + 3) >> 2;
+	// XXX all chips expect this value to be computed using 4 as entry size,
+	// not the real entry size
+	unsigned stack_entries = (elems + 3) >> 2;
 
 	if (nstack < stack_entries)
 		nstack = stack_entries;
 }
 
 void bc_finalizer::cf_peephole() {
+	if (ctx.stack_workaround_8xx || ctx.stack_workaround_9xx) {
+		for (node_iterator N, I = sh.root->begin(), E = sh.root->end(); I != E;
+				I = N) {
+			N = I; ++N;
+			cf_node *c = static_cast<cf_node*>(*I);
+
+			if (c->bc.op == CF_OP_ALU_PUSH_BEFORE &&
+					(c->flags & NF_ALU_STACK_WORKAROUND)) {
+				cf_node *push = sh.create_cf(CF_OP_PUSH);
+				c->insert_before(push);
+				push->jump(c);
+				c->bc.set_op(CF_OP_ALU);
+			}
+		}
+	}
 
 	for (node_iterator N, I = sh.root->begin(), E = sh.root->end(); I != E;
 			I = N) {

src/gallium/drivers/r600/sb/sb_context.cpp

@@ -66,20 +66,27 @@ int sb_context::init(r600_isa *isa, sb_hw_chip chip, sb_hw_class cclass) {
 	case HW_CHIP_RS780:
 	case HW_CHIP_RV620:
 	case HW_CHIP_RS880:
-
+		wavefront_size = 16;
+		stack_entry_size = 8;
+		break;
 	case HW_CHIP_RV630:
 	case HW_CHIP_RV635:
 	case HW_CHIP_RV730:
 	case HW_CHIP_RV710:
 	case HW_CHIP_PALM:
 	case HW_CHIP_CEDAR:
+		wavefront_size = 32;
 		stack_entry_size = 8;
 		break;
 	default:
+		wavefront_size = 64;
 		stack_entry_size = 4;
 		break;
 	}
 
+	stack_workaround_8xx = needs_8xx_stack_workaround();
+	stack_workaround_9xx = needs_9xx_stack_workaround();
+
 	return 0;
 }
 

src/gallium/drivers/r600/sb/sb_ir.h

@@ -700,7 +700,10 @@ enum node_flags {
 	NF_DONT_MOVE = (1 << 8),
 
 	// for KILLxx - we want to schedule them as early as possible
-	NF_SCHEDULE_EARLY = (1 << 9)
+	NF_SCHEDULE_EARLY = (1 << 9),
+
+	// for ALU_PUSH_BEFORE - when set, replace with PUSH + ALU
+	NF_ALU_STACK_WORKAROUND = (1 << 10)
 };
 
 inline node_flags operator |(node_flags l, node_flags r) {

src/gallium/drivers/r600/sb/sb_pass.h

@@ -708,6 +708,9 @@ public:
 	void update_ngpr(unsigned gpr);
 	void update_nstack(region_node *r, unsigned add = 0);
 
+	unsigned get_stack_depth(node *n, unsigned &loops, unsigned &ifs,
+	                         unsigned add = 0);
+
 	void cf_peephole();
 
 };

src/gallium/drivers/radeon/r600_pipe_common.c

@@ -234,6 +234,13 @@ bool r600_common_context_init(struct r600_common_context *rctx,
 
 void r600_common_context_cleanup(struct r600_common_context *rctx)
 {
+	if (rctx->rings.gfx.cs) {
+		rctx->ws->cs_destroy(rctx->rings.gfx.cs);
+	}
+	if (rctx->rings.dma.cs) {
+		rctx->ws->cs_destroy(rctx->rings.dma.cs);
+	}
+
 	if (rctx->allocator_so_filled_size) {
 		u_suballocator_destroy(rctx->allocator_so_filled_size);
 	}

src/gallium/drivers/radeon/radeon_llvm_util.c

@@ -30,13 +30,13 @@
 #include <llvm-c/BitReader.h>
 #include <llvm-c/Core.h>
 #include <llvm-c/Target.h>
+#include <llvm-c/Transforms/IPO.h>
 #include <llvm-c/Transforms/PassManagerBuilder.h>
 
-LLVMModuleRef radeon_llvm_parse_bitcode(const unsigned char * bitcode,
-							unsigned bitcode_len)
+LLVMModuleRef radeon_llvm_parse_bitcode(LLVMContextRef ctx,
+							const unsigned char * bitcode, unsigned bitcode_len)
 {
 	LLVMMemoryBufferRef buf;
-	LLVMContextRef ctx = LLVMContextCreate();
 	LLVMModuleRef module;
 
 	buf = LLVMCreateMemoryBufferWithMemoryRangeCopy((const char*)bitcode,
@@ -46,10 +46,10 @@ LLVMModuleRef radeon_llvm_parse_bitcode(const unsigned char * bitcode,
 	return module;
 }
 
-unsigned radeon_llvm_get_num_kernels(const unsigned char *bitcode,
-				unsigned bitcode_len)
+unsigned radeon_llvm_get_num_kernels(LLVMContextRef ctx,
+				const unsigned char *bitcode, unsigned bitcode_len)
 {
-	LLVMModuleRef mod = radeon_llvm_parse_bitcode(bitcode, bitcode_len);
+	LLVMModuleRef mod = radeon_llvm_parse_bitcode(ctx, bitcode, bitcode_len);
 	return LLVMGetNamedMetadataNumOperands(mod, "opencl.kernels");
 }
 
@@ -59,17 +59,35 @@ static void radeon_llvm_optimize(LLVMModuleRef mod)
 	LLVMTargetDataRef TD = LLVMCreateTargetData(data_layout);
 	LLVMPassManagerBuilderRef builder = LLVMPassManagerBuilderCreate();
 	LLVMPassManagerRef pass_manager = LLVMCreatePassManager();
-	LLVMAddTargetData(TD, pass_manager);
 
-	LLVMPassManagerBuilderUseInlinerWithThreshold(builder, 1000000000);
+	/* Functions calls are not supported yet, so we need to inline
+	 * everything.  The most efficient way to do this is to add
+	 * the always_inline attribute to all non-kernel functions
+	 * and then run the Always Inline pass.  The Always Inline
+	 * pass will automaically inline functions with this attribute
+	 * and does not perform the expensive cost analysis that the normal
+	 * inliner does.
+	 */
+
+	LLVMValueRef fn;
+	for (fn = LLVMGetFirstFunction(mod); fn; fn = LLVMGetNextFunction(fn)) {
+		/* All the non-kernel functions have internal linkage */
+		if (LLVMGetLinkage(fn) == LLVMInternalLinkage) {
+			LLVMAddFunctionAttr(fn, LLVMAlwaysInlineAttribute);
+		}
+	}
+
+	LLVMAddTargetData(TD, pass_manager);
+	LLVMAddAlwaysInlinerPass(pass_manager);
 	LLVMPassManagerBuilderPopulateModulePassManager(builder, pass_manager);
 
 	LLVMRunPassManager(pass_manager, mod);
 	LLVMPassManagerBuilderDispose(builder);
 	LLVMDisposePassManager(pass_manager);
+	LLVMDisposeTargetData(TD);
 }
 
-LLVMModuleRef radeon_llvm_get_kernel_module(unsigned index,
+LLVMModuleRef radeon_llvm_get_kernel_module(LLVMContextRef ctx, unsigned index,
 		const unsigned char *bitcode, unsigned bitcode_len)
 {
 	LLVMModuleRef mod;
@@ -77,7 +95,7 @@ LLVMModuleRef radeon_llvm_get_kernel_module(unsigned index,
 	LLVMValueRef *kernel_metadata;
 	unsigned i;
 
-	mod = radeon_llvm_parse_bitcode(bitcode, bitcode_len);
+	mod = radeon_llvm_parse_bitcode(ctx, bitcode, bitcode_len);
 	num_kernels = LLVMGetNamedMetadataNumOperands(mod, "opencl.kernels");
 	kernel_metadata = MALLOC(num_kernels * sizeof(LLVMValueRef));
 	LLVMGetNamedMetadataOperands(mod, "opencl.kernels", kernel_metadata);

src/gallium/drivers/radeon/radeon_llvm_util.h

@@ -29,10 +29,11 @@
 
 #include <llvm-c/Core.h>
 
-LLVMModuleRef radeon_llvm_parse_bitcode(const unsigned char * bitcode,
-							unsigned bitcode_len);
-unsigned radeon_llvm_get_num_kernels(const unsigned char *bitcode, unsigned bitcode_len);
-LLVMModuleRef radeon_llvm_get_kernel_module(unsigned index,
+LLVMModuleRef radeon_llvm_parse_bitcode(LLVMContextRef ctx,
+			const unsigned char * bitcode, unsigned bitcode_len);
+unsigned radeon_llvm_get_num_kernels(LLVMContextRef ctx,
+			const unsigned char *bitcode, unsigned bitcode_len);
+LLVMModuleRef radeon_llvm_get_kernel_module(LLVMContextRef ctx, unsigned index,
 			const unsigned char *bitcode, unsigned bitcode_len);
 
 #endif

src/gallium/drivers/radeonsi/radeonsi_compute.c

@@ -20,6 +20,7 @@ struct si_pipe_compute {
 
         struct pipe_resource *global_buffers[MAX_GLOBAL_BUFFERS];
 
+	LLVMContextRef llvm_ctx;
 };
 
 static void *radeonsi_create_compute_state(
@@ -33,6 +34,8 @@ static void *radeonsi_create_compute_state(
 	const unsigned char *code;
 	unsigned i;
 
+	program->llvm_ctx = LLVMContextCreate();
+
 	header = cso->prog;
 	code = cso->prog + sizeof(struct pipe_llvm_program_header);
 
@@ -41,13 +44,13 @@ static void *radeonsi_create_compute_state(
 	program->private_size = cso->req_private_mem;
 	program->input_size = cso->req_input_mem;
 
-	program->num_kernels = radeon_llvm_get_num_kernels(code,
+	program->num_kernels = radeon_llvm_get_num_kernels(program->llvm_ctx, code,
 							header->num_bytes);
 	program->kernels = CALLOC(sizeof(struct si_pipe_shader),
 							program->num_kernels);
 	for (i = 0; i < program->num_kernels; i++) {
-		LLVMModuleRef mod = radeon_llvm_get_kernel_module(i, code,
-							header->num_bytes);
+		LLVMModuleRef mod = radeon_llvm_get_kernel_module(program->llvm_ctx, i,
+							code, header->num_bytes);
 		si_compile_llvm(rctx, &program->kernels[i], mod);
 		LLVMDisposeModule(mod);
 	}
@@ -103,6 +106,7 @@ static void radeonsi_launch_grid(
 	unsigned arg_user_sgpr_count = 2;
 	unsigned i;
 	struct si_pipe_shader *shader = &program->kernels[pc];
+	unsigned lds_blocks;
 
 	pm4->compute_pkt = true;
 	si_cmd_context_control(pm4);
@@ -194,6 +198,20 @@ static void radeonsi_launch_grid(
 		                        shader->num_sgprs)) - 1) / 8))
 		;
 
+	lds_blocks = shader->lds_size;
+	/* XXX: We are over allocating LDS.  For SI, the shader reports LDS in
+	 * blocks of 256 bytes, so if there are 4 bytes lds allocated in
+	 * the shader and 4 bytes allocated by the state tracker, then
+	 * we will set LDS_SIZE to 512 bytes rather than 256.
+	 */
+	if (rctx->b.chip_class <= SI) {
+		lds_blocks += align(program->local_size, 256) >> 8;
+	} else {
+		lds_blocks += align(program->local_size, 512) >> 9;
+	}
+
+	assert(lds_blocks <= 0xFF);
+
 	si_pm4_set_reg(pm4, R_00B84C_COMPUTE_PGM_RSRC2,
 		S_00B84C_SCRATCH_EN(0)
 		| S_00B84C_USER_SGPR(arg_user_sgpr_count)
@@ -202,7 +220,7 @@ static void radeonsi_launch_grid(
 		| S_00B84C_TGID_Z_EN(1)
 		| S_00B84C_TG_SIZE_EN(1)
 		| S_00B84C_TIDIG_COMP_CNT(2)
-		| S_00B84C_LDS_SIZE(shader->lds_size)
+		| S_00B84C_LDS_SIZE(lds_blocks)
 		| S_00B84C_EXCP_EN(0))
 		;
 	si_pm4_set_reg(pm4, R_00B854_COMPUTE_RESOURCE_LIMITS, 0);
@@ -257,6 +275,10 @@ static void si_delete_compute_state(struct pipe_context *ctx, void* state){
 		FREE(program->kernels);
 	}
 
+	if (program->llvm_ctx){
+		LLVMContextDispose(program->llvm_ctx);
+	}
+
 	//And then free the program itself.
 	FREE(program);
 }

src/gallium/drivers/softpipe/sp_quad_blend.c

@@ -860,8 +860,8 @@ clamp_colors(float (*quadColor)[4])
 {
    unsigned i, j;
 
-   for (j = 0; j < TGSI_QUAD_SIZE; j++) {
-      for (i = 0; i < 4; i++) {
+   for (i = 0; i < 4; i++) {
+      for (j = 0; j < TGSI_QUAD_SIZE; j++) {
          quadColor[i][j] = CLAMP(quadColor[i][j], 0.0F, 1.0F);
       }
    }

src/gallium/state_trackers/clover/api/program.cpp

@@ -173,11 +173,15 @@ clGetProgramInfo(cl_program d_prog, cl_program_info param,
       break;
 
    case CL_PROGRAM_NUM_DEVICES:
-      buf.as_scalar<cl_uint>() = prog.devices().size();
+      buf.as_scalar<cl_uint>() = prog.devices().size() ?
+                                 prog.devices().size() :
+                                 prog.ctx.devs().size();
       break;
 
    case CL_PROGRAM_DEVICES:
-      buf.as_vector<cl_device_id>() = descs(prog.devices());
+      buf.as_vector<cl_device_id>() = prog.devices().size() ?
+                                      descs(prog.devices()) :
+                                      descs(prog.ctx.devs());
       break;
 
    case CL_PROGRAM_SOURCE:

src/gallium/state_trackers/clover/llvm/invocation.cpp

@@ -117,7 +117,6 @@ namespace {
            const std::string &opts, clang::LangAS::Map& address_spaces) {
 
       clang::CompilerInstance c;
-      clang::CompilerInvocation invocation;
       clang::EmitLLVMOnlyAction act(&llvm::getGlobalContext());
       std::string log;
       llvm::raw_string_ostream s_log(log);

src/gallium/state_trackers/dri/common/dri_screen.c

@@ -360,6 +360,12 @@ dri_destroy_option_cache(struct dri_screen * screen)
    }
 
    free(screen->optionCache.values);
+
+   /* Default values are copied to screen->optionCache->values in
+    * initOptionCache. The info field, however, is a pointer copy, so don't free
+    * that twice.
+    */
+   free(screen->optionCacheDefaults.values);
 }
 
 void

src/gallium/state_trackers/vdpau/device.c

@@ -72,6 +72,11 @@ vdp_imp_device_create_x11(Display *display, int screen, VdpDevice *device,
       goto no_context;
    }
 
+   if (!pscreen->get_param(pscreen, PIPE_CAP_NPOT_TEXTURES)) {
+      ret = VDP_STATUS_NO_IMPLEMENTATION;
+      goto no_context;
+   }
+
    *device = vlAddDataHTAB(dev);
    if (*device == 0) {
       ret = VDP_STATUS_ERROR;
@@ -86,6 +91,7 @@ vdp_imp_device_create_x11(Display *display, int screen, VdpDevice *device,
    return VDP_STATUS_OK;
 
 no_handle:
+   dev->context->destroy(dev->context);
    /* Destroy vscreen */
 no_context:
    vl_screen_destroy(dev->vscreen);

src/gallium/state_trackers/xa/xa_tracker.h

@@ -36,7 +36,7 @@
 
 #include <stdint.h>
 
-#define XA_TRACKER_VERSION_MAJOR 1
+#define XA_TRACKER_VERSION_MAJOR 2
 #define XA_TRACKER_VERSION_MINOR 0
 #define XA_TRACKER_VERSION_PATCH 0
 

src/gallium/targets/egl-static/Makefile.am

@@ -30,7 +30,7 @@
 #
 include $(top_srcdir)/src/gallium/Automake.inc
 
-LDFLAGS += -Wl,$(top_srcdir)/src/gallium/targets/egl-static/egl.link
+LDFLAGS += -Wl,--version-script=$(top_srcdir)/src/gallium/targets/egl-static/egl.link
 
 AM_CFLAGS = $(PTHREAD_CFLAGS)
 AM_CPPFLAGS = \

src/gallium/targets/egl-static/egl.link

@@ -1,3 +1,3 @@
 VERSION {
-{ global: _eglMain; local: *; };
+	global: _eglMain; local: *;
 };

src/gallium/targets/pipe-loader/Makefile.am

@@ -22,7 +22,7 @@
 
 include $(top_srcdir)/src/gallium/Automake.inc
 
-LDFLAGS += -Wl,$(top_srcdir)/src/gallium/targets/pipe-loader/pipe.link
+LDFLAGS += -Wl,--version-script=$(top_srcdir)/src/gallium/targets/pipe-loader/pipe.link
 
 AM_CPPFLAGS = \
 	$(GALLIUM_CFLAGS) \

src/gallium/targets/pipe-loader/pipe.link

@@ -1,3 +1,3 @@
 VERSION {
-	{global: driver_descriptor; local: *;  };
+	global: driver_descriptor; local: *;
 };

src/gbm/backends/dri/gbm_dri.c

@@ -131,14 +131,14 @@ static const __DRIimageLookupExtension image_lookup_extension = {
    dri_lookup_egl_image
 };
 
-const __DRIdri2LoaderExtension dri2_loader_extension = {
+static const __DRIdri2LoaderExtension dri2_loader_extension = {
    { __DRI_DRI2_LOADER, 3 },
    dri_get_buffers,
    dri_flush_front_buffer,
    dri_get_buffers_with_format,
 };
 
-const __DRIimageLoaderExtension image_loader_extension = {
+static const __DRIimageLoaderExtension image_loader_extension = {
    { __DRI_IMAGE_LOADER, 1 },
    image_get_buffers,
    dri_flush_front_buffer,

src/glsl/ast_function.cpp

@@ -1651,7 +1651,7 @@ ast_function_expression::hir(exec_list *instructions,
    } else {
       const ast_expression *id = subexpressions[0];
       const char *func_name = id->primary_expression.identifier;
-      YYLTYPE loc = id->get_location();
+      YYLTYPE loc = get_location();
       exec_list actual_parameters;
 
       process_parameters(instructions, &actual_parameters, &this->expressions,

src/glsl/ast_to_hir.cpp

@@ -734,14 +734,15 @@ do_assignment(exec_list *instructions, struct _mesa_glsl_parse_state *state,
 {
    void *ctx = state;
    bool error_emitted = (lhs->type->is_error() || rhs->type->is_error());
+   ir_rvalue *extract_channel = NULL;
 
    /* If the assignment LHS comes back as an ir_binop_vector_extract
     * expression, move it to the RHS as an ir_triop_vector_insert.
     */
    if (lhs->ir_type == ir_type_expression) {
-      ir_expression *const expr = lhs->as_expression();
+      ir_expression *const lhs_expr = lhs->as_expression();
 
-      if (unlikely(expr->operation == ir_binop_vector_extract)) {
+      if (unlikely(lhs_expr->operation == ir_binop_vector_extract)) {
          ir_rvalue *new_rhs =
             validate_assignment(state, lhs_loc, lhs->type,
                                 rhs, is_initializer);
@@ -749,12 +750,24 @@ do_assignment(exec_list *instructions, struct _mesa_glsl_parse_state *state,
          if (new_rhs == NULL) {
             return lhs;
          } else {
+            /* This converts:
+             * - LHS: (expression float vector_extract <vec> <channel>)
+             * - RHS: <scalar>
+             * into:
+             * - LHS: <vec>
+             * - RHS: (expression vec2 vector_insert <vec> <channel> <scalar>)
+             *
+             * The LHS type is now a vector instead of a scalar.  Since GLSL
+             * allows assignments to be used as rvalues, we need to re-extract
+             * the channel from assignment_temp when returning the rvalue.
+             */
+            extract_channel = lhs_expr->operands[1];
             rhs = new(ctx) ir_expression(ir_triop_vector_insert,
-                                         expr->operands[0]->type,
-                                         expr->operands[0],
+                                         lhs_expr->operands[0]->type,
+                                         lhs_expr->operands[0],
                                          new_rhs,
-                                         expr->operands[1]);
-            lhs = expr->operands[0]->clone(ctx, NULL);
+                                         extract_channel);
+            lhs = lhs_expr->operands[0]->clone(ctx, NULL);
          }
       }
    }
@@ -848,6 +861,11 @@ do_assignment(exec_list *instructions, struct _mesa_glsl_parse_state *state,
    if (!error_emitted)
       instructions->push_tail(new(ctx) ir_assignment(lhs, deref_var));
 
+   if (extract_channel) {
+      return new(ctx) ir_expression(ir_binop_vector_extract,
+                                    new(ctx) ir_dereference_variable(var),
+                                    extract_channel->clone(ctx, NULL));
+   }
    return new(ctx) ir_dereference_variable(var);
 }
 
@@ -2934,7 +2952,7 @@ ast_declarator_list::hir(exec_list *instructions,
                                precision_names[this->type->qualifier.precision],
                                type_name);
          }
-      } else {
+      } else if (this->type->specifier->structure == NULL) {
          _mesa_glsl_warning(&loc, state, "empty declaration");
       }
    }
@@ -3355,6 +3373,15 @@ ast_declarator_list::hir(exec_list *instructions,
       ir_variable *earlier =
          get_variable_being_redeclared(var, decl->get_location(), state,
                                        false /* allow_all_redeclarations */);
+      if (earlier != NULL) {
+         if (strncmp(var->name, "gl_", 3) == 0 &&
+             earlier->how_declared == ir_var_declared_in_block) {
+            _mesa_glsl_error(&loc, state,
+                             "`%s' has already been redeclared using "
+                             "gl_PerVertex", var->name);
+         }
+         earlier->how_declared = ir_var_declared_normally;
+      }
 
       if (decl->initializer != NULL) {
 	 result = process_initializer((earlier == NULL) ? var : earlier,
@@ -5048,6 +5075,7 @@ ast_interface_block::hir(exec_list *instructions,
             _mesa_glsl_error(&loc, state, "`%s' redeclared",
                              this->instance_name);
          }
+         earlier->how_declared = ir_var_declared_normally;
          earlier->type = var->type;
          earlier->reinit_interface_type(block_type);
          delete var;
@@ -5078,7 +5106,11 @@ ast_interface_block::hir(exec_list *instructions,
                _mesa_glsl_error(&loc, state,
                                 "redeclaration of gl_PerVertex can only "
                                 "include built-in variables");
+            } else if (earlier->how_declared == ir_var_declared_normally) {
+               _mesa_glsl_error(&loc, state,
+                                "`%s' has already been redeclared", var->name);
             } else {
+               earlier->how_declared = ir_var_declared_in_block;
                earlier->reinit_interface_type(block_type);
             }
             continue;
@@ -5125,6 +5157,12 @@ ast_interface_block::hir(exec_list *instructions,
             if (var != NULL &&
                 var->get_interface_type() == earlier_per_vertex &&
                 var->mode == var_mode) {
+               if (var->how_declared == ir_var_declared_normally) {
+                  _mesa_glsl_error(&loc, state,
+                                   "redeclaration of gl_PerVertex cannot "
+                                   "follow a redeclaration of `%s'",
+                                   var->name);
+               }
                state->symbols->disable_variable(var->name);
                var->remove();
             }

src/glsl/builtin_functions.cpp

@@ -3864,7 +3864,7 @@ builtin_builder::_fma(const glsl_type *type)
    ir_variable *c = in_var(type, "c");
    MAKE_SIG(type, gpu_shader5, 3, a, b, c);
 
-   body.emit(ret(fma(a, b, c)));
+   body.emit(ret(ir_builder::fma(a, b, c)));
 
    return sig;
 }

src/glsl/builtin_variables.cpp

@@ -434,6 +434,7 @@ builtin_variable_generator::add_variable(const char *name,
                                          enum ir_variable_mode mode, int slot)
 {
    ir_variable *var = new(symtab) ir_variable(type, name, mode);
+   var->how_declared = ir_var_declared_implicitly;
 
    switch (var->mode) {
    case ir_var_auto:

src/glsl/glcpp/glcpp-parse.y

@@ -310,6 +310,11 @@ control_line:
 			_glcpp_parser_expand_and_lex_from (parser,
 							   ELIF_EXPANDED, $2);
 		}
+		else if (parser->skip_stack &&
+		    parser->skip_stack->has_else)
+		{
+			glcpp_error(& @1, parser, "#elif after #else");
+		}
 		else
 		{
 			_glcpp_parser_skip_stack_change_if (parser, & @1,
@@ -324,6 +329,11 @@ control_line:
 		{
 			glcpp_error(& @1, parser, "#elif with no expression");
 		}
+		else if (parser->skip_stack &&
+		    parser->skip_stack->has_else)
+		{
+			glcpp_error(& @1, parser, "#elif after #else");
+		}
 		else
 		{
 			_glcpp_parser_skip_stack_change_if (parser, & @1,
@@ -332,7 +342,17 @@ control_line:
 		}
 	}
 |	HASH_ELSE {
-		_glcpp_parser_skip_stack_change_if (parser, & @1, "else", 1);
+		if (parser->skip_stack &&
+		    parser->skip_stack->has_else)
+		{
+			glcpp_error(& @1, parser, "multiple #else");
+		}
+		else
+		{
+			_glcpp_parser_skip_stack_change_if (parser, & @1, "else", 1);
+			if (parser->skip_stack)
+				parser->skip_stack->has_else = true;
+		}
 	} NEWLINE
 |	HASH_ENDIF {
 		_glcpp_parser_skip_stack_pop (parser, & @1);
@@ -1252,9 +1272,6 @@ glcpp_parser_create (const struct gl_extensions *extensions, int api)
 	      if (extensions->ARB_sample_shading)
 	         add_builtin_define(parser, "GL_ARB_sample_shading", 1);
 
-	      if (extensions->EXT_shader_integer_mix)
-	         add_builtin_define(parser, "GL_EXT_shader_integer_mix", 1);
-
 	      if (extensions->ARB_texture_gather)
 	         add_builtin_define(parser, "GL_ARB_texture_gather", 1);
 
@@ -1263,6 +1280,11 @@ glcpp_parser_create (const struct gl_extensions *extensions, int api)
 	   }
 	}
 
+	if (extensions != NULL) {
+	   if (extensions->EXT_shader_integer_mix)
+	      add_builtin_define(parser, "GL_EXT_shader_integer_mix", 1);
+	}
+
 	language_version = 110;
 	add_builtin_define(parser, "__VERSION__", language_version);
 
@@ -2024,6 +2046,7 @@ _glcpp_parser_skip_stack_push_if (glcpp_parser_t *parser, YYLTYPE *loc,
 		node->type = SKIP_TO_ENDIF;
 	}
 
+	node->has_else = false;
 	node->next = parser->skip_stack;
 	parser->skip_stack = node;
 }

src/glsl/glcpp/glcpp.h

@@ -153,6 +153,7 @@ typedef enum skip_type {
 
 typedef struct skip_node {
 	skip_type_t type;
+	bool has_else;
 	YYLTYPE loc; /* location of the initial #if/#elif/... */
 	struct skip_node *next;
 } skip_node_t;

src/glsl/glcpp/tests/118-multiple-else.c

@@ -0,0 +1,6 @@
+#if 0
+#else
+int foo;
+#else
+int bar;
+#endif

src/glsl/glcpp/tests/118-multiple-else.c.expected

@@ -0,0 +1,8 @@
+0:4(1): preprocessor error: multiple #else
+
+
+int foo;
+
+int bar;
+
+

src/glsl/glcpp/tests/119-elif-after-else.c

@@ -0,0 +1,6 @@
+#if 0
+#else
+int foo;
+#elif 0
+int bar;
+#endif

src/glsl/glcpp/tests/119-elif-after-else.c.expected

@@ -0,0 +1,8 @@
+0:4(1): preprocessor error: #elif after #else
+
+
+int foo;
+
+int bar;
+
+

src/glsl/glsl_parser_extras.cpp

@@ -191,6 +191,8 @@ _mesa_glsl_parse_state::_mesa_glsl_parse_state(struct gl_context *_ctx,
    this->gs_input_prim_type = GL_POINTS;
    this->gs_input_size = 0;
    this->out_qualifier = new(this) ast_type_qualifier();
+   memset(this->atomic_counter_offsets, 0,
+          sizeof(this->atomic_counter_offsets));
 }
 
 /**
@@ -290,6 +292,10 @@ _mesa_glsl_parse_state::process_version_directive(YYLTYPE *locp, int version,
       }
    }
 
+   if (this->es_shader) {
+      this->ARB_texture_rectangle_enable = false;
+   }
+
    this->language_version = version;
 
    bool supported = false;

src/glsl/ir.cpp

@@ -1586,7 +1586,8 @@ ir_variable::ir_variable(const struct glsl_type *type, const char *name,
 			 ir_variable_mode mode)
    : max_array_access(0), max_ifc_array_access(NULL),
      read_only(false), centroid(false), invariant(false),
-        mode(mode), interpolation(INTERP_QUALIFIER_NONE), atomic()
+     how_declared(ir_var_declared_normally), mode(mode),
+     interpolation(INTERP_QUALIFIER_NONE), atomic()
 {
    this->ir_type = ir_type_variable;
    this->type = type;

src/glsl/ir.h

@@ -283,6 +283,34 @@ enum ir_variable_mode {
    ir_var_mode_count	/**< Number of variable modes */
 };
 
+/**
+ * Enum keeping track of how a variable was declared.  For error checking of
+ * the gl_PerVertex redeclaration rules.
+ */
+enum ir_var_declaration_type {
+   /**
+    * Normal declaration (for most variables, this means an explicit
+    * declaration.  Exception: temporaries are always implicitly declared, but
+    * they still use ir_var_declared_normally).
+    *
+    * Note: an ir_variable that represents a named interface block uses
+    * ir_var_declared_normally.
+    */
+   ir_var_declared_normally = 0,
+
+   /**
+    * Variable was explicitly declared (or re-declared) in an unnamed
+    * interface block.
+    */
+   ir_var_declared_in_block,
+
+   /**
+    * Variable is an implicitly declared built-in that has not been explicitly
+    * re-declared by the shader.
+    */
+   ir_var_declared_implicitly,
+};
+
 /**
  * \brief Layout qualifiers for gl_FragDepth.
  *
@@ -515,6 +543,14 @@ public:
     */
    unsigned assigned:1;
 
+   /**
+    * Enum indicating how the variable was declared.  See
+    * ir_var_declaration_type.
+    *
+    * This is used to detect certain kinds of illegal variable redeclarations.
+    */
+   unsigned how_declared:2;
+
    /**
     * Storage class of the variable.
     *

src/glsl/ir_clone.cpp

@@ -68,6 +68,7 @@ ir_variable::clone(void *mem_ctx, struct hash_table *ht) const
    var->has_initializer = this->has_initializer;
    var->depth_layout = this->depth_layout;
    var->assigned = this->assigned;
+   var->how_declared = this->how_declared;
    var->used = this->used;
 
    var->num_state_slots = this->num_state_slots;
@@ -162,7 +163,8 @@ ir_loop::clone(void *mem_ctx, struct hash_table *ht) const
       new_loop->to = this->to->clone(mem_ctx, ht);
    if (this->increment)
       new_loop->increment = this->increment->clone(mem_ctx, ht);
-   new_loop->counter = counter;
+   if (this->counter)
+      new_loop->counter = this->counter->clone(mem_ctx, ht);
 
    foreach_iter(exec_list_iterator, iter, this->body_instructions) {
       ir_instruction *ir = (ir_instruction *)iter.get();

src/glsl/ir_constant_expression.cpp

@@ -1402,7 +1402,7 @@ ir_expression::constant_expression_value(struct hash_table *variable_context)
          data.f[c] = ldexp(op[0]->value.f[c], op[1]->value.i[c]);
          /* Flush subnormal values to zero. */
          if (!isnormal(data.f[c]))
-            data.f[c] = copysign(0.0, op[0]->value.f[c]);
+            data.f[c] = copysign(0.0f, op[0]->value.f[c]);
       }
       break;
 

src/glsl/ir_hv_accept.cpp

@@ -87,6 +87,12 @@ ir_loop::accept(ir_hierarchical_visitor *v)
    if (s != visit_continue)
       return (s == visit_continue_with_parent) ? visit_continue : s;
 
+   if (this->counter) {
+      s = this->counter->accept(v);
+      if (s != visit_continue)
+         return (s == visit_continue_with_parent) ? visit_continue : s;
+   }
+
    s = visit_list_elements(v, &this->body_instructions);
    if (s == visit_stop)
       return s;

src/glsl/ir_variable_refcount.cpp

@@ -132,3 +132,24 @@ ir_variable_refcount_visitor::visit_leave(ir_assignment *ir)
 
    return visit_continue;
 }
+
+
+ir_visitor_status
+ir_variable_refcount_visitor::visit_leave(ir_loop *ir)
+{
+   /* If the loop has a counter variable, it is implicitly referenced and
+    * assigned to.  Note that since the LHS of an assignment is counted as a
+    * reference, we actually have to increment referenced_count by 2 so that
+    * later code will know that the variable isn't just assigned to.
+    */
+   if (ir->counter != NULL) {
+      ir_variable_refcount_entry *entry =
+         this->get_variable_entry(ir->counter);
+      if (entry) {
+         entry->referenced_count += 2;
+         entry->assigned_count++;
+      }
+   }
+
+   return visit_continue;
+}

src/glsl/ir_variable_refcount.h

@@ -60,6 +60,7 @@ public:
 
    virtual ir_visitor_status visit_enter(ir_function_signature *);
    virtual ir_visitor_status visit_leave(ir_assignment *);
+   virtual ir_visitor_status visit_leave(ir_loop *);
 
    ir_variable_refcount_entry *get_variable_entry(ir_variable *var);
 

src/glsl/link_interface_blocks.cpp

@@ -60,6 +60,7 @@ struct interface_block_definition
          if (var->type->is_array())
             array_size = var->type->length;
       }
+      explicitly_declared = (var->how_declared != ir_var_declared_implicitly);
    }
 
    /**
@@ -77,6 +78,12 @@ struct interface_block_definition
     * Otherwise -1.
     */
    int array_size;
+
+   /**
+    * True if this interface block was explicitly declared in the shader;
+    * false if it was an implicitly declared built-in interface block.
+    */
+   bool explicitly_declared;
 };
 
 
@@ -91,8 +98,14 @@ intrastage_match(interface_block_definition *a,
                  ir_variable_mode mode)
 {
    /* Types must match. */
-   if (a->type != b->type)
-      return false;
+   if (a->type != b->type) {
+      /* Exception: if both the interface blocks are implicitly declared,
+       * don't force their types to match.  They might mismatch due to the two
+       * shaders using different GLSL versions, and that's ok.
+       */
+      if (a->explicitly_declared || b->explicitly_declared)
+         return false;
+   }
 
    /* Presence/absence of interface names must match. */
    if ((a->instance_name == NULL) != (b->instance_name == NULL))
@@ -144,8 +157,14 @@ interstage_match(const interface_block_definition *producer,
    assert(producer->array_size != 0);
 
    /* Types must match. */
-   if (consumer->type != producer->type)
-      return false;
+   if (consumer->type != producer->type) {
+      /* Exception: if both the interface blocks are implicitly declared,
+       * don't force their types to match.  They might mismatch due to the two
+       * shaders using different GLSL versions, and that's ok.
+       */
+      if (consumer->explicitly_declared || producer->explicitly_declared)
+         return false;
+   }
    if (extra_array_level) {
       /* Consumer must be an array, and producer must not. */
       if (consumer->array_size == -1)
@@ -289,57 +308,78 @@ validate_intrastage_interface_blocks(struct gl_shader_program *prog,
 }
 
 void
-validate_interstage_interface_blocks(struct gl_shader_program *prog,
-                                     const gl_shader *producer,
-                                     const gl_shader *consumer)
+validate_interstage_inout_blocks(struct gl_shader_program *prog,
+                                 const gl_shader *producer,
+                                 const gl_shader *consumer)
 {
-   interface_block_definitions inout_interfaces;
-   interface_block_definitions uniform_interfaces;
+   interface_block_definitions definitions;
    const bool extra_array_level = consumer->Type == GL_GEOMETRY_SHADER;
 
-   /* Add non-output interfaces from the consumer to the symbol table. */
+   /* Add input interfaces from the consumer to the symbol table. */
    foreach_list(node, consumer->ir) {
       ir_variable *var = ((ir_instruction *) node)->as_variable();
-      if (!var || !var->get_interface_type() || var->mode == ir_var_shader_out)
+      if (!var || !var->get_interface_type() || var->mode != ir_var_shader_in)
          continue;
 
-      interface_block_definitions *definitions = var->mode == ir_var_uniform ?
-         &uniform_interfaces : &inout_interfaces;
-      definitions->store(interface_block_definition(var));
+      definitions.store(interface_block_definition(var));
    }
 
-   /* Verify that the producer's interfaces match. */
+   /* Verify that the producer's output interfaces match. */
    foreach_list(node, producer->ir) {
       ir_variable *var = ((ir_instruction *) node)->as_variable();
-      if (!var || !var->get_interface_type() || var->mode == ir_var_shader_in)
+      if (!var || !var->get_interface_type() || var->mode != ir_var_shader_out)
          continue;
 
-      interface_block_definitions *definitions = var->mode == ir_var_uniform ?
-         &uniform_interfaces : &inout_interfaces;
       interface_block_definition *consumer_def =
-         definitions->lookup(var->get_interface_type()->name);
+         definitions.lookup(var->get_interface_type()->name);
 
       /* The consumer doesn't use this output block.  Ignore it. */
       if (consumer_def == NULL)
          continue;
 
       const interface_block_definition producer_def(var);
-      bool match;
-      if (var->mode == ir_var_uniform) {
-         /* Uniform matching rules are the same for interstage and intrastage
-          * linking.
-          */
-         match = intrastage_match(consumer_def, &producer_def,
-                                  (ir_variable_mode) var->mode);
-      } else {
-         match = interstage_match(&producer_def, consumer_def,
-                                  extra_array_level);
-      }
 
-      if (!match) {
+      if (!interstage_match(&producer_def, consumer_def, extra_array_level)) {
          linker_error(prog, "definitions of interface block `%s' do not "
                       "match\n", var->get_interface_type()->name);
          return;
       }
    }
 }
+
+
+void
+validate_interstage_uniform_blocks(struct gl_shader_program *prog,
+                                   gl_shader **stages, int num_stages)
+{
+   interface_block_definitions definitions;
+
+   for (int i = 0; i < num_stages; i++) {
+      if (stages[i] == NULL)
+         continue;
+
+      const gl_shader *stage = stages[i];
+      foreach_list(node, stage->ir) {
+         ir_variable *var = ((ir_instruction *) node)->as_variable();
+         if (!var || !var->get_interface_type() || var->mode != ir_var_uniform)
+            continue;
+
+         interface_block_definition *old_def =
+            definitions.lookup(var->get_interface_type()->name);
+         const interface_block_definition new_def(var);
+         if (old_def == NULL) {
+            definitions.store(new_def);
+         } else {
+            /* Interstage uniform matching rules are the same as intrastage
+             * uniform matchin rules (for uniforms, it is as though all
+             * shaders are in the same shader stage).
+             */
+            if (!intrastage_match(old_def, &new_def, ir_var_uniform)) {
+               linker_error(prog, "definitions of interface block `%s' do not "
+                            "match\n", var->get_interface_type()->name);
+               return;
+            }
+         }
+      }
+   }
+}

src/glsl/linker.cpp

@@ -2154,8 +2154,8 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
       if (prog->_LinkedShaders[i] == NULL)
          continue;
 
-      validate_interstage_interface_blocks(prog, prog->_LinkedShaders[prev],
-                                           prog->_LinkedShaders[i]);
+      validate_interstage_inout_blocks(prog, prog->_LinkedShaders[prev],
+                                       prog->_LinkedShaders[i]);
       if (!prog->LinkStatus)
          goto done;
 
@@ -2168,6 +2168,11 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
       prev = i;
    }
 
+   /* Cross-validate uniform blocks between shader stages */
+   validate_interstage_uniform_blocks(prog, prog->_LinkedShaders,
+                                      MESA_SHADER_TYPES);
+   if (!prog->LinkStatus)
+      goto done;
 
    for (unsigned int i = 0; i < MESA_SHADER_TYPES; i++) {
       if (prog->_LinkedShaders[i] != NULL)

src/glsl/linker.h

@@ -65,9 +65,13 @@ validate_intrastage_interface_blocks(struct gl_shader_program *prog,
                                      unsigned num_shaders);
 
 void
-validate_interstage_interface_blocks(struct gl_shader_program *prog,
-                                     const gl_shader *producer,
-                                     const gl_shader *consumer);
+validate_interstage_inout_blocks(struct gl_shader_program *prog,
+                                 const gl_shader *producer,
+                                 const gl_shader *consumer);
+
+void
+validate_interstage_uniform_blocks(struct gl_shader_program *prog,
+                                   gl_shader **stages, int num_stages);
 
 extern void
 link_assign_atomic_counter_resources(struct gl_context *ctx,

src/glsl/loop_controls.cpp

@@ -254,7 +254,7 @@ loop_control_visitor::visit_leave(ir_loop *ir)
 		     ir->from = init->clone(ir, NULL);
 		     ir->to = limit->clone(ir, NULL);
 		     ir->increment = lv->increment->clone(ir, NULL);
-		     ir->counter = lv->var;
+		     ir->counter = lv->var->clone(ir, NULL);
 		     ir->cmp = cmp;
 
 		     max_iterations = iterations;

src/glsl/lower_clip_distance.cpp

@@ -381,6 +381,11 @@ lower_clip_distance_visitor::fix_lhs(ir_assignment *ir)
 ir_visitor_status
 lower_clip_distance_visitor::visit_leave(ir_assignment *ir)
 {
+   /* First invoke the base class visitor.  This causes handle_rvalue() to be
+    * called on ir->rhs and ir->condition.
+    */
+   ir_rvalue_visitor::visit_leave(ir);
+
    if (this->is_clip_distance_vec8(ir->lhs) ||
        this->is_clip_distance_vec8(ir->rhs)) {
       /* LHS or RHS of the assignment is the entire 1D gl_ClipDistance array

src/glx/glxcmds.c

@@ -183,7 +183,7 @@ GetGLXPrivScreenConfig(Display * dpy, int scrn, struct glx_display ** ppriv,
 
    /* Check to see if the GL is supported on this screen */
    *ppsc = (*ppriv)->screens[scrn];
-   if ((*ppsc)->configs == NULL) {
+   if ((*ppsc)->configs == NULL && (*ppsc)->visuals == NULL) {
       /* No support for GL on this screen regardless of visual */
       return GLX_BAD_VISUAL;
    }

src/mesa/Makefile.am

@@ -95,6 +95,12 @@ AM_CPPFLAGS = $(DEFINES) $(INCLUDE_DIRS)
 AM_CFLAGS = $(LLVM_CFLAGS) $(VISIBILITY_CFLAGS)
 AM_CXXFLAGS = $(LLVM_CFLAGS) $(VISIBILITY_CXXFLAGS)
 
+ARCH_LIBS =
+
+if SSE41_SUPPORTED
+ARCH_LIBS += libmesa_sse41.la
+endif
+
 MESA_ASM_FILES_FOR_ARCH =
 
 if HAVE_X86_ASM
@@ -103,8 +109,6 @@ noinst_PROGRAMS = gen_matypes
 gen_matypes_SOURCES = x86/gen_matypes.c
 BUILT_SOURCES += matypes.h
 
-ARCH_LIBS = libmesa_sse41.la
-
 if HAVE_X86_64_ASM
 MESA_ASM_FILES_FOR_ARCH += $(X86_64_FILES)
 AM_CPPFLAGS += -I$(builddir)/x86-64 -I$(srcdir)/x86-64

src/mesa/drivers/dri/common/dri_util.c

@@ -407,8 +407,10 @@ driCreateContextAttribs(__DRIscreen *screen, int api,
        mesa_api = API_OPENGL_CORE;
     }
 
-    if ((flags & ~(__DRI_CTX_FLAG_DEBUG | __DRI_CTX_FLAG_FORWARD_COMPATIBLE))
-        != 0) {
+    const uint32_t allowed_flags = (__DRI_CTX_FLAG_DEBUG
+                                    | __DRI_CTX_FLAG_FORWARD_COMPATIBLE
+                                    | __DRI_CTX_FLAG_ROBUST_BUFFER_ACCESS);
+    if (flags & ~allowed_flags) {
 	*error = __DRI_CTX_ERROR_UNKNOWN_FLAG;
 	return NULL;
     }
@@ -436,16 +438,19 @@ driCreateContextAttribs(__DRIscreen *screen, int api,
         return NULL;
     }
 
-    struct gl_context *ctx = context->driverPrivate;
+    *error = __DRI_CTX_ERROR_SUCCESS;
+    return context;
+}
+
+void
+driContextSetFlags(struct gl_context *ctx, uint32_t flags)
+{
     if ((flags & __DRI_CTX_FLAG_FORWARD_COMPATIBLE) != 0)
         ctx->Const.ContextFlags |= GL_CONTEXT_FLAG_FORWARD_COMPATIBLE_BIT;
     if ((flags & __DRI_CTX_FLAG_DEBUG) != 0) {
         ctx->Const.ContextFlags |= GL_CONTEXT_FLAG_DEBUG_BIT;
         ctx->Debug.DebugOutput = GL_TRUE;
     }
-
-    *error = __DRI_CTX_ERROR_SUCCESS;
-    return context;
 }
 
 static __DRIcontext *

src/mesa/drivers/dri/common/dri_util.h

@@ -289,6 +289,9 @@ dri2InvalidateDrawable(__DRIdrawable *drawable);
 extern void
 driUpdateFramebufferSize(struct gl_context *ctx, const __DRIdrawable *dPriv);
 
+extern void
+driContextSetFlags(struct gl_context *ctx, uint32_t flags);
+
 extern const __DRIimageDriverExtension driImageDriverExtension;
 
 #endif /* _DRI_UTIL_H_ */

src/mesa/drivers/dri/common/megadriver_stub.c

@@ -23,6 +23,132 @@
 
 #include <stdio.h>
 #include "dri_util.h"
+#include <dlfcn.h>
+#include "main/macros.h"
+
+/* We need GNU extensions to dlfcn.h in order to provide backward
+ * compatibility for the older DRI driver loader mechanism. (dladdr,
+ * Dl_info, and RTLD_DEFAULT are only defined when _GNU_SOURCE is
+ * defined.)
+ */
+#ifdef _GNU_SOURCE
+
+#define MEGADRIVER_STUB_MAX_EXTENSIONS 10
+#define LIB_PATH_SUFFIX "_dri.so"
+#define LIB_PATH_SUFFIX_LENGTH (sizeof(LIB_PATH_SUFFIX)-1)
+
+/* This is the table of extensions that the loader will dlsym() for.
+ *
+ * Initially it is empty for the megadriver stub, but the library
+ * constructor may initialize it based on the name of the library that
+ * is being loaded.
+ */
+PUBLIC const __DRIextension *
+__driDriverExtensions[MEGADRIVER_STUB_MAX_EXTENSIONS] = {
+   NULL
+};
+
+/**
+ * This is a constructor function for the megadriver dynamic library.
+ *
+ * When the driver is dlopen'ed, this function will run. It will
+ * search for the name of the foo_dri.so file that was opened using
+ * the dladdr function.
+ *
+ * After finding foo's name, it will call __driDriverGetExtensions_foo
+ * and use the return to update __driDriverExtensions to enable
+ * compatibility with older DRI driver loaders.
+ */
+__attribute__((constructor)) static void
+megadriver_stub_init(void)
+{
+   Dl_info info;
+   char *driver_name;
+   size_t name_len;
+   char *get_extensions_name;
+   const __DRIextension **(*get_extensions)(void);
+   const __DRIextension **extensions;
+   int i;
+
+   /* Call dladdr on __driDriverExtensions. We are really
+    * interested in the returned info.dli_fname so we can
+    * figure out the path name of the library being loaded.
+    */
+   i = dladdr((void*) __driDriverExtensions, &info);
+   if (i == 0)
+      return;
+
+   /* Search for the last '/' character in the path. */
+   driver_name = strrchr(info.dli_fname, '/');
+   if (driver_name != NULL) {
+      /* Skip '/' character */
+      driver_name++;
+   } else {
+      /* Try using the start of the path */
+      driver_name = (char*) info.dli_fname;
+   }
+
+   /* Make sure the path ends with _dri.so */
+   name_len = strlen(driver_name);
+   i = name_len - LIB_PATH_SUFFIX_LENGTH;
+   if (i < 0 || strcmp(driver_name + i, LIB_PATH_SUFFIX) != 0)
+      return;
+
+   /* Duplicate the string so we can modify it.
+    * So far we've been using info.dli_fname.
+    */
+   driver_name = strdup(driver_name);
+   if (!driver_name)
+      return;
+
+   /* The path ends with _dri.so. Chop this part of the
+    * string off. Then we'll have the driver's final name.
+    */
+   driver_name[i] = '\0';
+
+   i = asprintf(&get_extensions_name, "%s_%s",
+                __DRI_DRIVER_GET_EXTENSIONS, driver_name);
+   free(driver_name);
+   if (i == -1)
+      return;
+
+   /* dlsym to get the driver's get extensions function. We
+    * don't have the dlopen handle, so we have to use
+    * RTLD_DEFAULT. It seems unlikely that the symbol will
+    * be found in another library, but this isn't optimal.
+    */
+   get_extensions = dlsym(RTLD_DEFAULT, get_extensions_name);
+   free(get_extensions_name);
+   if (!get_extensions)
+      return;
+
+   /* Use the newer DRI loader entrypoint to find extensions.
+    * We will then expose these extensions via the older
+    * __driDriverExtensions symbol.
+    */
+   extensions = get_extensions();
+
+   /* Copy the extensions into the __driDriverExtensions array
+    * we declared.
+    */
+   for (i = 0; i < ARRAY_SIZE(__driDriverExtensions); i++) {
+      __driDriverExtensions[i] = extensions[i];
+      if (extensions[i] == NULL)
+         break;
+   }
+
+   /* If the driver had more extensions than we reserved, then
+    * bail out.
+    */
+   if (i == ARRAY_SIZE(__driDriverExtensions)) {
+      __driDriverExtensions[0] = NULL;
+      fprintf(stderr, "Megadriver stub did not reserve enough extension "
+              "slots.\n");
+      return;
+   }
+}
+
+#endif /* _GNU_SOURCE */
 
 static const
 __DRIconfig **stub_error_init_screen(__DRIscreen *psp)

src/mesa/drivers/dri/i915/i830_context.c

@@ -56,6 +56,7 @@ i830CreateContext(int api,
                   __DRIcontext * driContextPriv,
                   unsigned major_version,
                   unsigned minor_version,
+                  uint32_t flags,
                   unsigned *error,
                   void *sharedContextPrivate)
 {
@@ -73,7 +74,7 @@ i830CreateContext(int api,
    i830InitDriverFunctions(&functions);
 
    if (!intelInitContext(intel, __DRI_API_OPENGL,
-                         major_version, minor_version,
+                         major_version, minor_version, flags,
                          mesaVis, driContextPriv,
                          sharedContextPrivate, &functions,
                          error)) {

src/mesa/drivers/dri/i915/i830_context.h

@@ -183,6 +183,7 @@ i830CreateContext(int api,
                   __DRIcontext * driContextPriv,
                   unsigned major_version,
                   unsigned minor_version,
+                  uint32_t flags,
                   unsigned *error,
                   void *sharedContextPrivate);
 

src/mesa/drivers/dri/i915/i915_context.c

@@ -151,6 +151,7 @@ i915CreateContext(int api,
                   __DRIcontext * driContextPriv,
                   unsigned major_version,
                   unsigned minor_version,
+                  uint32_t flags,
                   unsigned *error,
                   void *sharedContextPrivate)
 {
@@ -168,7 +169,7 @@ i915CreateContext(int api,
 
    i915InitDriverFunctions(&functions);
 
-   if (!intelInitContext(intel, api, major_version, minor_version,
+   if (!intelInitContext(intel, api, major_version, minor_version, flags,
                          mesaVis, driContextPriv,
                          sharedContextPrivate, &functions,
                          error)) {

src/mesa/drivers/dri/i915/i915_context.h

@@ -324,6 +324,7 @@ extern bool i915CreateContext(int api,
 			      __DRIcontext * driContextPriv,
                               unsigned major_version,
                               unsigned minor_version,
+                              uint32_t flags,
                               unsigned *error,
 			      void *sharedContextPrivate);
 

src/mesa/drivers/dri/i915/i915_fragprog.c

@@ -146,6 +146,7 @@ src_vector(struct i915_fragment_program *p,
    case PROGRAM_OUTPUT:
       switch (source->Index) {
       case FRAG_RESULT_COLOR:
+      case FRAG_RESULT_DATA0:
 	 src = UREG(REG_TYPE_OC, 0);
 	 break;
       case FRAG_RESULT_DEPTH:

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

@@ -409,6 +409,7 @@ intelInitContext(struct intel_context *intel,
                  int api,
                  unsigned major_version,
                  unsigned minor_version,
+                 uint32_t flags,
                  const struct gl_config * mesaVis,
                  __DRIcontext * driContextPriv,
                  void *sharedContextPrivate,

src/mesa/drivers/dri/i915/intel_context.h

@@ -401,6 +401,7 @@ extern bool intelInitContext(struct intel_context *intel,
                              int api,
                              unsigned major_version,
                              unsigned minor_version,
+                             uint32_t flags,
                              const struct gl_config * mesaVis,
                              __DRIcontext * driContextPriv,
                              void *sharedContextPrivate,

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

@@ -930,6 +930,7 @@ i830CreateContext(int api,
 		  __DRIcontext *driContextPriv,
 		  unsigned major_version,
 		  unsigned minor_version,
+                  uint32_t flags,
 		  unsigned *error,
 		  void *sharedContextPrivate);
 
@@ -939,6 +940,7 @@ i915CreateContext(int api,
 		  __DRIcontext *driContextPriv,
                   unsigned major_version,
                   unsigned minor_version,
+                  uint32_t flags,
                   unsigned *error,
 		  void *sharedContextPrivate);
 
@@ -970,13 +972,13 @@ intelCreateContext(gl_api api,
 
    if (IS_9XX(intelScreen->deviceID)) {
       success = i915CreateContext(api, mesaVis, driContextPriv,
-                                  major_version, minor_version, error,
-                                  sharedContextPrivate);
+                                  major_version, minor_version, flags,
+                                  error, sharedContextPrivate);
    } else {
       intelScreen->no_vbo = true;
       success = i830CreateContext(api, mesaVis, driContextPriv,
-                                  major_version, minor_version, error,
-                                  sharedContextPrivate);
+                                  major_version, minor_version, flags,
+                                  error, sharedContextPrivate);
    }
 
    if (success)

src/mesa/drivers/dri/i965/Makefile.am

@@ -46,8 +46,8 @@ TEST_LIBS = \
 	libi965_dri.la \
 	../common/libdricommon.la \
 	../common/libmegadriver_stub.la \
-	$(DRI_LIB_DEPS) \
         ../../../libmesa.la \
+	$(DRI_LIB_DEPS) \
         -lrt \
 	../common/libdri_test_stubs.la
 

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

@@ -128,7 +128,7 @@ static void
 brw_gs_upload_binding_table(struct brw_context *brw)
 {
    /* If there's no GS, skip changing anything. */
-   if (!brw->gs.prog_data)
+   if (brw->geometry_program == NULL)
       return;
 
    brw_upload_binding_table(brw, BRW_NEW_GS_BINDING_TABLE, &brw->gs.base);

src/mesa/drivers/dri/i965/brw_blorp.cpp

@@ -266,8 +266,6 @@ retry:
     */
    brw->state.dirty.brw = ~0;
    brw->state.dirty.cache = ~0;
-   brw->state_batch_count = 0;
-   brw->batch.need_workaround_flush = true;
    brw->ib.type = -1;
    intel_batchbuffer_clear_cache(brw);