docs: Add release notes for the 10.1.6 release.

In preparation for the 10.1.6 release.

VERSION

@@ -1 +1 @@
-10.1.0
+10.1.6

bin/.cherry-ignore

@@ -0,0 +1,28 @@
+# This patch does not apply cleanly, author says it can be skipped.
+dff3eccd158d648482bb47118ef5d57a9186e5a4
+
+# And this one depends on the above, author says it too can be skipped.
+ac35ded4733883037316d556af596524e5e02535
+
+# This patch introduces some regressions. See:
+# https://bugs.freedesktop.org/show_bug.cgi?id=77443
+1afe3359258a9e89b62c8638761f52d78f6d1cbc
+
+# Author retracted this from consideration for stable branch
+3e817e7e56806d8adb8f16c35136045c29908944
+
+# And this one was simply a bug fix for the previously-retracted commit
+2bab95973d8ad3a84f62670143d6f26c230d9582
+
+# Here we have a commit, and its subsequent "revert" both proposed within a
+# single window of the stable release. So we can achieve the same final effect
+# by ignoring both of the commits.
+e3cc0d90e14e62a0a787b6c07a6df0f5c84039be
+0d5ec2c615784929be095951f9269773a790a2dd
+
+# The function being modified here (_eglCreateWindowSurfaceCommon) does not
+# exist in the 10.1 branch.
+91ff0d4c6510dc38f279c586ced17fba917873e7
+
+# This patch is not needed (modifies work only in 10.2)
+6980cae6aeb6671b6b0245e20a2d34957c1fff0a

bin/get-pick-list.sh

@@ -14,7 +14,7 @@ git log --reverse --grep="cherry picked from commit" origin/master..HEAD |\
 	sed -e 's/^[[:space:]]*(cherry picked from commit[[:space:]]*//' -e 's/)//' > already_picked
 
 # Grep for commits that were marked as a candidate for the stable tree.
-git log --reverse --pretty=%H -i --grep='^\([[:space:]]*NOTE: .*[Cc]andidate\|CC:.*mesa-stable\)' HEAD..origin/master |\
+git log --reverse --pretty=%H -i --grep='^\([[:space:]]*NOTE: .*[Cc]andidate\|CC:.*10\.1.*mesa-stable\)' HEAD..origin/master |\
 while read sha
 do
 	# Check to see whether the patch is on the ignore list.

configure.ac

@@ -542,11 +542,20 @@ AC_ARG_ENABLE([dri],
         [enable DRI modules @<:@default=enabled@:>@])],
     [enable_dri="$enableval"],
     [enable_dri=yes])
+
+case "$host_os" in
+linux*)
+    dri3_default=yes
+    ;;
+*)
+    dri3_default=no
+    ;;
+esac
 AC_ARG_ENABLE([dri3],
     [AS_HELP_STRING([--enable-dri3],
-        [enable DRI3 @<:@default=enabled@:>@])],
+        [enable DRI3 @<:@default=auto@:>@])],
     [enable_dri3="$enableval"],
-    [enable_dri3=yes])
+    [enable_dri3="$dri3_default"])
 AC_ARG_ENABLE([glx],
     [AS_HELP_STRING([--enable-glx],
         [enable GLX library @<:@default=enabled@:>@])],
@@ -771,6 +780,13 @@ if test "x$have_libdrm" = xyes; then
 	DEFINES="$DEFINES -DHAVE_LIBDRM"
 fi
 
+case "$host_os" in
+linux*)
+    need_libudev=yes ;;
+*)
+    need_libudev=no ;;
+esac
+
 PKG_CHECK_MODULES([LIBUDEV], [libudev >= $LIBUDEV_REQUIRED],
                   have_libudev=yes, have_libudev=no)
 
@@ -830,9 +846,6 @@ xyesno)
         PKG_CHECK_MODULES([DRI2PROTO], [dri2proto >= $DRI2PROTO_REQUIRED])
         GL_PC_REQ_PRIV="$GL_PC_REQ_PRIV libdrm >= $LIBDRM_REQUIRED"
         if test x"$enable_dri3" = xyes; then
-            if test x"$have_libudev" != xyes; then
-              AC_MSG_ERROR([DRI3 requires libudev >= $LIBUDEV_REQUIRED])
-            fi
             PKG_CHECK_MODULES([DRI3PROTO], [dri3proto >= $DRI3PROTO_REQUIRED])
             PKG_CHECK_MODULES([PRESENTPROTO], [presentproto >= $PRESENTPROTO_REQUIRED])
         fi
@@ -1017,7 +1030,7 @@ if test "x$enable_dri" = xyes; then
     gnu*)
         DEFINES="$DEFINES -DUSE_EXTERNAL_DXTN_LIB=1"
         DEFINES="$DEFINES -DHAVE_ALIAS"
-	;;
+        ;;
     solaris*)
         DEFINES="$DEFINES -DUSE_EXTERNAL_DXTN_LIB=1"
         ;;
@@ -1037,7 +1050,7 @@ if test "x$enable_dri" = xyes; then
     DRI_DIRS=`echo "$DRI_DIRS" | $SED 's/  */ /g'`
 
     # Check for expat
-    PKG_CHECK_EXISTS([EXPAT], [have_expat=yes], [have_expat=no])
+    PKG_CHECK_EXISTS([expat], [have_expat=yes], [have_expat=no])
     if test "x$have_expat" = "xyes"; then
        PKG_CHECK_MODULES([EXPAT], [expat], [],
                          AC_MSG_ERROR([Expat required for DRI.]))
@@ -1178,8 +1191,8 @@ if test "x$enable_gbm" = xauto; then
     esac
 fi
 if test "x$enable_gbm" = xyes; then
-    if test x"$have_libudev" != xyes; then
-        AC_MSG_ERROR([gbm needs udev])
+    if test "x$need_libudev$have_libudev" = xyesno; then
+        AC_MSG_ERROR([gbm requires udev >= $LIBUDEV_REQUIRED])
     fi
 
     if test "x$enable_dri" = xyes; then
@@ -1187,10 +1200,21 @@ if test "x$enable_gbm" = xyes; then
         if test "x$enable_shared_glapi" = xno; then
             AC_MSG_ERROR([gbm_dri requires --enable-shared-glapi])
         fi
+    else
+        # Strictly speaking libgbm does not require --enable-dri, although
+        # both of its backends do. Thus one can build libgbm without any
+        # backends if --disable-dri is set.
+        # To avoid unnecessary complexity of checking if at least one backend
+        # is available when building, just mandate --enable-dri.
+        AC_MSG_ERROR([gbm requires --enable-dri])
     fi
 fi
 AM_CONDITIONAL(HAVE_GBM, test "x$enable_gbm" = xyes)
-GBM_PC_REQ_PRIV="libudev"
+if test "x$need_libudev" = xyes; then
+    GBM_PC_REQ_PRIV="libudev >= $LIBUDEV_REQUIRED"
+else
+    GBM_PC_REQ_PRIV=""
+fi
 GBM_PC_LIB_PRIV="$DLOPEN_LIBS"
 AC_SUBST([GBM_PC_REQ_PRIV])
 AC_SUBST([GBM_PC_LIB_PRIV])
@@ -1461,9 +1485,9 @@ for plat in $egl_platforms; do
 		;;
 	esac
 
-        case "$plat$have_libudev" in
-                waylandno|drmno)
-                    AC_MSG_ERROR([cannot build $plat platfrom without udev]) ;;
+        case "$plat$need_libudev$have_libudev" in
+                waylandyesno|drmyesno)
+                    AC_MSG_ERROR([cannot build $plat platform without udev >= $LIBUDEV_REQUIRED]) ;;
         esac
 done
 
@@ -1529,11 +1553,11 @@ AC_ARG_ENABLE([gallium-llvm],
     [enable_gallium_llvm="$enableval"],
     [enable_gallium_llvm=auto])
 
-AC_ARG_WITH([llvm-shared-libs],
-    [AS_HELP_STRING([--with-llvm-shared-libs],
-        [link with LLVM shared libraries @<:@default=disabled@:>@])],
+AC_ARG_ENABLE([llvm-shared-libs],
+    [AS_HELP_STRING([--enable-llvm-shared-libs],
+        [link with LLVM shared libraries @<:@default=enabled@:>@])],
     [],
-    [with_llvm_shared_libs=no])
+    [with_llvm_shared_libs=yes])
 
 AC_ARG_WITH([llvm-prefix],
     [AS_HELP_STRING([--with-llvm-prefix],
@@ -1588,6 +1612,12 @@ if test "x$enable_gallium_llvm" = xyes; then
         AC_COMPUTE_INT([LLVM_VERSION_MINOR], [LLVM_VERSION_MINOR],
             [#include "${LLVM_INCLUDEDIR}/llvm/Config/llvm-config.h"])
 
+        dnl In LLVM 3.4.1 patch level was defined in config.h and not
+        dnl llvm-config.h
+        AC_COMPUTE_INT([LLVM_VERSION_PATCH], [LLVM_VERSION_PATCH],
+            [#include "${LLVM_INCLUDEDIR}/llvm/Config/config.h"],
+            LLVM_VERSION_PATCH=0) dnl Default if LLVM_VERSION_PATCH not found
+
         if test "x${LLVM_VERSION_MAJOR}" != x; then
             LLVM_VERSION_INT="${LLVM_VERSION_MAJOR}0${LLVM_VERSION_MINOR}"
         else
@@ -1610,7 +1640,7 @@ if test "x$enable_gallium_llvm" = xyes; then
                 LLVM_COMPONENTS="${LLVM_COMPONENTS} option"
             fi
         fi
-        DEFINES="${DEFINES} -DHAVE_LLVM=0x0$LLVM_VERSION_INT"
+        DEFINES="${DEFINES} -DHAVE_LLVM=0x0$LLVM_VERSION_INT -DLLVM_VERSION_PATCH=$LLVM_VERSION_PATCH"
         MESA_LLVM=1
 
         dnl Check for Clang internal headers
@@ -1629,6 +1659,10 @@ if test "x$enable_gallium_llvm" = xyes; then
 else
     MESA_LLVM=0
     LLVM_VERSION_INT=0
+
+    if test "x$enable_opencl" = xyes; then
+        AC_MSG_ERROR([cannot enable OpenCL without LLVM])
+    fi
 fi
 
 dnl Directory for XVMC libs
@@ -1702,8 +1736,9 @@ gallium_require_llvm() {
 
 gallium_require_drm_loader() {
     if test "x$enable_gallium_loader" = xyes; then
-        PKG_CHECK_MODULES([LIBUDEV], [libudev], [],
-                          AC_MSG_ERROR([Gallium drm loader requires libudev]))
+        if test "x$need_libudev$have_libudev" = xyesno; then
+            AC_MSG_ERROR([Gallium drm loader requires libudev >= $LIBUDEV_REQUIRED])
+        fi
         if test "x$have_libdrm" != xyes; then
             AC_MSG_ERROR([Gallium drm loader requires libdrm >= $LIBDRM_REQUIRED])
         fi

docs/relnotes/10.1.1.html

@@ -0,0 +1,254 @@
+<!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 10.1.1 Release Notes / (April 18, 2014)</h1>
+
+<p>
+Mesa 10.1.1 is a bug fix release which fixes bugs found since the 10.1 release.
+</p>
+<p>
+Mesa 10.1.1 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 are not supported.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+96e63674ccfa98e7ec6eb4fee3f770c3  MesaLib-10.1.1.tar.gz
+1fde7ed079df7aeb9b6a744ca033de8d  MesaLib-10.1.1.tar.bz2
+e64d0a562638664b13d2edf22321df59  MesaLib-10.1.1.zip
+</pre>
+
+
+<h2>New features</h2>
+<p>None</p>
+
+<h2>Bug fixes</h2>
+
+<ul>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=71547">Bug 71547</a> - compilation failure :#error &quot;SSE4.1 instruction set not enabled&quot;</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=74868">Bug 74868</a> - r600g: Diablo III Crashes After a few minutes</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=74988">Bug 74988</a> - Buffer overrun (segfault) decompressing ETC2 texture in GLBenchmark 3.0 Manhattan</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=75279">Bug 75279</a> - XCloseDisplay() takes one minute around nouveau_dri.so, freezing Firefox startup</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=75543">Bug 75543</a> - OSMesa Gallium OSMesaMakeCurrent</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=75660">Bug 75660</a> - u_inlines.h:277:pipe_buffer_map_range: Assertion `length' failed.</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=76323">Bug 76323</a> - GLSL compiler ignores layout(binding=N) on uniform blocks</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=76377">Bug 76377</a> - DRI3 should only be enabled on Linux due to a udev dependency</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=76749">Bug 76749</a> - [HSW] DOTA world lighting has no effect</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=77102">Bug 77102</a> - gallium nouveau has no profile in vdpau and libva</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=77207">Bug 77207</a> - [ivb/hsw] batch overwritten with garbage</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>Aaron Watry (1):</p>
+<ul>
+  <li>gallium/util: Fix memory leak</li>
+</ul>
+
+<p>Alexander von Gluck IV (1):</p>
+<ul>
+  <li>haiku: Fix build through scons corrections and viewport fixes</li>
+</ul>
+
+<p>Anuj Phogat (2):</p>
+<ul>
+  <li>mesa: Set initial internal format of a texture to GL_RGBA</li>
+  <li>mesa: Allow GL_DEPTH_COMPONENT and GL_DEPTH_STENCIL combinations in glTexImage{123}D()</li>
+</ul>
+
+<p>Brian Paul (12):</p>
+<ul>
+  <li>softpipe: use 64-bit arithmetic in softpipe_resource_layout()</li>
+  <li>mesa: don't call ctx-&gt;Driver.ClearBufferSubData() if size==0</li>
+  <li>st/osmesa: check buffer size when searching for buffers</li>
+  <li>mesa: fix copy &amp; paste bugs in pack_ubyte_SARGB8()</li>
+  <li>mesa: fix copy &amp; paste bugs in pack_ubyte_SRGB8()</li>
+  <li>c11/threads: don't include assert.h if the assert macro is already defined</li>
+  <li>mesa: fix unpack_Z32_FLOAT_X24S8() / unpack_Z32_FLOAT() mix-up</li>
+  <li>st/mesa: add null pointer checking in query object functions</li>
+  <li>mesa: fix glMultiDrawArrays inside a display list</li>
+  <li>cso: fix sampler view count in cso_set_sampler_views()</li>
+  <li>svga: replace sampler assertion with conditional</li>
+  <li>svga: move LIST_INITHEAD(dirty_buffers) earlier in svga_context_create()</li>
+</ul>
+
+<p>Carl Worth (3):</p>
+<ul>
+  <li>cherry-ignore: Ignore a few patches</li>
+  <li>glsl: Allow explicit binding on atomics again</li>
+  <li>Update VERSION to 10.1.1</li>
+</ul>
+
+<p>Chia-I Wu (1):</p>
+<ul>
+  <li>i965/vec4: fix record clearing in copy propagation</li>
+</ul>
+
+<p>Christian König (2):</p>
+<ul>
+  <li>st/mesa: recreate sampler view on context change v3</li>
+  <li>st/mesa: fix sampler view handling with shared textures v4</li>
+</ul>
+
+<p>Courtney Goeltzenleuchter (1):</p>
+<ul>
+  <li>mesa: add bounds checking to eliminate buffer overrun</li>
+</ul>
+
+<p>Emil Velikov (5):</p>
+<ul>
+  <li>nv50: add missing brackets when handling the samplers array</li>
+  <li>mesa: return v.value_int64 when the requested type is TYPE_INT64</li>
+  <li>configure: enable dri3 only for linux</li>
+  <li>glx: drop obsolete _XUnlock_Mutex in __glXInitialize error path</li>
+  <li>configure: cleanup libudev handling</li>
+</ul>
+
+<p>Eric Anholt (1):</p>
+<ul>
+  <li>i965: Fix buffer overruns in MSAA MCS buffer clearing.</li>
+</ul>
+
+<p>Hans (2):</p>
+<ul>
+  <li>util: don't define isfinite(), isnan() for MSVC &gt;= 1800</li>
+  <li>mesa: don't define c99 math functions for MSVC &gt;= 1800</li>
+</ul>
+
+<p>Ian Romanick (7):</p>
+<ul>
+  <li>linker: Split set_uniform_binding into separate functions for blocks and samplers</li>
+  <li>linker: Various trivial clean-ups in set_sampler_binding</li>
+  <li>linker: Fold set_uniform_binding into call site</li>
+  <li>linker: Clean up "unused parameter" warnings</li>
+  <li>linker: Set block bindings based on UniformBlocks rather than UniformStorage</li>
+  <li>linker: Set binding for all elements of UBO array</li>
+  <li>glsl: Propagate explicit binding information from the AST all the way to the linker</li>
+</ul>
+
+<p>Ilia Mirkin (8):</p>
+<ul>
+  <li>nouveau: fix fence waiting logic in screen destroy</li>
+  <li>nv50: adjust blit_3d handling of ms output textures</li>
+  <li>loader: add special logic to distinguish nouveau from nouveau_vieux</li>
+  <li>mesa/main: condition GL_DEPTH_STENCIL on ARB_depth_texture</li>
+  <li>nouveau: add forgotten GL_COMPRESSED_INTENSITY to texture format list</li>
+  <li>nouveau: there may not have been a texture if the fbo was incomplete</li>
+  <li>nvc0/ir: move sample id to second source arg to fix sampler2DMS</li>
+  <li>nouveau: fix firmware check on nvd7/nvd9</li>
+</ul>
+
+<p>Johannes Nixdorf (1):</p>
+<ul>
+  <li>configure.ac: fix the detection of expat with pkg-config</li>
+</ul>
+
+<p>Jonathan Gray (7):</p>
+<ul>
+  <li>gallium: add endian detection for OpenBSD</li>
+  <li>loader: use 0 instead of FALSE which isn't defined</li>
+  <li>loader: don't limit the non-udev path to only android</li>
+  <li>megadriver_stub.c: don't use _GNU_SOURCE to gate the compat code</li>
+  <li>egl/dri2: don't require libudev to build drm/wayland platforms</li>
+  <li>egl/dri2: use drm macros to construct device name</li>
+  <li>configure: don't require libudev for gbm or egl drm/wayland</li>
+</ul>
+
+<p>José Fonseca (4):</p>
+<ul>
+  <li>c11/threads: Fix nano to milisecond conversion.</li>
+  <li>mapi/u_thread: Use GetCurrentThreadId</li>
+  <li>c11/threads: Don't implement thrd_current on Windows.</li>
+  <li>draw: Duplicate TGSI tokens in draw_pipe_pstipple module.</li>
+</ul>
+
+<p>Kenneth Graunke (4):</p>
+<ul>
+  <li>i965/fs: Fix register comparisons in saturate propagation.</li>
+  <li>glsl: Fix lack of i2u in lower_ubo_reference.</li>
+  <li>i965: Stop advertising GL_MESA_ycbcr_texture.</li>
+  <li>glsl: Try vectorizing when seeing a repeated assignment to a channel.</li>
+</ul>
+
+<p>Marek Olšák (13):</p>
+<ul>
+  <li>r600g: fix texelFetchOffset GLSL functions</li>
+  <li>r600g: fix blitting the last 2 mipmap levels for Evergreen</li>
+  <li>mesa: fix the format of glEdgeFlagPointer</li>
+  <li>r600g,radeonsi: fix MAX_TEXTURE_3D_LEVELS and MAX_TEXTURE_ARRAY_LAYERS limits</li>
+  <li>st/mesa: fix per-vertex edge flags and GLSL support (v2)</li>
+  <li>mesa: mark GL_RGB9_E5 as not color-renderable</li>
+  <li>mesa: fix texture border handling for cube arrays</li>
+  <li>mesa: allow generating mipmaps for cube arrays</li>
+  <li>mesa: fix software fallback for generating mipmaps for cube arrays</li>
+  <li>mesa: fix software fallback for generating mipmaps for 3D textures</li>
+  <li>st/mesa: fix generating mipmaps for cube arrays</li>
+  <li>st/mesa: drop the lowering of quad strips to triangle strips</li>
+  <li>r600g: implement edge flags</li>
+</ul>
+
+<p>Matt Turner (4):</p>
+<ul>
+  <li>mesa: Wrap SSE4.1 code in #ifdef __SSE4_1__.</li>
+  <li>i965/fs: Fix off-by-one in saturate propagation.</li>
+  <li>i965/fs: Don't propagate saturate modifiers into partial writes.</li>
+  <li>i965/fs: Don't propagate saturation modifiers if there are source modifiers.</li>
+</ul>
+
+<p>Michel Dänzer (1):</p>
+<ul>
+  <li>r600g: Don't leak bytecode on shader compile failure</li>
+</ul>
+
+<p>Mike Stroyan (1):</p>
+<ul>
+  <li>i965: Avoid dependency hints on math opcodes</li>
+</ul>
+
+<p>Thomas Hellstrom (5):</p>
+<ul>
+  <li>winsys/svga: Replace the query mm buffer pool with a slab pool v3</li>
+  <li>winsys/svga: Update the vmwgfx_drm.h header to latest version from kernel</li>
+  <li>winsys/svga: Fix prime surface references also for guest-backed surfaces</li>
+  <li>st/xa: Bind destination before setting new state</li>
+  <li>st/xa: Make sure unused samplers are set to NULL</li>
+</ul>
+
+<p>Tom Stellard (1):</p>
+<ul>
+  <li>configure: Use LLVM shared libraries by default</li>
+</ul>
+
+</div>
+</body>
+</html>

docs/relnotes/10.1.2.html

@@ -0,0 +1,179 @@
+<!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 10.1.2 Release Notes / (May 5, 2014)</h1>
+
+<p>
+Mesa 10.1.2 is a bug fix release which fixes bugs found since the 10.1.1 release.
+</p>
+<p>
+Mesa 10.1.2 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 are not supported.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+37d79f94b1f41852a89d1fc3900bea76  MesaLib-10.1.2.tar.gz
+28b60d15ac9f364da1e0155911eaf44e  MesaLib-10.1.2.tar.bz2
+05300039085a65fc53c5472c4bb5747a  MesaLib-10.1.2.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=27499">Bug 27499</a> - [855GM i915] GL_LINE_STIPPLE displays incorrect colors</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=75723">Bug 75723</a> - (regression since Linux 3.14?) brw_get_graphics_reset_status: Assertion `brw-&gt;hw_ctx != ((void *)0)' failed</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=76894">Bug 76894</a> - Piglit/spec/EXT_framebuffer_object/fbo-bind-renderbuffer failed</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=77702">Bug 77702</a> - [i965 Bisected]Piglit spec/NV_conditional_render_blitframebuffer fails</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>Ander Conselvan de Oliveira (2):</p>
+<ul>
+  <li>gbm/dri: Fix out-of-memory error path in dri_device_create()</li>
+  <li>egl: Protect use of gbm_dri with ifdef HAVE_DRM_PLATFORM</li>
+</ul>
+
+<p>Anuj Phogat (27):</p>
+<ul>
+  <li>mesa: Fix glGetVertexAttribi(GL_VERTEX_ATTRIB_ARRAY_SIZE)</li>
+  <li>swrast: Add glBlitFramebuffer to commands affected by conditional rendering</li>
+  <li>mesa: Fix error condition for multisample proxy texture targets</li>
+  <li>i965: Put an assertion to check valid varying_to_slot[varying]</li>
+  <li>i965: Fix component mask and varying_to_slot mapping for gl_Layer</li>
+  <li>i965: Fix component mask and varying_to_slot mapping for gl_ViewportIndex</li>
+  <li>mesa: Add helper function _mesa_is_format_integer()</li>
+  <li>mesa: Add error condition for integer formats in glGetTexImage()</li>
+  <li>mesa: Add an error condition in glGetFramebufferAttachmentParameteriv()</li>
+  <li>mesa: Fix error code generation in glReadPixels()</li>
+  <li>glsl: Allow overlapping locations for vertex input attributes</li>
+  <li>mesa: Fix querying location of nth element of an array variable</li>
+  <li>mesa: Use location VERT_ATTRIB_GENERIC0 for vertex attribute 0</li>
+  <li>glsl: Compile error if fs defines conflicting qualifiers for gl_FragCoord</li>
+  <li>glsl: Compile error if fs uses gl_FragCoord before first redeclaration</li>
+  <li>mesa: Add entry for extension ARB_texture_stencil8</li>
+  <li>mesa: Add error condition for format=STENCIL_INDEX in glGetTexImage()</li>
+  <li>i965: Fix crash in do_blit_readpixels()</li>
+  <li>mesa: Add missing types in _mesa_texstore_xx_xx() functions</li>
+  <li>mesa: Allow srcFormat=GL_DEPTH_STENCIL in _mesa_texstore_xx_xx() functions</li>
+  <li>mesa: Add new helper function _mesa_unpack_depth_stencil_row()</li>
+  <li>mesa: Add support to unpack depth-stencil texture in to FLOAT_32_UNSIGNED_INT_24_8_REV</li>
+  <li>mesa: Allow FLOAT_32_UNSIGNED_INT_24_8_REV in get_tex_depth_stencil()</li>
+  <li>i965: Add glBlitFramebuffer to commands affected by conditional rendering</li>
+  <li>glsl: Use switch to allow adding more shader types</li>
+  <li>glsl: Link error if fs defines conflicting qualifiers for gl_FragCoord</li>
+  <li>glsl: Apply the link error conditions to GL_ARB_fragment_coord_conventions</li>
+</ul>
+
+<p>Benjamin Bellec (1):</p>
+<ul>
+  <li>mesa: fix GetStringi error message with correct function name</li>
+</ul>
+
+<p>Brian Paul (1):</p>
+<ul>
+  <li>swrast: allocate swrast_texture_image::ImageSlices array if needed</li>
+</ul>
+
+<p>Carl Worth (4):</p>
+<ul>
+  <li>docs: Add the MD5 sums for the 10.1.1 release tar files.</li>
+  <li>cherry-ignore: Ignore a patch causing a regression</li>
+  <li>cherry-ignore: Drop an ignored patch now that piglit has been updated.</li>
+  <li>Update VERSION to 10.1.2</li>
+</ul>
+
+<p>Chris Forbes (1):</p>
+<ul>
+  <li>glsl: Only allow `invariant` on shader in/out between stages.</li>
+</ul>
+
+<p>Eric Anholt (1):</p>
+<ul>
+  <li>i965: Fix render-to-texture in non-FinishRenderTexture cases.</li>
+</ul>
+
+<p>Ian Romanick (1):</p>
+<ul>
+  <li>dri3: Enable GLX_MESA_query_renderer on DRI3 too</li>
+</ul>
+
+<p>Kenneth Graunke (2):</p>
+<ul>
+  <li>i965: Don't enable reset notification support on Gen4-5.</li>
+  <li>i965: Actually emit PIPELINE_SELECT and 3DSTATE_VF_STATISTICS.</li>
+</ul>
+
+<p>Marek Olšák (10):</p>
+<ul>
+  <li>r300g: don't crash when getting NULL colorbuffers</li>
+  <li>st/mesa: remove trailing NULL colorbuffers</li>
+  <li>r600g: fix edge flags and layered rendering on R600-R700</li>
+  <li>r600g: disable async DMA on R700</li>
+  <li>r600g: fix MSAA resolve on R6xx when the destination is 1D-tiled</li>
+  <li>r600g: fix flushing on RV670, RS780, RS880 again</li>
+  <li>r600g: fix buffer copying on R600-R700</li>
+  <li>r600g: fix for broken CULL_FRONT behavior on R6xx</li>
+  <li>r600g: fix for an MSAA hang on RV770</li>
+  <li>r600g: fix hang on RV740 by using DX_RASTERIZATION_KILL instead of SX_MISC</li>
+</ul>
+
+<p>Michel Dänzer (2):</p>
+<ul>
+  <li>r600g: Disable LLVM by default at runtime for graphics</li>
+  <li>st/mesa: Fix NULL pointer dereference for incomplete framebuffers</li>
+</ul>
+
+<p>Neil Roberts (1):</p>
+<ul>
+  <li>wayland: Fix the logic in disabling the prime capability</li>
+</ul>
+
+<p>Samuel Iglesias Gonsalvez (1):</p>
+<ul>
+  <li>mesa: fix check for dummy renderbuffer in _mesa_FramebufferRenderbufferEXT()</li>
+</ul>
+
+<p>Thomas Hellstrom (1):</p>
+<ul>
+  <li>st/xa: Cache render target surface</li>
+</ul>
+
+<p>nick (1):</p>
+<ul>
+  <li>swrast: Fix vertex color in _swsetup_Translate()</li>
+</ul>
+
+</div>
+</body>
+</html>

docs/relnotes/10.1.3.html

@@ -0,0 +1,90 @@
+<!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 10.1.3 Release Notes / (May 9, 2014)</h1>
+
+<p>
+Mesa 10.1.3 is a bug fix release which fixes bugs found since the 10.1.2 release.
+</p>
+
+<p>
+Note: Mesa 10.1.3 is being released sooner than originally scheduled to make
+available a fix for a performance rgression that was inadvertently introduced
+to Mesa 10.1.2. The performance regression is reported to make vmware
+swapbuffers fall back to software. 
+</p>
+
+<p>
+Mesa 10.1.3 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 are not supported.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+665fe1656aaa2c37b32042068aff92cb  MesaLib-10.1.3.tar.gz
+ba6dbe2b9cab0b4de840c996b9b6a3ad  MesaLib-10.1.3.tar.bz2
+4e6f26330a63d3c47e62ac4bdead39e8  MesaLib-10.1.3.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=77245">Bug 77245</a> - Bogus GL_ARB_explicit_attrib_location layout identifier warnings</li>
+
+</ul>
+
+
+<h2>Changes</h2>
+
+<p>Carl Worth (3):</p>
+<ul>
+  <li>docs: Add MD5 sums for Mesa 10.1.2</li>
+  <li>get-pick-list.sh: Require explicit "10.1" for nominating stable patches</li>
+  <li>VERSION: Update to 10.1.3</li>
+</ul>
+
+<p>Kenneth Graunke (2):</p>
+<ul>
+  <li>mesa: Fix MaxNumLayers for 1D array textures.</li>
+  <li>i965: Fix depth (array slices) computation for 1D_ARRAY render targets.</li>
+</ul>
+
+<p>Tapani Pälli (1):</p>
+<ul>
+  <li>glsl: fix bogus layout qualifier warnings</li>
+</ul>
+
+<p>Thomas Hellstrom (1):</p>
+<ul>
+  <li>st/xa: Fix performance regression introduced by commit "Cache render target surface"</li>
+</ul>
+
+
+</div>
+</body>
+</html>

docs/relnotes/10.1.4.html

@@ -0,0 +1,100 @@
+<!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 10.1.4 Release Notes / (May 20, 2014)</h1>
+
+<p>
+Mesa 10.1.4 is a bug fix release which fixes bugs found since the 10.1.3 release.
+</p>
+
+<p>
+Mesa 10.1.4 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 are not supported.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+e934365d77f384bfaec844999440bef8  MesaLib-10.1.4.tar.gz
+6fddee101f49b7409cd29994c34ddee7  MesaLib-10.1.4.tar.bz2
+ba5f48e7d5e373922c804c2651fec6c1  MesaLib-10.1.4.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=78225">Bug 78225</a> - Compile error due to undefined reference to `gbm_dri_backend', fix attached</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=78537">Bug 78537</a> - no anisotropic filtering in a native Half-Life 2</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>Brian Paul (1):</p>
+<ul>
+  <li>mesa: fix double-freeing of dispatch tables inside glBegin/End.</li>
+</ul>
+
+<p>Carl Worth (3):</p>
+<ul>
+  <li>docs: Add MD5 sums for 10.1.3</li>
+  <li>cherry-ignore: Roland and Michel agreed to drop these patches.</li>
+  <li>VERSION: Update to 10.1.4</li>
+</ul>
+
+<p>Emil Velikov (1):</p>
+<ul>
+  <li>configure: error out if building GBM without dri</li>
+</ul>
+
+<p>Eric Anholt (1):</p>
+<ul>
+  <li>i965/vs: Use samplers for UBOs in the VS like we do for non-UBO pulls.</li>
+</ul>
+
+<p>Ilia Mirkin (3):</p>
+<ul>
+  <li>nv50/ir: make sure to reverse cond codes on all the OP_SET variants</li>
+  <li>nv50: fix setting of texture ms info to be per-stage</li>
+  <li>nv50/ir: fix integer mul lowering for u32 x u32 -&gt; high u32</li>
+</ul>
+
+<p>Michel Dänzer (1):</p>
+<ul>
+  <li>radeonsi: Fix anisotropic filtering state setup</li>
+</ul>
+
+<p>Tom Stellard (2):</p>
+<ul>
+  <li>configure.ac: Add LLVM_VERSION_PATCH to DEFINES</li>
+  <li>radeonsi: Enable geometry shaders with LLVM 3.4.1</li>
+</ul>
+
+</div>
+</body>
+</html>

docs/relnotes/10.1.5.html

@@ -0,0 +1,105 @@
+<!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 10.1.5 Release Notes / (June 6, 2014)</h1>
+
+<p>
+Mesa 10.1.5 is a bug fix release which fixes bugs found since the 10.1.4 release.
+</p>
+
+<p>
+Mesa 10.1.5 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 are not supported.
+</p>
+
+<h2>SHA256 checksums</h2>
+<pre>
+b0aceaa75bc9a9b2d9215a113e2ad488b5cf85c99005a7624f8cf7c37c5d0eaa  MesaLib-10.1.5.tar.gz
+bc6c5ec7836f254a49d055a29d9aa34c97c54c038f47ad3a00fa57a5fef15bbc  MesaLib-10.1.5.tar.bz2
+78b7255cab0af7918945452a84de7989096ebcdd27e99b31c56c0589274cbc77  MesaLib-10.1.5.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=79115">Bug 79115</a> - </li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=79421">Bug 79421</a> - </li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>Brian Paul (1):</p>
+<ul>
+  <li>glsl: fix use-after free bug/crash in ast_declarator_list::hir()</li>
+</ul>
+
+<p>Carl Worth (5):</p>
+<ul>
+  <li>docs: Add md5sums for 10.1.4 release</li>
+  <li>Merge remote-tracking branch 'origin/10.1' into 10.1</li>
+  <li>cherry-ignore: Ignore two commits.</li>
+  <li>Ignore a patch that is not needed for the 10.1 branch.</li>
+  <li>Update version to 10.1.5</li>
+</ul>
+
+<p>Emil Velikov (1):</p>
+<ul>
+  <li>glx: do not leak dri3Display</li>
+</ul>
+
+<p>Ilia Mirkin (2):</p>
+<ul>
+  <li>nv50/ir: fix s32 x s32 -&gt; high s32 multiply logic</li>
+  <li>nv50/ir: fix constant folding for OP_MUL subop HIGH</li>
+</ul>
+
+<p>James Legg (1):</p>
+<ul>
+  <li>mesa: Fix unbinding GL_DEPTH_STENCIL_ATTACHMENT</li>
+</ul>
+
+<p>Jeremy Huddleston Sequoia (2):</p>
+<ul>
+  <li>glapi: Avoid heap corruption in _glapi_table</li>
+  <li>darwin: Fix test for kCGLPFAOpenGLProfile support at runtime</li>
+</ul>
+
+<p>Pavel Popov (2):</p>
+<ul>
+  <li>i965: Properly return *RESET* status in glGetGraphicsResetStatusARB</li>
+  <li>i965: Fix Line Stipple enable bit in 3DSTATE_SF for Haswell.</li>
+</ul>
+
+<p>Roland Scheidegger (1):</p>
+<ul>
+  <li>llvmpipe: fix crash when not all attachments are populated in a fb</li>
+</ul>
+
+</div>
+</body>
+</html>

docs/relnotes/10.1.6.html

@@ -0,0 +1,135 @@
+<!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 10.1.6 Release Notes / (June 24, 2014)</h1>
+
+<p>
+Mesa 10.1.6 is a bug fix release which fixes bugs found since the 10.1.5 release.
+</p>
+
+<p>
+Mesa 10.1.6 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 are not supported.
+</p>
+
+<h2>SHA256 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=54372">Bug 54372</a> - GLX_INTEL_swap_event crashes driver when swapping window buffers</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=74005">Bug 74005</a> - [i965 Bisected]Piglit/glx_glx-make-glxdrawable-current fails</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=78581">Bug 78581</a> - </li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=79729">Bug 79729</a> - [i965] glClear on a multisample texture doesn't work</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>Adrian Negreanu (7):</p>
+<ul>
+  <li>add megadriver_stub_FILES</li>
+  <li>android: adapt to the megadriver mechanism</li>
+  <li>android: add libloader to libGLES_mesa and libmesa_egl_dri2</li>
+  <li>android: add src/gallium/auxiliary as include path for libmesa_dricore</li>
+  <li>android, egl: add correct drm include for libmesa_egl_dri2</li>
+  <li>android, mesa_gen_matypes: pull in timespec POSIX definition</li>
+  <li>android, dricore: undefined reference to _mesa_streaming_load_memcpy</li>
+</ul>
+
+<p>Beren Minor (1):</p>
+<ul>
+  <li>egl/main: Fix eglMakeCurrent when releasing context from current thread.</li>
+</ul>
+
+<p>Carl Worth (3):</p>
+<ul>
+  <li>docs: Add SHA256 checksums for the 10.1.5 release</li>
+  <li>cherry-ignore: Add a patch to ignore</li>
+  <li>Update VERSION to 10.1.6</li>
+</ul>
+
+<p>Daniel Manjarres (1):</p>
+<ul>
+  <li>glx: Don't crash on swap event for a Window (non-GLXWindow)</li>
+</ul>
+
+<p>Emil Velikov (1):</p>
+<ul>
+  <li>configure: error out when building opencl without LLVM</li>
+</ul>
+
+<p>Iago Toral Quiroga (1):</p>
+<ul>
+  <li>mesa: Copy Geom.UsesEndPrimitive when cloning a geometry program.</li>
+</ul>
+
+<p>José Fonseca (3):</p>
+<ul>
+  <li>mesa/main: Make get_hash.c values constant.</li>
+  <li>mesa: Make glGetIntegerv(GL_*_ARRAY_SIZE) return GL_BGRA.</li>
+  <li>mesa/main: Prevent sefgault on glGetIntegerv(GL_ATOMIC_COUNTER_BUFFER_BINDING).</li>
+</ul>
+
+<p>Kristian Høgsberg (1):</p>
+<ul>
+  <li>mesa: Remove glClear optimization based on drawable size</li>
+</ul>
+
+<p>Michel Dänzer (1):</p>
+<ul>
+  <li>configure: Only check for OpenCL without LLVM when the latter is certain</li>
+</ul>
+
+<p>Neil Roberts (1):</p>
+<ul>
+  <li>i965: Set the fast clear color value for texture surfaces</li>
+</ul>
+
+<p>Roland Scheidegger (1):</p>
+<ul>
+  <li>draw: (trivial) fix clamping of viewport index</li>
+</ul>
+
+<p>Tobias Klausmann (1):</p>
+<ul>
+  <li>nv50/ir: clear subop when folding constant expressions</li>
+</ul>
+
+<p>Tom Stellard (2):</p>
+<ul>
+  <li>clover: Prevent Clang from printing number of errors and warnings to stderr.</li>
+  <li>clover: Don't use llvm's global context</li>
+</ul>
+
+</div>
+</body>
+</html>

include/c11/threads_posix.h

@@ -27,7 +27,9 @@
  * DEALINGS IN THE SOFTWARE.
  */
 #include <stdlib.h>
+#ifndef assert
 #include <assert.h>
+#endif
 #include <limits.h>
 #include <errno.h>
 #include <unistd.h>

include/c11/threads_win32.h

@@ -26,7 +26,9 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  */
+#ifndef assert
 #include <assert.h>
+#endif
 #include <limits.h>
 #include <errno.h>
 #include <process.h>  // MSVCRT
@@ -146,7 +148,7 @@ static unsigned __stdcall impl_thrd_routine(void *p)
 
 static DWORD impl_xtime2msec(const xtime *xt)
 {
-    return (DWORD)((xt->sec * 1000u) + (xt->nsec / 1000));
+    return (DWORD)((xt->sec * 1000U) + (xt->nsec / 1000000L));
 }
 
 #ifdef EMULATED_THREADS_USE_NATIVE_CALL_ONCE
@@ -492,12 +494,42 @@ thrd_create(thrd_t *thr, thrd_start_t func, void *arg)
     return thrd_success;
 }
 
+#if 0
 // 7.25.5.2
 static inline thrd_t
 thrd_current(void)
 {
-    return GetCurrentThread();
+    HANDLE hCurrentThread;
+    BOOL bRet;
+
+    /* GetCurrentThread() returns a pseudo-handle, which is useless.  We need
+     * to call DuplicateHandle to get a real handle.  However the handle value
+     * will not match the one returned by thread_create.
+     *
+     * Other potential solutions would be:
+     * - define thrd_t as a thread Ids, but this would mean we'd need to OpenThread for many operations
+     * - use malloc'ed memory for thrd_t. This would imply using TLS for current thread.
+     *
+     * Neither is particularly nice.
+     *
+     * Life would be much easier if C11 threads had different abstractions for
+     * threads and thread IDs, just like C++11 threads does...
+     */
+
+    bRet = DuplicateHandle(GetCurrentProcess(), // source process (pseudo) handle
+                           GetCurrentThread(), // source (pseudo) handle
+                           GetCurrentProcess(), // target process
+                           &hCurrentThread, // target handle
+                           0,
+                           FALSE,
+                           DUPLICATE_SAME_ACCESS);
+    assert(bRet);
+    if (!bRet) {
+	hCurrentThread = GetCurrentThread();
+    }
+    return hCurrentThread;
 }
+#endif
 
 // 7.25.5.3
 static inline int
@@ -511,7 +543,7 @@ thrd_detach(thrd_t thr)
 static inline int
 thrd_equal(thrd_t thr0, thrd_t thr1)
 {
-    return (thr0 == thr1);
+    return GetThreadId(thr0) == GetThreadId(thr1);
 }
 
 // 7.25.5.5

scons/gallium.py

@@ -269,6 +269,11 @@ def generate(env):
             cppdefines += ['HAVE_ALIAS']
         else:
             cppdefines += ['GLX_ALIAS_UNSUPPORTED']
+    if env['platform'] == 'haiku':
+        cppdefines += [
+            'HAVE_PTHREAD',
+            'HAVE_POSIX_MEMALIGN'
+        ]
     if platform == 'windows':
         cppdefines += [
             'WIN32',

src/egl/drivers/dri2/Android.mk

@@ -40,8 +40,12 @@ LOCAL_C_INCLUDES := \
 	$(MESA_TOP)/src/mapi \
 	$(MESA_TOP)/src/egl/main \
 	$(MESA_TOP)/src/loader \
+	$(DRM_TOP)/include/drm \
 	$(DRM_GRALLOC_TOP)
 
+LOCAL_STATIC_LIBRARIES := \
+	libloader
+
 LOCAL_MODULE := libmesa_egl_dri2
 
 include $(MESA_COMMON_MK)

src/egl/drivers/dri2/egl_dri2.c

@@ -626,7 +626,6 @@ dri2_initialize(_EGLDriver *drv, _EGLDisplay *disp)
       return dri2_initialize_x11(drv, disp);
 #endif
 
-#ifdef HAVE_LIBUDEV
 #ifdef HAVE_DRM_PLATFORM
    case _EGL_PLATFORM_DRM:
       if (disp->Options.TestOnly)
@@ -639,7 +638,6 @@ dri2_initialize(_EGLDriver *drv, _EGLDisplay *disp)
          return EGL_TRUE;
       return dri2_initialize_wayland(drv, disp);
 #endif
-#endif
 #ifdef HAVE_ANDROID_PLATFORM
    case _EGL_PLATFORM_ANDROID:
       if (disp->Options.TestOnly)
@@ -1894,10 +1892,12 @@ dri2_bind_wayland_display_wl(_EGLDriver *drv, _EGLDisplay *disp,
    if (!dri2_dpy->wl_server_drm)
 	   return EGL_FALSE;
 
+#ifdef HAVE_DRM_PLATFORM
    /* We have to share the wl_drm instance with gbm, so gbm can convert
     * wl_buffers to gbm bos. */
    if (dri2_dpy->gbm_dri)
       dri2_dpy->gbm_dri->wl_drm = dri2_dpy->wl_server_drm;
+#endif
 
    return EGL_TRUE;
 }

src/egl/drivers/dri2/platform_drm.c

@@ -458,7 +458,12 @@ dri2_initialize_drm(_EGLDriver *drv, _EGLDisplay *disp)
 
    gbm = disp->PlatformDisplay;
    if (gbm == NULL) {
-      fd = open("/dev/dri/card0", O_RDWR);
+      char buf[64];
+      int n = snprintf(buf, sizeof(buf), DRM_DEV_NAME, DRM_DIR_NAME, 0);
+      if (n != -1 && n < sizeof(buf))
+         fd = open(buf, O_RDWR);
+      if (fd < 0)
+         fd = open("/dev/dri/card0", O_RDWR);
       dri2_dpy->own_device = 1;
       gbm = gbm_create_device(fd);
       if (gbm == NULL)

src/egl/drivers/dri2/platform_wayland.c

@@ -1045,7 +1045,7 @@ dri2_initialize_wayland(_EGLDriver *drv, _EGLDisplay *disp)
 
    if (dri2_dpy->image->base.version < 7 ||
        dri2_dpy->image->createImageFromFds == NULL)
-      dri2_dpy->capabilities &= WL_DRM_CAPABILITY_PRIME;
+      dri2_dpy->capabilities &= ~WL_DRM_CAPABILITY_PRIME;
 
    types = EGL_WINDOW_BIT;
    for (i = 0; dri2_dpy->driver_configs[i]; i++) {

src/egl/main/Android.mk

@@ -154,11 +154,14 @@ LOCAL_STATIC_LIBRARIES := \
 	libmesa_glsl \
 	libmesa_glsl_utils \
 	libmesa_gallium \
-	libloader \
 	$(LOCAL_STATIC_LIBRARIES)
 
 endif # MESA_BUILD_GALLIUM
 
+LOCAL_STATIC_LIBRARIES := \
+	$(LOCAL_STATIC_LIBRARIES) \
+	libloader
+
 LOCAL_MODULE := libGLES_mesa
 LOCAL_MODULE_PATH := $(TARGET_OUT_SHARED_LIBRARIES)/egl
 

src/egl/main/eglapi.c

@@ -490,8 +490,12 @@ eglMakeCurrent(EGLDisplay dpy, EGLSurface draw, EGLSurface read,
    if (!context && ctx != EGL_NO_CONTEXT)
       RETURN_EGL_ERROR(disp, EGL_BAD_CONTEXT, EGL_FALSE);
    if (!draw_surf || !read_surf) {
-      /* surfaces may be NULL if surfaceless */
-      if (!disp->Extensions.KHR_surfaceless_context)
+      /* From the EGL 1.4 (20130211) spec:
+       *
+       *    To release the current context without assigning a new one, set ctx
+       *    to EGL_NO_CONTEXT and set draw and read to EGL_NO_SURFACE.
+       */
+      if (!disp->Extensions.KHR_surfaceless_context && ctx != EGL_NO_CONTEXT)
          RETURN_EGL_ERROR(disp, EGL_BAD_SURFACE, EGL_FALSE);
 
       if ((!draw_surf && draw != EGL_NO_SURFACE) ||

src/gallium/auxiliary/cso_cache/cso_context.c

@@ -1187,11 +1187,12 @@ cso_set_sampler_views(struct cso_context *ctx,
       pipe_sampler_view_reference(&info->views[i], NULL);
    }
 
-   info->nr_views = count;
-
    /* bind the new sampler views */
-   ctx->pipe->set_sampler_views(ctx->pipe, shader_stage, 0, count,
+   ctx->pipe->set_sampler_views(ctx->pipe, shader_stage, 0,
+                                MAX2(info->nr_views, count),
                                 info->views);
+
+   info->nr_views = count;
 }
 
 

src/gallium/auxiliary/draw/draw_pipe_pstipple.c

@@ -673,7 +673,7 @@ pstip_create_fs_state(struct pipe_context *pipe,
    struct pstip_fragment_shader *pstipfs = CALLOC_STRUCT(pstip_fragment_shader);
 
    if (pstipfs) {
-      pstipfs->state = *fs;
+      pstipfs->state.tokens = tgsi_dup_tokens(fs->tokens);
 
       /* pass-through */
       pstipfs->driver_fs = pstip->driver_create_fs_state(pstip->pipe, fs);
@@ -707,6 +707,7 @@ pstip_delete_fs_state(struct pipe_context *pipe, void *fs)
    if (pstipfs->pstip_fs)
       pstip->driver_delete_fs_state(pstip->pipe, pstipfs->pstip_fs);
 
+   FREE((void*)pstipfs->state.tokens);
    FREE(pstipfs);
 }
 

src/gallium/auxiliary/draw/draw_private.h

@@ -495,7 +495,7 @@ draw_stats_clipper_primitives(struct draw_context *draw,
 static INLINE unsigned
 draw_clamp_viewport_idx(int idx)
 {
-   return ((PIPE_MAX_VIEWPORTS > idx || idx < 0) ? idx : 0);
+   return ((PIPE_MAX_VIEWPORTS > idx && idx >= 0) ? idx : 0);
 }
 
 /**

src/gallium/auxiliary/util/u_blitter.c

@@ -364,6 +364,8 @@ void util_blitter_destroy(struct blitter_context *blitter)
    pipe->delete_vs_state(pipe, ctx->vs);
    if (ctx->vs_pos_only)
       pipe->delete_vs_state(pipe, ctx->vs_pos_only);
+   if (ctx->vs_layered)
+      pipe->delete_vs_state(pipe, ctx->vs_layered);
    pipe->delete_vertex_elements_state(pipe, ctx->velem_state);
    for (i = 0; i < 4; i++) {
       if (ctx->velem_state_readbuf[i]) {

src/gallium/auxiliary/util/u_gen_mipmap.c

@@ -1382,7 +1382,7 @@ get_next_slot(struct gen_mipmap_state *ctx)
 static unsigned
 set_vertex_data(struct gen_mipmap_state *ctx,
                 enum pipe_texture_target tex_target,
-                uint layer, float r)
+                uint face, float r)
 {
    unsigned offset;
 
@@ -1403,14 +1403,21 @@ set_vertex_data(struct gen_mipmap_state *ctx,
    ctx->vertices[3][0][1] = 1.0f;
 
    /* Setup vertex texcoords.  This is a little tricky for cube maps. */
-   if (tex_target == PIPE_TEXTURE_CUBE) {
+   if (tex_target == PIPE_TEXTURE_CUBE ||
+       tex_target == PIPE_TEXTURE_CUBE_ARRAY) {
       static const float st[4][2] = {
          {0.0f, 0.0f}, {1.0f, 0.0f}, {1.0f, 1.0f}, {0.0f, 1.0f}
       };
 
-      util_map_texcoords2d_onto_cubemap(layer, &st[0][0], 2,
+      util_map_texcoords2d_onto_cubemap(face, &st[0][0], 2,
                                         &ctx->vertices[0][1][0], 8,
                                         FALSE);
+
+      /* set the layer for cube arrays */
+      ctx->vertices[0][1][3] = r;
+      ctx->vertices[1][1][3] = r;
+      ctx->vertices[2][1][3] = r;
+      ctx->vertices[3][1][3] = r;
    }
    else if (tex_target == PIPE_TEXTURE_1D_ARRAY) {
       /* 1D texture array  */
@@ -1520,29 +1527,7 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
    assert(filter == PIPE_TEX_FILTER_LINEAR ||
           filter == PIPE_TEX_FILTER_NEAREST);
 
-   switch (pt->target) {
-   case PIPE_TEXTURE_1D:
-      type = TGSI_TEXTURE_1D;
-      break;
-   case PIPE_TEXTURE_2D:
-      type = TGSI_TEXTURE_2D;
-      break;
-   case PIPE_TEXTURE_3D:
-      type = TGSI_TEXTURE_3D;
-      break;
-   case PIPE_TEXTURE_CUBE:
-      type = TGSI_TEXTURE_CUBE;
-      break;
-   case PIPE_TEXTURE_1D_ARRAY:
-      type = TGSI_TEXTURE_1D_ARRAY;
-      break;
-   case PIPE_TEXTURE_2D_ARRAY:
-      type = TGSI_TEXTURE_2D_ARRAY;
-      break;
-   default:
-      assert(0);
-      type = TGSI_TEXTURE_2D;
-   }
+   type = util_pipe_tex_to_tgsi_tex(pt->target, 1);
 
    /* check if we can render in the texture's format */
    if (!screen->is_format_supported(screen, psv->format, pt->target,
@@ -1600,7 +1585,9 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
 
       if (pt->target == PIPE_TEXTURE_3D)
          nr_layers = u_minify(pt->depth0, dstLevel);
-      else if (pt->target == PIPE_TEXTURE_2D_ARRAY || pt->target == PIPE_TEXTURE_1D_ARRAY)
+      else if (pt->target == PIPE_TEXTURE_2D_ARRAY ||
+               pt->target == PIPE_TEXTURE_1D_ARRAY ||
+               pt->target == PIPE_TEXTURE_CUBE_ARRAY)
 	 nr_layers = pt->array_size;
       else
          nr_layers = 1;
@@ -1613,9 +1600,14 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
             layer = i;
             /* XXX hmm really? */
             rcoord = (float)layer / (float)nr_layers + 1.0f / (float)(nr_layers * 2);
-         } else if (pt->target == PIPE_TEXTURE_2D_ARRAY || pt->target == PIPE_TEXTURE_1D_ARRAY) {
+         } else if (pt->target == PIPE_TEXTURE_2D_ARRAY ||
+                    pt->target == PIPE_TEXTURE_1D_ARRAY) {
 	    layer = i;
 	    rcoord = (float)layer;
+         } else if (pt->target == PIPE_TEXTURE_CUBE_ARRAY) {
+            layer = i;
+            face = layer % 6;
+            rcoord = layer / 6;
 	 } else
             layer = face;
 

src/gallium/auxiliary/util/u_math.h

@@ -112,10 +112,13 @@ static INLINE float logf( float f )
 #define logf(x) ((float)log((double)(x)))
 #endif /* logf */
 
+#if _MSC_VER < 1800
 #define isfinite(x) _finite((double)(x))
 #define isnan(x) _isnan((double)(x))
+#endif /* _MSC_VER < 1800 */
 #endif /* _MSC_VER < 1400 && !defined(__cplusplus) */
 
+#if _MSC_VER < 1800
 static INLINE double log2( double x )
 {
    const double invln2 = 1.442695041;
@@ -133,6 +136,7 @@ roundf(float x)
 {
    return x >= 0.0f ? floorf(x + 0.5f) : ceilf(x - 0.5f);
 }
+#endif
 
 #define INFINITY (DBL_MAX + DBL_MAX)
 #define NAN (INFINITY - INFINITY)

src/gallium/drivers/llvmpipe/lp_rast.c

@@ -376,9 +376,15 @@ lp_rast_shade_tile(struct lp_rasterizer_task *task,
 
          /* color buffer */
          for (i = 0; i < scene->fb.nr_cbufs; i++){
-            stride[i] = scene->cbufs[i].stride;
-            color[i] = lp_rast_get_unswizzled_color_block_pointer(task, i, tile_x + x,
-                                                                  tile_y + y, inputs->layer);
+            if (scene->fb.cbufs[i]) {
+               stride[i] = scene->cbufs[i].stride;
+               color[i] = lp_rast_get_unswizzled_color_block_pointer(task, i, tile_x + x,
+                                                                     tile_y + y, inputs->layer);
+            }
+            else {
+               stride[i] = 0;
+               color[i] = NULL;
+            }
          }
 
          /* depth buffer */

src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp

@@ -37,18 +37,25 @@ namespace nv50_ir {
 //    ah*bl 00
 //
 // fffe0001 + fffe0001
+//
+// Note that this sort of splitting doesn't work for signed values, so we
+// compute the sign on those manually and then perform an unsigned multiply.
 static bool
 expandIntegerMUL(BuildUtil *bld, Instruction *mul)
 {
    const bool highResult = mul->subOp == NV50_IR_SUBOP_MUL_HIGH;
 
-   DataType fTy = mul->sType; // full type
-   DataType hTy;
+   DataType fTy; // full type
+   switch (mul->sType) {
+   case TYPE_S32: fTy = TYPE_U32; break;
+   case TYPE_S64: fTy = TYPE_U64; break;
+   default: fTy = mul->sType; break;
+   }
+
+   DataType hTy; // half type
    switch (fTy) {
-   case TYPE_S32: hTy = TYPE_S16; break;
    case TYPE_U32: hTy = TYPE_U16; break;
    case TYPE_U64: hTy = TYPE_U32; break;
-   case TYPE_S64: hTy = TYPE_S32; break;
    default:
       return false;
    }
@@ -59,15 +66,25 @@ expandIntegerMUL(BuildUtil *bld, Instruction *mul)
 
    bld->setPosition(mul, true);
 
+   Value *s[2];
    Value *a[2], *b[2];
-   Value *c[2];
    Value *t[4];
    for (int j = 0; j < 4; ++j)
       t[j] = bld->getSSA(fullSize);
 
+   s[0] = mul->getSrc(0);
+   s[1] = mul->getSrc(1);
+
+   if (isSignedType(mul->sType)) {
+      s[0] = bld->getSSA(fullSize);
+      s[1] = bld->getSSA(fullSize);
+      bld->mkOp1(OP_ABS, mul->sType, s[0], mul->getSrc(0));
+      bld->mkOp1(OP_ABS, mul->sType, s[1], mul->getSrc(1));
+   }
+
    // split sources into halves
-   i[0] = bld->mkSplit(a, halfSize, mul->getSrc(0));
-   i[1] = bld->mkSplit(b, halfSize, mul->getSrc(1));
+   i[0] = bld->mkSplit(a, halfSize, s[0]);
+   i[1] = bld->mkSplit(b, halfSize, s[1]);
 
    i[2] = bld->mkOp2(OP_MUL, fTy, t[0], a[0], b[1]);
    i[3] = bld->mkOp3(OP_MAD, fTy, t[1], a[1], b[0], t[0]);
@@ -75,23 +92,76 @@ expandIntegerMUL(BuildUtil *bld, Instruction *mul)
    i[4] = bld->mkOp3(OP_MAD, fTy, t[3], a[0], b[0], t[2]);
 
    if (highResult) {
-      Value *r[3];
+      Value *c[2];
+      Value *r[5];
       Value *imm = bld->loadImm(NULL, 1 << (halfSize * 8));
       c[0] = bld->getSSA(1, FILE_FLAGS);
       c[1] = bld->getSSA(1, FILE_FLAGS);
-      for (int j = 0; j < 3; ++j)
+      for (int j = 0; j < 5; ++j)
          r[j] = bld->getSSA(fullSize);
 
       i[8] = bld->mkOp2(OP_SHR, fTy, r[0], t[1], bld->mkImm(halfSize * 8));
       i[6] = bld->mkOp2(OP_ADD, fTy, r[1], r[0], imm);
-      bld->mkOp2(OP_UNION, TYPE_U32, r[2], r[1], r[0]);
-      i[5] = bld->mkOp3(OP_MAD, fTy, mul->getDef(0), a[1], b[1], r[2]);
+      bld->mkMov(r[3], r[0])->setPredicate(CC_NC, c[0]);
+      bld->mkOp2(OP_UNION, TYPE_U32, r[2], r[1], r[3]);
+      i[5] = bld->mkOp3(OP_MAD, fTy, r[4], a[1], b[1], r[2]);
 
       // set carry defs / sources
       i[3]->setFlagsDef(1, c[0]);
-      i[4]->setFlagsDef(0, c[1]); // actual result not required, just the carry
+      // actual result required in negative case, but ignored for
+      // unsigned. for some reason the compiler ends up dropping the whole
+      // instruction if the destination is unused but the flags are.
+      if (isSignedType(mul->sType))
+         i[4]->setFlagsDef(1, c[1]);
+      else
+         i[4]->setFlagsDef(0, c[1]);
       i[6]->setPredicate(CC_C, c[0]);
       i[5]->setFlagsSrc(3, c[1]);
+
+      if (isSignedType(mul->sType)) {
+         Value *cc[2];
+         Value *rr[7];
+         Value *one = bld->getSSA(fullSize);
+         bld->loadImm(one, 1);
+         for (int j = 0; j < 7; j++)
+            rr[j] = bld->getSSA(fullSize);
+
+         // NOTE: this logic uses predicates because splitting basic blocks is
+         // ~impossible during the SSA phase. The RA relies on a correlation
+         // between edge order and phi node sources.
+
+         // Set the sign of the result based on the inputs
+         bld->mkOp2(OP_XOR, fTy, NULL, mul->getSrc(0), mul->getSrc(1))
+            ->setFlagsDef(0, (cc[0] = bld->getSSA(1, FILE_FLAGS)));
+
+         // 1s complement of 64-bit value
+         bld->mkOp1(OP_NOT, fTy, rr[0], r[4])
+            ->setPredicate(CC_S, cc[0]);
+         bld->mkOp1(OP_NOT, fTy, rr[1], t[3])
+            ->setPredicate(CC_S, cc[0]);
+
+         // add to low 32-bits, keep track of the carry
+         Instruction *n = bld->mkOp2(OP_ADD, fTy, NULL, rr[1], one);
+         n->setPredicate(CC_S, cc[0]);
+         n->setFlagsDef(0, (cc[1] = bld->getSSA(1, FILE_FLAGS)));
+
+         // If there was a carry, add 1 to the upper 32 bits
+         // XXX: These get executed even if they shouldn't be
+         bld->mkOp2(OP_ADD, fTy, rr[2], rr[0], one)
+            ->setPredicate(CC_C, cc[1]);
+         bld->mkMov(rr[3], rr[0])
+            ->setPredicate(CC_NC, cc[1]);
+         bld->mkOp2(OP_UNION, fTy, rr[4], rr[2], rr[3]);
+
+         // Merge the results from the negative and non-negative paths
+         bld->mkMov(rr[5], rr[4])
+            ->setPredicate(CC_S, cc[0]);
+         bld->mkMov(rr[6], r[4])
+            ->setPredicate(CC_NS, cc[0]);
+         bld->mkOp2(OP_UNION, mul->sType, mul->getDef(0), rr[5], rr[6]);
+      } else {
+         bld->mkMov(mul->getDef(0), r[4]);
+      }
    } else {
       bld->mkMov(mul->getDef(0), t[3]);
    }
@@ -590,6 +660,10 @@ void NV50LoweringPreSSA::loadTexMsInfo(uint32_t off, Value **ms,
    Value *tmp = new_LValue(func, FILE_GPR);
    uint8_t b = prog->driver->io.resInfoCBSlot;
    off += prog->driver->io.suInfoBase;
+   if (prog->getType() > Program::TYPE_VERTEX)
+      off += 16 * 2 * 4;
+   if (prog->getType() > Program::TYPE_GEOMETRY)
+      off += 16 * 2 * 4;
    *ms_x = bld.mkLoadv(TYPE_U32, bld.mkSymbol(
                              FILE_MEMORY_CONST, b, TYPE_U32, off + 0), NULL);
    *ms_y = bld.mkLoadv(TYPE_U32, bld.mkSymbol(

src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp

@@ -666,8 +666,9 @@ NVC0LoweringPass::handleTEX(TexInstruction *i)
    const int dim = i->tex.target.getDim() + i->tex.target.isCube();
    const int arg = i->tex.target.getArgCount();
    const int lyr = arg - (i->tex.target.isMS() ? 2 : 1);
+   const int chipset = prog->getTarget()->getChipset();
 
-   if (prog->getTarget()->getChipset() >= NVISA_GK104_CHIPSET) {
+   if (chipset >= NVISA_GK104_CHIPSET) {
       if (i->tex.rIndirectSrc >= 0 || i->tex.sIndirectSrc >= 0) {
          WARN("indirect TEX not implemented\n");
       }
@@ -697,7 +698,7 @@ NVC0LoweringPass::handleTEX(TexInstruction *i)
       }
    } else
    // (nvc0) generate and move the tsc/tic/array source to the front
-   if (dim != arg || i->tex.rIndirectSrc >= 0 || i->tex.sIndirectSrc >= 0) {
+   if (i->tex.target.isArray() || i->tex.rIndirectSrc >= 0 || i->tex.sIndirectSrc >= 0) {
       LValue *src = new_LValue(func, FILE_GPR); // 0xttxsaaaa
 
       Value *arrayIndex = i->tex.target.isArray() ? i->getSrc(lyr) : NULL;
@@ -728,6 +729,13 @@ NVC0LoweringPass::handleTEX(TexInstruction *i)
       i->setSrc(0, src);
    }
 
+   // For nvc0, the sample id has to be in the second operand, as the offset
+   // does. Right now we don't know how to pass both in, and this case can't
+   // happen with OpenGL. On nve0, the sample id is part of the texture
+   // coordinate argument.
+   assert(chipset >= NVISA_GK104_CHIPSET ||
+          !i->tex.useOffsets || !i->tex.target.isMS());
+
    // offset is last source (lod 1st, dc 2nd)
    if (i->tex.useOffsets) {
       uint32_t value = 0;
@@ -741,7 +749,7 @@ NVC0LoweringPass::handleTEX(TexInstruction *i)
       i->setSrc(s, bld.loadImm(NULL, value));
    }
 
-   if (prog->getTarget()->getChipset() >= NVISA_GK104_CHIPSET) {
+   if (chipset >= NVISA_GK104_CHIPSET) {
       //
       // If TEX requires more than 4 sources, the 2nd register tuple must be
       // aligned to 4, even if it consists of just a single 4-byte register.

src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp

@@ -187,7 +187,8 @@ LoadPropagation::checkSwapSrc01(Instruction *insn)
       return;
    }
 
-   if (insn->op == OP_SET)
+   if (insn->op == OP_SET || insn->op == OP_SET_AND ||
+       insn->op == OP_SET_OR || insn->op == OP_SET_XOR)
       insn->asCmp()->setCond = reverseCondCode(insn->asCmp()->setCond);
    else
    if (insn->op == OP_SLCT)
@@ -417,7 +418,17 @@ ConstantFolding::expr(Instruction *i,
       case TYPE_F32: res.data.f32 = a->data.f32 * b->data.f32; break;
       case TYPE_F64: res.data.f64 = a->data.f64 * b->data.f64; break;
       case TYPE_S32:
-      case TYPE_U32: res.data.u32 = a->data.u32 * b->data.u32; break;
+         if (i->subOp == NV50_IR_SUBOP_MUL_HIGH) {
+            res.data.s32 = ((int64_t)a->data.s32 * b->data.s32) >> 32;
+            break;
+         }
+         /* fallthrough */
+      case TYPE_U32:
+         if (i->subOp == NV50_IR_SUBOP_MUL_HIGH) {
+            res.data.u32 = ((uint64_t)a->data.u32 * b->data.u32) >> 32;
+            break;
+         }
+         res.data.u32 = a->data.u32 * b->data.u32; break;
       default:
          return;
       }
@@ -524,6 +535,7 @@ ConstantFolding::expr(Instruction *i,
    } else {
       i->op = OP_MOV;
    }
+   i->subOp = 0;
 }
 
 void
@@ -625,12 +637,41 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s)
 {
    const int t = !s;
    const operation op = i->op;
+   Instruction *newi = i;
 
    switch (i->op) {
    case OP_MUL:
       if (i->dType == TYPE_F32)
          tryCollapseChainedMULs(i, s, imm0);
 
+      if (i->subOp == NV50_IR_SUBOP_MUL_HIGH) {
+         assert(!isFloatType(i->sType));
+         if (imm0.isInteger(1) && i->dType == TYPE_S32) {
+            bld.setPosition(i, false);
+            // Need to set to the sign value, which is a compare.
+            newi = bld.mkCmp(OP_SET, CC_LT, TYPE_S32, i->getDef(0),
+                             TYPE_S32, i->getSrc(t), bld.mkImm(0));
+            delete_Instruction(prog, i);
+         } else if (imm0.isInteger(0) || imm0.isInteger(1)) {
+            // The high bits can't be set in this case (either mul by 0 or
+            // unsigned by 1)
+            i->op = OP_MOV;
+            i->subOp = 0;
+            i->setSrc(0, new_ImmediateValue(prog, 0u));
+            i->src(0).mod = Modifier(0);
+            i->setSrc(1, NULL);
+         } else if (!imm0.isNegative() && imm0.isPow2()) {
+            // Translate into a shift
+            imm0.applyLog2();
+            i->op = OP_SHR;
+            i->subOp = 0;
+            imm0.reg.data.u32 = 32 - imm0.reg.data.u32;
+            i->setSrc(0, i->getSrc(t));
+            i->src(0).mod = i->src(t).mod;
+            i->setSrc(1, new_ImmediateValue(prog, imm0.reg.data.u32));
+            i->src(1).mod = 0;
+         }
+      } else
       if (imm0.isInteger(0)) {
          i->op = OP_MOV;
          i->setSrc(0, new_ImmediateValue(prog, 0u));
@@ -721,7 +762,7 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s)
          else
             tA = tB;
          tB = s ? bld.getSSA() : i->getDef(0);
-         bld.mkOp2(OP_ADD, TYPE_U32, tB, mul->getDef(0), tA);
+         newi = bld.mkOp2(OP_ADD, TYPE_U32, tB, mul->getDef(0), tA);
          if (s)
             bld.mkOp2(OP_SHR, TYPE_U32, i->getDef(0), tB, bld.mkImm(s));
 
@@ -753,7 +794,7 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s)
          tA = bld.getSSA();
          bld.mkCmp(OP_SET, CC_LT, TYPE_S32, tA, TYPE_S32, i->getSrc(0), bld.mkImm(0));
          tD = (d < 0) ? bld.getSSA() : i->getDef(0)->asLValue();
-         bld.mkOp2(OP_SUB, TYPE_U32, tD, tB, tA);
+         newi = bld.mkOp2(OP_SUB, TYPE_U32, tD, tB, tA);
          if (d < 0)
             bld.mkOp1(OP_NEG, TYPE_S32, i->getDef(0), tB);
 
@@ -831,7 +872,7 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s)
    default:
       return;
    }
-   if (i->op != op)
+   if (newi->op != op)
       foldCount++;
 }
 

src/gallium/drivers/nouveau/codegen/nv50_ir_ra.cpp

@@ -1914,7 +1914,7 @@ RegAlloc::InsertConstraintsPass::texConstraintNVC0(TexInstruction *tex)
       s = tex->srcCount(0xff);
       n = 0;
    } else {
-      s = tex->tex.target.getArgCount();
+      s = tex->tex.target.getArgCount() - tex->tex.target.isMS();
       if (!tex->tex.target.isArray() &&
           (tex->tex.rIndirectSrc >= 0 || tex->tex.sIndirectSrc >= 0))
          ++s;

src/gallium/drivers/nouveau/codegen/nv50_ir_target_nv50.cpp

@@ -329,6 +329,8 @@ TargetNV50::insnCanLoad(const Instruction *i, int s,
          return false;
       if (sf == FILE_IMMEDIATE)
          return false;
+      if (i->subOp == NV50_IR_SUBOP_MUL_HIGH && sf == FILE_MEMORY_CONST)
+         return false;
       ldSize = 2;
    } else {
       ldSize = typeSizeof(ld->dType);

src/gallium/drivers/nouveau/nouveau_vp3_video.c

@@ -371,10 +371,10 @@ firmware_present(struct pipe_screen *pscreen, enum pipe_video_profile profile)
       int size, oclass;
       if (chipset < 0xc0)
          oclass = 0x85b1;
-      else if (vp5)
-         oclass = 0x95b1;
-      else
+      else if (chipset < 0xe0)
          oclass = 0x90b1;
+      else
+         oclass = 0x95b1;
 
       if (chipset < 0xc0) {
          data = &nv04_data;

src/gallium/drivers/nouveau/nv30/nv30_screen.c

@@ -302,10 +302,16 @@ nv30_screen_destroy(struct pipe_screen *pscreen)
 {
    struct nv30_screen *screen = nv30_screen(pscreen);
 
-   if (screen->base.fence.current &&
-       screen->base.fence.current->state >= NOUVEAU_FENCE_STATE_EMITTED) {
-      nouveau_fence_wait(screen->base.fence.current);
-      nouveau_fence_ref (NULL, &screen->base.fence.current);
+   if (screen->base.fence.current) {
+      struct nouveau_fence *current = NULL;
+
+      /* nouveau_fence_wait will create a new current fence, so wait on the
+       * _current_ one, and remove both.
+       */
+      nouveau_fence_ref(screen->base.fence.current, &current);
+      nouveau_fence_wait(current);
+      nouveau_fence_ref(NULL, &current);
+      nouveau_fence_ref(NULL, &screen->base.fence.current);
    }
 
    nouveau_object_del(&screen->query);

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

@@ -77,13 +77,13 @@
 /* 8 user clip planes, at 4 32-bit floats each */
 #define NV50_CB_AUX_UCP_OFFSET    0x0000
 #define NV50_CB_AUX_UCP_SIZE      (8 * 4 * 4)
-/* 256 textures, each with ms_x, ms_y u32 pairs */
+/* 16 textures * 3 shaders, each with ms_x, ms_y u32 pairs */
 #define NV50_CB_AUX_TEX_MS_OFFSET 0x0080
-#define NV50_CB_AUX_TEX_MS_SIZE   (256 * 2 * 4)
+#define NV50_CB_AUX_TEX_MS_SIZE   (16 * 3 * 2 * 4)
 /* For each MS level (4), 8 sets of 32-bit integer pairs sample offsets */
-#define NV50_CB_AUX_MS_OFFSET     0x880
+#define NV50_CB_AUX_MS_OFFSET     0x200
 #define NV50_CB_AUX_MS_SIZE       (4 * 8 * 4 * 2)
-/* next spot: 0x980 */
+/* next spot: 0x300 */
 /* 4 32-bit floats for the vertex runout, put at the end */
 #define NV50_CB_AUX_RUNOUT_OFFSET (NV50_CB_AUX_SIZE - 0x10)
 

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

@@ -290,8 +290,15 @@ nv50_screen_destroy(struct pipe_screen *pscreen)
    struct nv50_screen *screen = nv50_screen(pscreen);
 
    if (screen->base.fence.current) {
-      nouveau_fence_wait(screen->base.fence.current);
-      nouveau_fence_ref (NULL, &screen->base.fence.current);
+      struct nouveau_fence *current = NULL;
+
+      /* nouveau_fence_wait will create a new current fence, so wait on the
+       * _current_ one, and remove both.
+       */
+      nouveau_fence_ref(screen->base.fence.current, &current);
+      nouveau_fence_wait(current);
+      nouveau_fence_ref(NULL, &current);
+      nouveau_fence_ref(NULL, &screen->base.fence.current);
    }
    if (screen->base.pushbuf)
       screen->base.pushbuf->user_priv = NULL;

src/gallium/drivers/nouveau/nv50/nv50_state.c

@@ -556,11 +556,12 @@ nv50_sampler_state_delete(struct pipe_context *pipe, void *hwcso)
 {
    unsigned s, i;
 
-   for (s = 0; s < 3; ++s)
+   for (s = 0; s < 3; ++s) {
       assert(nv50_context(pipe)->num_samplers[s] <= PIPE_MAX_SAMPLERS);
       for (i = 0; i < nv50_context(pipe)->num_samplers[s]; ++i)
          if (nv50_context(pipe)->samplers[s][i] == hwcso)
             nv50_context(pipe)->samplers[s][i] = NULL;
+   }
 
    nv50_screen_tsc_free(nv50_context(pipe)->screen, nv50_tsc_entry(hwcso));
 

src/gallium/drivers/nouveau/nv50/nv50_surface.c

@@ -977,6 +977,7 @@ nv50_blit_3d(struct nv50_context *nv50, const struct pipe_blit_info *info)
    float x0, x1, y0, y1, z;
    float dz;
    float x_range, y_range;
+   float tri_x, tri_y;
 
    blit->mode = nv50_blit_select_mode(info);
    blit->color_mask = nv50_blit_derive_color_mask(info);
@@ -996,11 +997,14 @@ nv50_blit_3d(struct nv50_context *nv50, const struct pipe_blit_info *info)
    x_range = (float)info->src.box.width / (float)info->dst.box.width;
    y_range = (float)info->src.box.height / (float)info->dst.box.height;
 
+   tri_x = 16384 << nv50_miptree(dst)->ms_x;
+   tri_y = 16384 << nv50_miptree(dst)->ms_y;
+
    x0 = (float)info->src.box.x - x_range * (float)info->dst.box.x;
    y0 = (float)info->src.box.y - y_range * (float)info->dst.box.y;
 
-   x1 = x0 + 16384.0f * x_range;
-   y1 = y0 + 16384.0f * y_range;
+   x1 = x0 + tri_x * x_range;
+   y1 = y0 + tri_y * y_range;
 
    x0 *= (float)(1 << nv50_miptree(src)->ms_x);
    x1 *= (float)(1 << nv50_miptree(src)->ms_x);
@@ -1069,7 +1073,7 @@ nv50_blit_3d(struct nv50_context *nv50, const struct pipe_blit_info *info)
       PUSH_DATAf(push, y0);
       PUSH_DATAf(push, z);
       BEGIN_NV04(push, NV50_3D(VTX_ATTR_2F_X(0)), 2);
-      PUSH_DATAf(push, 16384 << nv50_miptree(dst)->ms_x);
+      PUSH_DATAf(push, tri_x);
       PUSH_DATAf(push, 0.0f);
       BEGIN_NV04(push, NV50_3D(VTX_ATTR_3F_X(1)), 3);
       PUSH_DATAf(push, x0);
@@ -1077,7 +1081,7 @@ nv50_blit_3d(struct nv50_context *nv50, const struct pipe_blit_info *info)
       PUSH_DATAf(push, z);
       BEGIN_NV04(push, NV50_3D(VTX_ATTR_2F_X(0)), 2);
       PUSH_DATAf(push, 0.0f);
-      PUSH_DATAf(push, 16384 << nv50_miptree(dst)->ms_y);
+      PUSH_DATAf(push, tri_y);
       BEGIN_NV04(push, NV50_3D(VERTEX_END_GL), 1);
       PUSH_DATA (push, 0);
    }

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

@@ -286,7 +286,7 @@ nv50_validate_tic(struct nv50_context *nv50, int s)
    }
    if (nv50->num_textures[s]) {
       BEGIN_NV04(push, NV50_3D(CB_ADDR), 1);
-      PUSH_DATA (push, (NV50_CB_AUX_TEX_MS_OFFSET << (8 - 2)) | NV50_CB_AUX);
+      PUSH_DATA (push, ((NV50_CB_AUX_TEX_MS_OFFSET + 16 * s * 2 * 4) << (8 - 2)) | NV50_CB_AUX);
       BEGIN_NI04(push, NV50_3D(CB_DATA(0)), nv50->num_textures[s] * 2);
       for (i = 0; i < nv50->num_textures[s]; i++) {
          struct nv50_tic_entry *tic = nv50_tic_entry(nv50->textures[s][i]);

src/gallium/drivers/nouveau/nvc0/nvc0_screen.c

@@ -334,7 +334,14 @@ nvc0_screen_destroy(struct pipe_screen *pscreen)
    struct nvc0_screen *screen = nvc0_screen(pscreen);
 
    if (screen->base.fence.current) {
-      nouveau_fence_wait(screen->base.fence.current);
+      struct nouveau_fence *current = NULL;
+
+      /* nouveau_fence_wait will create a new current fence, so wait on the
+       * _current_ one, and remove both.
+       */
+      nouveau_fence_ref(screen->base.fence.current, &current);
+      nouveau_fence_wait(current);
+      nouveau_fence_ref(NULL, &current);
       nouveau_fence_ref(NULL, &screen->base.fence.current);
    }
    if (screen->base.pushbuf)

src/gallium/drivers/r300/r300_blit.c

@@ -130,7 +130,7 @@ static boolean r300_cbzb_clear_allowed(struct r300_context *r300,
         (struct pipe_framebuffer_state*)r300->fb_state.state;
 
     /* Only color clear allowed, and only one colorbuffer. */
-    if ((clear_buffers & ~PIPE_CLEAR_COLOR) != 0 || fb->nr_cbufs != 1)
+    if ((clear_buffers & ~PIPE_CLEAR_COLOR) != 0 || fb->nr_cbufs != 1 || !fb->cbufs[0])
         return FALSE;
 
     return r300_surface(fb->cbufs[0])->cbzb_allowed;
@@ -313,7 +313,7 @@ static void r300_clear(struct pipe_context* pipe,
     /* Use fast color clear for an AA colorbuffer.
      * The CMASK is shared between all colorbuffers, so we use it
      * if there is only one colorbuffer bound. */
-    if ((buffers & PIPE_CLEAR_COLOR) && fb->nr_cbufs == 1 &&
+    if ((buffers & PIPE_CLEAR_COLOR) && fb->nr_cbufs == 1 && fb->cbufs[0] &&
         r300_resource(fb->cbufs[0]->texture)->tex.cmask_dwords) {
         /* Try to obtain the access to the CMASK if we don't have one. */
         if (!r300->cmask_access) {

src/gallium/drivers/r300/r300_context.h

@@ -688,6 +688,20 @@ static INLINE void r300_mark_atom_dirty(struct r300_context *r300,
     }
 }
 
+static INLINE struct pipe_surface *
+r300_get_nonnull_cb(struct pipe_framebuffer_state *fb, unsigned i)
+{
+    if (fb->cbufs[i])
+        return fb->cbufs[i];
+
+    /* The i-th framebuffer is NULL, return any non-NULL one. */
+    for (i = 0; i < fb->nr_cbufs; i++)
+        if (fb->cbufs[i])
+            return fb->cbufs[i];
+
+    return NULL;
+}
+
 struct pipe_context* r300_create_context(struct pipe_screen* screen,
                                          void *priv);
 

src/gallium/drivers/r300/r300_emit.c

@@ -42,15 +42,18 @@ void r300_emit_blend_state(struct r300_context* r300,
     struct r300_blend_state* blend = (struct r300_blend_state*)state;
     struct pipe_framebuffer_state* fb =
         (struct pipe_framebuffer_state*)r300->fb_state.state;
+    struct pipe_surface *cb;
     CS_LOCALS(r300);
 
-    if (fb->nr_cbufs) {
-        if (fb->cbufs[0]->format == PIPE_FORMAT_R16G16B16A16_FLOAT) {
+    cb = fb->nr_cbufs ? r300_get_nonnull_cb(fb, 0) : NULL;
+
+    if (cb) {
+        if (cb->format == PIPE_FORMAT_R16G16B16A16_FLOAT) {
             WRITE_CS_TABLE(blend->cb_noclamp, size);
-        } else if (fb->cbufs[0]->format == PIPE_FORMAT_R16G16B16X16_FLOAT) {
+        } else if (cb->format == PIPE_FORMAT_R16G16B16X16_FLOAT) {
             WRITE_CS_TABLE(blend->cb_noclamp_noalpha, size);
         } else {
-            unsigned swz = r300_surface(fb->cbufs[0])->colormask_swizzle;
+            unsigned swz = r300_surface(cb)->colormask_swizzle;
             WRITE_CS_TABLE(blend->cb_clamp[swz], size);
         }
     } else {
@@ -88,9 +91,11 @@ void r300_emit_dsa_state(struct r300_context* r300, unsigned size, void* state)
     /* Choose the alpha ref value between 8-bit (FG_ALPHA_FUNC.AM_VAL) and
      * 16-bit (FG_ALPHA_VALUE). */
     if (is_r500 && (alpha_func & R300_FG_ALPHA_FUNC_ENABLE)) {
-        if (fb->nr_cbufs &&
-            (fb->cbufs[0]->format == PIPE_FORMAT_R16G16B16A16_FLOAT ||
-             fb->cbufs[0]->format == PIPE_FORMAT_R16G16B16X16_FLOAT)) {
+        struct pipe_surface *cb = fb->nr_cbufs ? r300_get_nonnull_cb(fb, 0) : NULL;
+
+        if (cb &&
+            (cb->format == PIPE_FORMAT_R16G16B16A16_FLOAT ||
+             cb->format == PIPE_FORMAT_R16G16B16X16_FLOAT)) {
             alpha_func |= R500_FG_ALPHA_FUNC_FP16_ENABLE;
         } else {
             alpha_func |= R500_FG_ALPHA_FUNC_8BIT;
@@ -419,7 +424,7 @@ void r300_emit_fb_state(struct r300_context* r300, unsigned size, void* state)
 
     /* Set up colorbuffers. */
     for (i = 0; i < fb->nr_cbufs; i++) {
-        surf = r300_surface(fb->cbufs[i]);
+        surf = r300_surface(r300_get_nonnull_cb(fb, i));
 
         OUT_CS_REG(R300_RB3D_COLOROFFSET0 + (4 * i), surf->offset);
         OUT_CS_RELOC(surf);
@@ -600,7 +605,7 @@ void r300_emit_fb_state_pipelined(struct r300_context *r300,
      * (must be written after unpipelined regs) */
     OUT_CS_REG_SEQ(R300_US_OUT_FMT_0, 4);
     for (i = 0; i < num_cbufs; i++) {
-        OUT_CS(r300_surface(fb->cbufs[i])->format);
+        OUT_CS(r300_surface(r300_get_nonnull_cb(fb, i))->format);
     }
     for (; i < 1; i++) {
         OUT_CS(R300_US_OUT_FMT_C4_8 |
@@ -1310,6 +1315,8 @@ validate:
     if (r300->fb_state.dirty) {
         /* Color buffers... */
         for (i = 0; i < fb->nr_cbufs; i++) {
+            if (!fb->cbufs[i])
+                continue;
             tex = r300_resource(fb->cbufs[i]->texture);
             assert(tex && tex->buf && "cbuf is marked, but NULL!");
             r300->rws->cs_add_reloc(r300->cs, tex->cs_buf,

src/gallium/drivers/r300/r300_state.c

@@ -579,16 +579,17 @@ static void r300_set_blend_color(struct pipe_context* pipe,
     struct r300_blend_color_state *state =
         (struct r300_blend_color_state*)r300->blend_color_state.state;
     struct pipe_blend_color c;
-    enum pipe_format format = fb->nr_cbufs ? fb->cbufs[0]->format : 0;
+    struct pipe_surface *cb;
     float tmp;
     CB_LOCALS;
 
     state->state = *color; /* Save it, so that we can reuse it in set_fb_state */
     c = *color;
+    cb = fb->nr_cbufs ? r300_get_nonnull_cb(fb, 0) : NULL;
 
     /* The blend color is dependent on the colorbuffer format. */
-    if (fb->nr_cbufs) {
-        switch (format) {
+    if (cb) {
+        switch (cb->format) {
         case PIPE_FORMAT_R8_UNORM:
         case PIPE_FORMAT_L8_UNORM:
         case PIPE_FORMAT_I8_UNORM:
@@ -623,7 +624,7 @@ static void r300_set_blend_color(struct pipe_context* pipe,
         BEGIN_CB(state->cb, 3);
         OUT_CB_REG_SEQ(R500_RB3D_CONSTANT_COLOR_AR, 2);
 
-        switch (format) {
+        switch (cb ? cb->format : 0) {
         case PIPE_FORMAT_R16G16B16A16_FLOAT:
         case PIPE_FORMAT_R16G16B16X16_FLOAT:
             OUT_CB(util_float_to_half(c.color[2]) |
@@ -858,6 +859,9 @@ static void r300_fb_set_tiling_flags(struct r300_context *r300,
 
     /* Set tiling flags for new surfaces. */
     for (i = 0; i < state->nr_cbufs; i++) {
+        if (!state->cbufs[i])
+            continue;
+
         r300_tex_set_tiling_flags(r300,
                                   r300_resource(state->cbufs[i]->texture),
                                   state->cbufs[i]->u.tex.level);
@@ -950,7 +954,8 @@ static unsigned r300_get_num_samples(struct r300_context *r300)
     num_samples = 6;
 
     for (i = 0; i < fb->nr_cbufs; i++)
-        num_samples = MIN2(num_samples, fb->cbufs[i]->texture->nr_samples);
+        if (fb->cbufs[i])
+            num_samples = MIN2(num_samples, fb->cbufs[i]->texture->nr_samples);
 
     if (fb->zsbuf)
         num_samples = MIN2(num_samples, fb->zsbuf->texture->nr_samples);
@@ -967,7 +972,7 @@ r300_set_framebuffer_state(struct pipe_context* pipe,
 {
     struct r300_context* r300 = r300_context(pipe);
     struct r300_aa_state *aa = (struct r300_aa_state*)r300->aa_state.state;
-    struct pipe_framebuffer_state *old_state = r300->fb_state.state;
+    struct pipe_framebuffer_state *current_state = r300->fb_state.state;
     unsigned max_width, max_height, i;
     uint32_t zbuffer_bpp = 0;
     boolean unlock_zbuffer = FALSE;
@@ -986,17 +991,17 @@ r300_set_framebuffer_state(struct pipe_context* pipe,
         return;
     }
 
-    if (old_state->zsbuf && r300->zmask_in_use && !r300->locked_zbuffer) {
+    if (current_state->zsbuf && r300->zmask_in_use && !r300->locked_zbuffer) {
         /* There is a zmask in use, what are we gonna do? */
         if (state->zsbuf) {
-            if (!pipe_surface_equal(old_state->zsbuf, state->zsbuf)) {
+            if (!pipe_surface_equal(current_state->zsbuf, state->zsbuf)) {
                 /* Decompress the currently bound zbuffer before we bind another one. */
                 r300_decompress_zmask(r300);
                 r300->hiz_in_use = FALSE;
             }
         } else {
             /* We don't bind another zbuffer, so lock the current one. */
-            pipe_surface_reference(&r300->locked_zbuffer, old_state->zsbuf);
+            pipe_surface_reference(&r300->locked_zbuffer, current_state->zsbuf);
         }
     } else if (r300->locked_zbuffer) {
         /* We have a locked zbuffer now, what are we gonna do? */
@@ -1014,9 +1019,20 @@ r300_set_framebuffer_state(struct pipe_context* pipe,
     }
     assert(state->zsbuf || (r300->locked_zbuffer && !unlock_zbuffer) || !r300->zmask_in_use);
 
+    /* If zsbuf is set from NULL to non-NULL or vice versa.. */
+    if (!!current_state->zsbuf != !!state->zsbuf) {
+        r300_mark_atom_dirty(r300, &r300->dsa_state);
+    }
+
+    util_copy_framebuffer_state(r300->fb_state.state, state);
+
+    /* Remove trailing NULL colorbuffers. */
+    while (current_state->nr_cbufs && !current_state->cbufs[current_state->nr_cbufs-1])
+        current_state->nr_cbufs--;
+
     /* Set whether CMASK can be used. */
     r300->cmask_in_use =
-        state->nr_cbufs == 1 &&
+        state->nr_cbufs == 1 && state->cbufs[0] &&
         r300->screen->cmask_resource == state->cbufs[0]->texture;
 
     /* Need to reset clamping or colormask. */
@@ -1025,11 +1041,6 @@ r300_set_framebuffer_state(struct pipe_context* pipe,
     /* Re-swizzle the blend color. */
     r300_set_blend_color(pipe, &((struct r300_blend_color_state*)r300->blend_color_state.state)->state);
 
-    /* If zsbuf is set from NULL to non-NULL or vice versa.. */
-    if (!!old_state->zsbuf != !!state->zsbuf) {
-        r300_mark_atom_dirty(r300, &r300->dsa_state);
-    }
-
     if (r300->screen->info.drm_minor < 12) {
        /* The tiling flags are dependent on the surface miplevel, unfortunately.
         * This workarounds a bad design decision in old kernels which were
@@ -1037,8 +1048,6 @@ r300_set_framebuffer_state(struct pipe_context* pipe,
         r300_fb_set_tiling_flags(r300, state);
     }
 
-    util_copy_framebuffer_state(r300->fb_state.state, state);
-
     if (unlock_zbuffer) {
         pipe_surface_reference(&r300->locked_zbuffer, NULL);
     }
@@ -1089,7 +1098,8 @@ r300_set_framebuffer_state(struct pipe_context* pipe,
     if (DBG_ON(r300, DBG_FB)) {
         fprintf(stderr, "r300: set_framebuffer_state:\n");
         for (i = 0; i < state->nr_cbufs; i++) {
-            r300_print_fb_surf_info(state->cbufs[i], i, "CB");
+            if (state->cbufs[i])
+                r300_print_fb_surf_info(state->cbufs[i], i, "CB");
         }
         if (state->zsbuf) {
             r300_print_fb_surf_info(state->zsbuf, 0, "ZB");

src/gallium/drivers/r600/evergreen_state.c

@@ -927,7 +927,8 @@ static void *evergreen_create_rs_state(struct pipe_context *ctx,
 		S_028810_PS_UCP_MODE(3) |
 		S_028810_ZCLIP_NEAR_DISABLE(!state->depth_clip) |
 		S_028810_ZCLIP_FAR_DISABLE(!state->depth_clip) |
-		S_028810_DX_LINEAR_ATTR_CLIP_ENA(1);
+		S_028810_DX_LINEAR_ATTR_CLIP_ENA(1) |
+		S_028810_DX_RASTERIZATION_KILL(state->rasterizer_discard);
 	rs->multisample_enable = state->multisample;
 
 	/* offset */
@@ -996,7 +997,6 @@ static void *evergreen_create_rs_state(struct pipe_context *ctx,
 						  state->fill_back != PIPE_POLYGON_MODE_FILL) |
 			       S_028814_POLYMODE_FRONT_PTYPE(r600_translate_fill(state->fill_front)) |
 			       S_028814_POLYMODE_BACK_PTYPE(r600_translate_fill(state->fill_back)));
-	r600_store_context_reg(&rs->buffer, R_028350_SX_MISC, S_028350_MULTIPASS(state->rasterizer_discard));
 	return rs;
 }
 
@@ -1097,7 +1097,8 @@ struct pipe_sampler_view *
 evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 				     struct pipe_resource *texture,
 				     const struct pipe_sampler_view *state,
-				     unsigned width0, unsigned height0)
+				     unsigned width0, unsigned height0,
+				     unsigned force_level)
 {
 	struct r600_screen *rscreen = (struct r600_screen*)ctx->screen;
 	struct r600_pipe_sampler_view *view = CALLOC_STRUCT(r600_pipe_sampler_view);
@@ -1109,6 +1110,8 @@ evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 	unsigned macro_aspect, tile_split, bankh, bankw, nbanks, fmask_bankh;
 	enum pipe_format pipe_format = state->format;
 	struct radeon_surface_level *surflevel;
+	unsigned base_level, first_level, last_level;
+	uint64_t va;
 
 	if (view == NULL)
 		return NULL;
@@ -1165,13 +1168,26 @@ evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 
 	endian = r600_colorformat_endian_swap(format);
 
+	base_level = 0;
+	first_level = state->u.tex.first_level;
+	last_level = state->u.tex.last_level;
 	width = width0;
 	height = height0;
 	depth = texture->depth0;
-	pitch = surflevel[0].nblk_x * util_format_get_blockwidth(pipe_format);
+
+	if (force_level) {
+		base_level = force_level;
+		first_level = 0;
+		last_level = 0;
+		width = u_minify(width, force_level);
+		height = u_minify(height, force_level);
+		depth = u_minify(depth, force_level);
+	}
+
+	pitch = surflevel[base_level].nblk_x * util_format_get_blockwidth(pipe_format);
 	non_disp_tiling = tmp->non_disp_tiling;
 
-	switch (surflevel[0].mode) {
+	switch (surflevel[base_level].mode) {
 	case RADEON_SURF_MODE_LINEAR_ALIGNED:
 		array_mode = V_028C70_ARRAY_LINEAR_ALIGNED;
 		break;
@@ -1210,6 +1226,8 @@ evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 	} else if (texture->target == PIPE_TEXTURE_CUBE_ARRAY)
 		depth = texture->array_size / 6;
 
+	va = r600_resource_va(ctx->screen, texture);
+
 	view->tex_resource = &tmp->resource;
 	view->tex_resource_words[0] = (S_030000_DIM(r600_tex_dim(texture->target, texture->nr_samples)) |
 				       S_030000_PITCH((pitch / 8) - 1) |
@@ -1221,7 +1239,7 @@ evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 	view->tex_resource_words[1] = (S_030004_TEX_HEIGHT(height - 1) |
 				       S_030004_TEX_DEPTH(depth - 1) |
 				       S_030004_ARRAY_MODE(array_mode));
-	view->tex_resource_words[2] = (surflevel[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+	view->tex_resource_words[2] = (surflevel[base_level].offset + va) >> 8;
 
 	/* TEX_RESOURCE_WORD3.MIP_ADDRESS */
 	if (texture->nr_samples > 1 && rscreen->has_compressed_msaa_texturing) {
@@ -1231,12 +1249,12 @@ evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 			view->skip_mip_address_reloc = true;
 		} else {
 			/* FMASK should be in MIP_ADDRESS for multisample textures */
-			view->tex_resource_words[3] = (tmp->fmask.offset + r600_resource_va(ctx->screen, texture)) >> 8;
+			view->tex_resource_words[3] = (tmp->fmask.offset + va) >> 8;
 		}
-	} else if (state->u.tex.last_level && texture->nr_samples <= 1) {
-		view->tex_resource_words[3] = (surflevel[1].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+	} else if (last_level && texture->nr_samples <= 1) {
+		view->tex_resource_words[3] = (surflevel[1].offset + va) >> 8;
 	} else {
-		view->tex_resource_words[3] = (surflevel[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+		view->tex_resource_words[3] = (surflevel[base_level].offset + va) >> 8;
 	}
 
 	view->tex_resource_words[4] = (word4 |
@@ -1255,8 +1273,8 @@ evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 		view->tex_resource_words[5] |= S_030014_LAST_LEVEL(log_samples);
 		view->tex_resource_words[6] |= S_030018_FMASK_BANK_HEIGHT(fmask_bankh);
 	} else {
-		view->tex_resource_words[4] |= S_030010_BASE_LEVEL(state->u.tex.first_level);
-		view->tex_resource_words[5] |= S_030014_LAST_LEVEL(state->u.tex.last_level);
+		view->tex_resource_words[4] |= S_030010_BASE_LEVEL(first_level);
+		view->tex_resource_words[5] |= S_030014_LAST_LEVEL(last_level);
 		/* aniso max 16 samples */
 		view->tex_resource_words[6] |= S_030018_MAX_ANISO(4);
 	}
@@ -1277,7 +1295,7 @@ evergreen_create_sampler_view(struct pipe_context *ctx,
 			      const struct pipe_sampler_view *state)
 {
 	return evergreen_create_sampler_view_custom(ctx, tex, state,
-						    tex->width0, tex->height0);
+						    tex->width0, tex->height0, 0);
 }
 
 static void evergreen_emit_clip_state(struct r600_context *rctx, struct r600_atom *atom)
@@ -1824,7 +1842,10 @@ static void evergreen_set_framebuffer_state(struct pipe_context *ctx,
 	}
 
 	log_samples = util_logbase2(rctx->framebuffer.nr_samples);
-	if (rctx->b.chip_class == CAYMAN && rctx->db_misc_state.log_samples != log_samples) {
+	/* This is for Cayman to program SAMPLE_RATE, and for RV770 to fix a hw bug. */
+	if ((rctx->b.chip_class == CAYMAN ||
+	     rctx->b.family == CHIP_RV770) &&
+	    rctx->db_misc_state.log_samples != log_samples) {
 		rctx->db_misc_state.log_samples = log_samples;
 		rctx->db_misc_state.atom.dirty = true;
 	}
@@ -2809,7 +2830,9 @@ void cayman_init_common_regs(struct r600_command_buffer *cb,
 
 	r600_store_context_reg(cb, R_028A4C_PA_SC_MODE_CNTL_1, 0);
 
-	r600_store_context_reg(cb, R_028354_SX_SURFACE_SYNC, S_028354_SURFACE_SYNC_MASK(0xf));
+	r600_store_context_reg_seq(cb, R_028350_SX_MISC, 2);
+	r600_store_value(cb, 0);
+	r600_store_value(cb, S_028354_SURFACE_SYNC_MASK(0xf));
 
 	r600_store_context_reg(cb, R_028800_DB_DEPTH_CONTROL, 0);
 }
@@ -3085,7 +3108,9 @@ void evergreen_init_common_regs(struct r600_command_buffer *cb,
 	/* The cs checker requires this register to be set. */
 	r600_store_context_reg(cb, R_028800_DB_DEPTH_CONTROL, 0);
 
-	r600_store_context_reg(cb, R_028354_SX_SURFACE_SYNC, S_028354_SURFACE_SYNC_MASK(0xf));
+	r600_store_context_reg_seq(cb, R_028350_SX_MISC, 2);
+	r600_store_value(cb, 0);
+	r600_store_value(cb, S_028354_SURFACE_SYNC_MASK(0xf));
 
 	return;
 }
@@ -3603,30 +3628,6 @@ void evergreen_update_es_state(struct pipe_context *ctx, struct r600_pipe_shader
 	/* After that, the NOP relocation packet must be emitted (shader->bo, RADEON_USAGE_READ). */
 }
 
-static unsigned r600_conv_prim_to_gs_out(unsigned mode)
-{
-	static const int prim_conv[] = {
-		V_028A6C_OUTPRIM_TYPE_POINTLIST,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP
-	};
-	assert(mode < Elements(prim_conv));
-
-	return prim_conv[mode];
-}
-
 void evergreen_update_gs_state(struct pipe_context *ctx, struct r600_pipe_shader *shader)
 {
 	struct r600_context *rctx = (struct r600_context *)ctx;
@@ -3727,6 +3728,7 @@ void evergreen_update_vs_state(struct pipe_context *ctx, struct r600_pipe_shader
 		S_02881C_VS_OUT_CCDIST1_VEC_ENA((rshader->clip_dist_write & 0xF0) != 0) |
 		S_02881C_VS_OUT_MISC_VEC_ENA(rshader->vs_out_misc_write) |
 		S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size) |
+		S_02881C_USE_VTX_EDGE_FLAG(rshader->vs_out_edgeflag) |
 		S_02881C_USE_VTX_RENDER_TARGET_INDX(rshader->vs_out_layer);
 }
 

src/gallium/drivers/r600/evergreend.h

@@ -582,6 +582,9 @@
 #define   S_028810_VTX_KILL_OR(x)                      (((x) & 0x1) << 21)
 #define   G_028810_VTX_KILL_OR(x)                      (((x) >> 21) & 0x1)
 #define   C_028810_VTX_KILL_OR                         0xFFDFFFFF
+#define   S_028810_DX_RASTERIZATION_KILL(x)            (((x) & 0x1) << 22)
+#define   G_028810_DX_RASTERIZATION_KILL(x)            (((x) >> 22) & 0x1)
+#define   C_028810_DX_RASTERIZATION_KILL               0xFFBFFFFF
 #define   S_028810_DX_LINEAR_ATTR_CLIP_ENA(x)          (((x) & 0x1) << 24)
 #define   G_028810_DX_LINEAR_ATTR_CLIP_ENA(x)          (((x) >> 24) & 0x1)
 #define   C_028810_DX_LINEAR_ATTR_CLIP_ENA             0xFEFFFFFF

src/gallium/drivers/r600/r600_blit.c

@@ -599,6 +599,12 @@ static void r600_copy_buffer(struct pipe_context *ctx, struct pipe_resource *dst
 	} else {
 		util_resource_copy_region(ctx, dst, 0, dstx, 0, 0, src, 0, src_box);
 	}
+
+	/* The index buffer (VGT) doesn't seem to see the result of the copying.
+	 * Can we somehow flush the index buffer cache? Starting a new IB seems
+	 * to do the trick. */
+	if (rctx->b.chip_class <= R700)
+		rctx->b.rings.gfx.flush(ctx, RADEON_FLUSH_ASYNC);
 }
 
 /**
@@ -679,6 +685,7 @@ static void r600_resource_copy_region(struct pipe_context *ctx,
 	struct pipe_surface *dst_view, dst_templ;
 	struct pipe_sampler_view src_templ, *src_view;
 	unsigned dst_width, dst_height, src_width0, src_height0, src_widthFL, src_heightFL;
+	unsigned src_force_level = 0;
 	struct pipe_box sbox, dstbox;
 
 	/* Handle buffers first. */
@@ -737,6 +744,8 @@ static void r600_resource_copy_region(struct pipe_context *ctx,
 		sbox.height = util_format_get_nblocksy(src->format, src_box->height);
 		sbox.depth = src_box->depth;
 		src_box = &sbox;
+
+		src_force_level = src_level;
 	} else if (!util_blitter_is_copy_supported(rctx->blitter, dst, src)) {
 		if (util_format_is_subsampled_2x1_32bpp(src->format)) {
 
@@ -789,7 +798,8 @@ static void r600_resource_copy_region(struct pipe_context *ctx,
 
 	if (rctx->b.chip_class >= EVERGREEN) {
 		src_view = evergreen_create_sampler_view_custom(ctx, src, &src_templ,
-								src_width0, src_height0);
+								src_width0, src_height0,
+								src_force_level);
 	} else {
 		src_view = r600_create_sampler_view_custom(ctx, src, &src_templ,
 							   src_widthFL, src_heightFL);

src/gallium/drivers/r600/r600_hw_context.c

@@ -81,7 +81,7 @@ void r600_need_cs_space(struct r600_context *ctx, unsigned num_dw,
 	}
 
 	/* SX_MISC */
-	if (ctx->b.chip_class <= R700) {
+	if (ctx->b.chip_class == R600) {
 		num_dw += 3;
 	}
 
@@ -210,6 +210,15 @@ void r600_flush_emit(struct r600_context *rctx)
 				S_0085F0_SMX_ACTION_ENA(1);
 	}
 
+	/* Workaround for buggy flushing on some R6xx chipsets. */
+	if (rctx->b.flags & R600_CONTEXT_FLUSH_AND_INV &&
+	    (rctx->b.family == CHIP_RV670 ||
+	     rctx->b.family == CHIP_RS780 ||
+	     rctx->b.family == CHIP_RS880)) {
+		cp_coher_cntl |=  S_0085F0_CB1_DEST_BASE_ENA(1) |
+				  S_0085F0_DEST_BASE_0_ENA(1);
+	}
+
 	if (cp_coher_cntl) {
 		cs->buf[cs->cdw++] = PKT3(PKT3_SURFACE_SYNC, 3, 0);
 		cs->buf[cs->cdw++] = cp_coher_cntl;   /* CP_COHER_CNTL */
@@ -260,7 +269,7 @@ void r600_context_flush(struct r600_context *ctx, unsigned flags)
 	r600_flush_emit(ctx);
 
 	/* old kernels and userspace don't set SX_MISC, so we must reset it to 0 here */
-	if (ctx->b.chip_class <= R700) {
+	if (ctx->b.chip_class == R600) {
 		r600_write_context_reg(cs, R_028350_SX_MISC, 0);
 	}
 

src/gallium/drivers/r600/r600_pipe.c

@@ -44,7 +44,7 @@
 static const struct debug_named_value r600_debug_options[] = {
 	/* features */
 #if defined(R600_USE_LLVM)
-	{ "nollvm", DBG_NO_LLVM, "Disable the LLVM shader compiler" },
+	{ "llvm", DBG_LLVM, "Enable the LLVM shader compiler" },
 #endif
 	{ "nocpdma", DBG_NO_CP_DMA, "Disable CP DMA" },
 	{ "nodma", DBG_NO_ASYNC_DMA, "Disable asynchronous DMA" },
@@ -419,15 +419,17 @@ static int r600_get_param(struct pipe_screen* pscreen, enum pipe_cap param)
 
 	/* Texturing. */
 	case PIPE_CAP_MAX_TEXTURE_2D_LEVELS:
-	case PIPE_CAP_MAX_TEXTURE_3D_LEVELS:
 	case PIPE_CAP_MAX_TEXTURE_CUBE_LEVELS:
 		if (family >= CHIP_CEDAR)
 			return 15;
 		else
 			return 14;
+	case PIPE_CAP_MAX_TEXTURE_3D_LEVELS:
+		/* textures support 8192, but layered rendering supports 2048 */
+		return 12;
 	case PIPE_CAP_MAX_TEXTURE_ARRAY_LAYERS:
-		return rscreen->b.info.drm_minor >= 9 ?
-			(family >= CHIP_CEDAR ? 16384 : 8192) : 0;
+		/* textures support 8192, but layered rendering supports 2048 */
+		return rscreen->b.info.drm_minor >= 9 ? 2048 : 0;
 	case PIPE_CAP_MAX_COMBINED_SAMPLERS:
 		return 48;
 
@@ -589,8 +591,8 @@ struct pipe_screen *r600_screen_create(struct radeon_winsys *ws)
 		rscreen->b.debug_flags |= DBG_FS | DBG_VS | DBG_GS | DBG_PS | DBG_CS;
 	if (debug_get_bool_option("R600_HYPERZ", FALSE))
 		rscreen->b.debug_flags |= DBG_HYPERZ;
-	if (!debug_get_bool_option("R600_LLVM", TRUE))
-		rscreen->b.debug_flags |= DBG_NO_LLVM;
+	if (debug_get_bool_option("R600_LLVM", FALSE))
+		rscreen->b.debug_flags |= DBG_LLVM;
 
 	if (rscreen->b.family == CHIP_UNKNOWN) {
 		fprintf(stderr, "r600: Unknown chipset 0x%04X\n", rscreen->b.info.pci_id);

src/gallium/drivers/r600/r600_pipe.h

@@ -42,7 +42,7 @@
 
 /* the number of CS dwords for flushing and drawing */
 #define R600_MAX_FLUSH_CS_DWORDS	16
-#define R600_MAX_DRAW_CS_DWORDS		34
+#define R600_MAX_DRAW_CS_DWORDS		37
 #define R600_TRACE_CS_DWORDS		7
 
 #define R600_MAX_USER_CONST_BUFFERS 13
@@ -195,7 +195,7 @@ struct r600_gs_rings_state {
 
 /* This must start from 16. */
 /* features */
-#define DBG_NO_LLVM		(1 << 17)
+#define DBG_LLVM		(1 << 17)
 #define DBG_NO_CP_DMA		(1 << 18)
 #define DBG_NO_ASYNC_DMA	(1 << 19)
 /* shader backend */
@@ -235,6 +235,7 @@ struct r600_rasterizer_state {
 	unsigned                        clip_plane_enable;
 	unsigned			pa_sc_line_stipple;
 	unsigned			pa_cl_clip_cntl;
+	unsigned			pa_su_sc_mode_cntl;
 	float				offset_units;
 	float				offset_scale;
 	bool				offset_enable;
@@ -511,7 +512,8 @@ struct pipe_sampler_view *
 evergreen_create_sampler_view_custom(struct pipe_context *ctx,
 				     struct pipe_resource *texture,
 				     const struct pipe_sampler_view *state,
-				     unsigned width0, unsigned height0);
+				     unsigned width0, unsigned height0,
+				     unsigned force_level);
 void evergreen_init_common_regs(struct r600_command_buffer *cb,
 				enum chip_class ctx_chip_class,
 				enum radeon_family ctx_family,
@@ -825,15 +827,6 @@ static INLINE uint32_t S_FIXED(float value, uint32_t frac_bits)
 }
 #define ALIGN_DIVUP(x, y) (((x) + (y) - 1) / (y))
 
-static inline unsigned r600_tex_aniso_filter(unsigned filter)
-{
-	if (filter <= 1)   return 0;
-	if (filter <= 2)   return 1;
-	if (filter <= 4)   return 2;
-	if (filter <= 8)   return 3;
-	 /* else */        return 4;
-}
-
 /* 12.4 fixed-point */
 static INLINE unsigned r600_pack_float_12p4(float x)
 {
@@ -841,4 +834,32 @@ static INLINE unsigned r600_pack_float_12p4(float x)
 	       x >= 4096 ? 0xffff : x * 16;
 }
 
+#define     V_028A6C_OUTPRIM_TYPE_POINTLIST            0
+#define     V_028A6C_OUTPRIM_TYPE_LINESTRIP            1
+#define     V_028A6C_OUTPRIM_TYPE_TRISTRIP             2
+
+static INLINE unsigned r600_conv_prim_to_gs_out(unsigned mode)
+{
+	static const int prim_conv[] = {
+		V_028A6C_OUTPRIM_TYPE_POINTLIST,
+		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
+		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
+		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
+		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
+		V_028A6C_OUTPRIM_TYPE_TRISTRIP
+	};
+	assert(mode < Elements(prim_conv));
+
+	return prim_conv[mode];
+}
+
 #endif

src/gallium/drivers/r600/r600_shader.c

@@ -155,7 +155,7 @@ int r600_pipe_shader_create(struct pipe_context *ctx,
 	r = r600_shader_from_tgsi(rctx, shader, key);
 	if (r) {
 		R600_ERR("translation from TGSI failed !\n");
-		return r;
+		goto error;
 	}
 
 	/* disable SB for geom shaders - it can't handle the CF_EMIT instructions */
@@ -169,7 +169,7 @@ int r600_pipe_shader_create(struct pipe_context *ctx,
 		r = r600_bytecode_build(&shader->shader.bc);
 		if (r) {
 			R600_ERR("building bytecode failed !\n");
-			return r;
+			goto error;
 		}
 	}
 
@@ -182,7 +182,7 @@ int r600_pipe_shader_create(struct pipe_context *ctx,
 		                             dump, use_sb);
 		if (r) {
 			R600_ERR("r600_sb_bytecode_process failed !\n");
-			return r;
+			goto error;
 		}
 	}
 
@@ -192,16 +192,16 @@ int r600_pipe_shader_create(struct pipe_context *ctx,
 			r = r600_sb_bytecode_process(rctx, &shader->gs_copy_shader->shader.bc,
 						     &shader->gs_copy_shader->shader, dump, 0);
 			if (r)
-				return r;
+				goto error;
 		}
 
 		if ((r = store_shader(ctx, shader->gs_copy_shader)))
-			return r;
+			goto error;
 	}
 
 	/* Store the shader in a buffer. */
 	if ((r = store_shader(ctx, shader)))
-		return r;
+		goto error;
 
 	/* Build state. */
 	switch (shader->shader.processor_type) {
@@ -235,9 +235,14 @@ int r600_pipe_shader_create(struct pipe_context *ctx,
 		}
 		break;
 	default:
-		return -EINVAL;
+		r = -EINVAL;
+		goto error;
 	}
 	return 0;
+
+error:
+	r600_pipe_shader_destroy(ctx, shader);
+	return r;
 }
 
 void r600_pipe_shader_destroy(struct pipe_context *ctx, struct r600_pipe_shader *shader)
@@ -286,6 +291,7 @@ struct r600_shader_ctx {
 	int					colors_used;
 	boolean                 clip_vertex_write;
 	unsigned                cv_output;
+	unsigned		edgeflag_output;
 	int					fragcoord_input;
 	int					native_integers;
 	int					next_ring_offset;
@@ -490,9 +496,10 @@ static int r600_spi_sid(struct r600_shader_io * io)
 	 * semantic indices, so we'll use 0 for them.
 	 */
 	if (name == TGSI_SEMANTIC_POSITION ||
-		name == TGSI_SEMANTIC_PSIZE ||
-		name == TGSI_SEMANTIC_LAYER ||
-		name == TGSI_SEMANTIC_FACE)
+	    name == TGSI_SEMANTIC_PSIZE ||
+	    name == TGSI_SEMANTIC_EDGEFLAG ||
+	    name == TGSI_SEMANTIC_LAYER ||
+	    name == TGSI_SEMANTIC_FACE)
 		index = 0;
 	else {
 		if (name == TGSI_SEMANTIC_GENERIC) {
@@ -623,6 +630,11 @@ static int tgsi_declaration(struct r600_shader_ctx *ctx)
 				ctx->shader->vs_out_misc_write = 1;
 				ctx->shader->vs_out_point_size = 1;
 				break;
+			case TGSI_SEMANTIC_EDGEFLAG:
+				ctx->shader->vs_out_misc_write = 1;
+				ctx->shader->vs_out_edgeflag = 1;
+				ctx->edgeflag_output = i;
+				break;
 			case TGSI_SEMANTIC_LAYER:
 				ctx->shader->vs_out_misc_write = 1;
 				ctx->shader->vs_out_layer = 1;
@@ -1157,6 +1169,35 @@ out_err:
 	return r;
 }
 
+static void convert_edgeflag_to_int(struct r600_shader_ctx *ctx)
+{
+	struct r600_bytecode_alu alu;
+	unsigned reg;
+
+	if (!ctx->shader->vs_out_edgeflag)
+		return;
+
+	reg = ctx->shader->output[ctx->edgeflag_output].gpr;
+
+	/* clamp(x, 0, 1) */
+	memset(&alu, 0, sizeof(alu));
+	alu.op = ALU_OP1_MOV;
+	alu.src[0].sel = reg;
+	alu.dst.sel = reg;
+	alu.dst.write = 1;
+	alu.dst.clamp = 1;
+	alu.last = 1;
+	r600_bytecode_add_alu(ctx->bc, &alu);
+
+	memset(&alu, 0, sizeof(alu));
+	alu.op = ALU_OP1_FLT_TO_INT;
+	alu.src[0].sel = reg;
+	alu.dst.sel = reg;
+	alu.dst.write = 1;
+	alu.last = 1;
+	r600_bytecode_add_alu(ctx->bc, &alu);
+}
+
 static int generate_gs_copy_shader(struct r600_context *rctx,
 				   struct r600_pipe_shader *gs,
 				   struct pipe_stream_output_info *so)
@@ -1480,7 +1521,7 @@ static int r600_shader_from_tgsi(struct r600_context *rctx,
 	bool pos_emitted = false;
 
 #ifdef R600_USE_LLVM
-	use_llvm = !(rscreen->b.debug_flags & DBG_NO_LLVM);
+	use_llvm = rscreen->b.debug_flags & DBG_LLVM;
 #endif
 	ctx.bc = &shader->bc;
 	ctx.shader = shader;
@@ -1901,6 +1942,8 @@ static int r600_shader_from_tgsi(struct r600_context *rctx,
 	    so.num_outputs && !use_llvm)
 		emit_streamout(&ctx, &so);
 
+	convert_edgeflag_to_int(&ctx);
+
 	if (ring_outputs) {
 		if (key.vs_as_es)
 			emit_gs_ring_writes(&ctx, FALSE);
@@ -1936,6 +1979,15 @@ static int r600_shader_from_tgsi(struct r600_context *rctx,
 					output[j].type = V_SQ_CF_ALLOC_EXPORT_WORD0_SQ_EXPORT_POS;
 					pos_emitted = true;
 					break;
+				case TGSI_SEMANTIC_EDGEFLAG:
+					output[j].array_base = 61;
+					output[j].swizzle_x = 7;
+					output[j].swizzle_y = 0;
+					output[j].swizzle_z = 7;
+					output[j].swizzle_w = 7;
+					output[j].type = V_SQ_CF_ALLOC_EXPORT_WORD0_SQ_EXPORT_POS;
+					pos_emitted = true;
+					break;
 				case TGSI_SEMANTIC_LAYER:
 					output[j].array_base = 61;
 					output[j].swizzle_x = 7;
@@ -4338,11 +4390,16 @@ static int tgsi_tex(struct r600_shader_ctx *ctx)
 				    (inst->Texture.Texture == TGSI_TEXTURE_2D_MSAA ||
 				     inst->Texture.Texture == TGSI_TEXTURE_2D_ARRAY_MSAA);
 
+	bool txf_add_offsets = inst->Texture.NumOffsets &&
+			     inst->Instruction.Opcode == TGSI_OPCODE_TXF &&
+			     inst->Texture.Texture != TGSI_TEXTURE_BUFFER;
+
 	/* Texture fetch instructions can only use gprs as source.
 	 * Also they cannot negate the source or take the absolute value */
 	const boolean src_requires_loading = (inst->Instruction.Opcode != TGSI_OPCODE_TXQ_LZ &&
                                               tgsi_tex_src_requires_loading(ctx, 0)) ||
-					     read_compressed_msaa;
+					     read_compressed_msaa || txf_add_offsets;
+
 	boolean src_loaded = FALSE;
 	unsigned sampler_src_reg = inst->Instruction.Opcode == TGSI_OPCODE_TXQ_LZ ? 0 : 1;
 	int8_t offset_x = 0, offset_y = 0, offset_z = 0;
@@ -4375,15 +4432,6 @@ static int tgsi_tex(struct r600_shader_ctx *ctx)
 		}
 	}
 
-	/* get offset values */
-	if (inst->Texture.NumOffsets) {
-		assert(inst->Texture.NumOffsets == 1);
-
-		offset_x = ctx->literals[4 * inst->TexOffsets[0].Index + inst->TexOffsets[0].SwizzleX] << 1;
-		offset_y = ctx->literals[4 * inst->TexOffsets[0].Index + inst->TexOffsets[0].SwizzleY] << 1;
-		offset_z = ctx->literals[4 * inst->TexOffsets[0].Index + inst->TexOffsets[0].SwizzleZ] << 1;
-	}
-
 	if (inst->Instruction.Opcode == TGSI_OPCODE_TXD) {
 		/* TGSI moves the sampler to src reg 3 for TXD */
 		sampler_src_reg = 3;
@@ -4736,6 +4784,77 @@ static int tgsi_tex(struct r600_shader_ctx *ctx)
 		src_gpr = ctx->temp_reg;
 	}
 
+	/* get offset values */
+	if (inst->Texture.NumOffsets) {
+		assert(inst->Texture.NumOffsets == 1);
+
+		if (txf_add_offsets) {
+			/* Add the offsets for texelFetch manually. */
+			const struct tgsi_texture_offset *off = inst->TexOffsets;
+
+			switch (inst->Texture.Texture) {
+			case TGSI_TEXTURE_3D:
+				memset(&alu, 0, sizeof(struct r600_bytecode_alu));
+				alu.op = ALU_OP2_ADD_INT;
+				alu.src[0].sel = src_gpr;
+				alu.src[0].chan = 2;
+				alu.src[1].sel = V_SQ_ALU_SRC_LITERAL;
+				alu.src[1].value = ctx->literals[4 * off[0].Index + off[0].SwizzleZ];
+				alu.dst.sel = src_gpr;
+				alu.dst.chan = 2;
+				alu.dst.write = 1;
+				alu.last = 1;
+				r = r600_bytecode_add_alu(ctx->bc, &alu);
+				if (r)
+					return r;
+				/* fall through */
+
+			case TGSI_TEXTURE_2D:
+			case TGSI_TEXTURE_SHADOW2D:
+			case TGSI_TEXTURE_RECT:
+			case TGSI_TEXTURE_SHADOWRECT:
+			case TGSI_TEXTURE_2D_ARRAY:
+			case TGSI_TEXTURE_SHADOW2D_ARRAY:
+				memset(&alu, 0, sizeof(struct r600_bytecode_alu));
+				alu.op = ALU_OP2_ADD_INT;
+				alu.src[0].sel = src_gpr;
+				alu.src[0].chan = 1;
+				alu.src[1].sel = V_SQ_ALU_SRC_LITERAL;
+				alu.src[1].value = ctx->literals[4 * off[0].Index + off[0].SwizzleY];
+				alu.dst.sel = src_gpr;
+				alu.dst.chan = 1;
+				alu.dst.write = 1;
+				alu.last = 1;
+				r = r600_bytecode_add_alu(ctx->bc, &alu);
+				if (r)
+					return r;
+				/* fall through */
+
+			case TGSI_TEXTURE_1D:
+			case TGSI_TEXTURE_SHADOW1D:
+			case TGSI_TEXTURE_1D_ARRAY:
+			case TGSI_TEXTURE_SHADOW1D_ARRAY:
+				memset(&alu, 0, sizeof(struct r600_bytecode_alu));
+				alu.op = ALU_OP2_ADD_INT;
+				alu.src[0].sel = src_gpr;
+				alu.src[1].sel = V_SQ_ALU_SRC_LITERAL;
+				alu.src[1].value = ctx->literals[4 * off[0].Index + off[0].SwizzleX];
+				alu.dst.sel = src_gpr;
+				alu.dst.write = 1;
+				alu.last = 1;
+				r = r600_bytecode_add_alu(ctx->bc, &alu);
+				if (r)
+					return r;
+				break;
+				/* texture offsets do not apply to other texture targets */
+			}
+		} else {
+			offset_x = ctx->literals[4 * inst->TexOffsets[0].Index + inst->TexOffsets[0].SwizzleX] << 1;
+			offset_y = ctx->literals[4 * inst->TexOffsets[0].Index + inst->TexOffsets[0].SwizzleY] << 1;
+			offset_z = ctx->literals[4 * inst->TexOffsets[0].Index + inst->TexOffsets[0].SwizzleZ] << 1;
+		}
+	}
+
 	/* Obtain the sample index for reading a compressed MSAA color texture.
 	 * To read the FMASK, we use the ldfptr instruction, which tells us
 	 * where the samples are stored.

src/gallium/drivers/r600/r600_shader.h

@@ -63,6 +63,7 @@ struct r600_shader {
 	boolean			vs_out_misc_write;
 	boolean			vs_out_point_size;
 	boolean			vs_out_layer;
+	boolean			vs_out_edgeflag;
 	boolean			has_txq_cube_array_z_comp;
 	boolean			uses_tex_buffers;
 	boolean                 gs_prim_id_input;

src/gallium/drivers/r600/r600_state.c

@@ -911,6 +911,10 @@ static void *r600_create_rs_state(struct pipe_context *ctx,
 		S_028810_ZCLIP_NEAR_DISABLE(!state->depth_clip) |
 		S_028810_ZCLIP_FAR_DISABLE(!state->depth_clip) |
 		S_028810_DX_LINEAR_ATTR_CLIP_ENA(1);
+	if (rctx->b.chip_class == R700) {
+		rs->pa_cl_clip_cntl |=
+			S_028810_DX_RASTERIZATION_KILL(state->rasterizer_discard);
+	}
 	rs->multisample_enable = state->multisample;
 
 	/* offset */
@@ -968,19 +972,25 @@ static void *r600_create_rs_state(struct pipe_context *ctx,
 			       S_028C08_PIX_CENTER_HALF(state->half_pixel_center) |
 			       S_028C08_QUANT_MODE(V_028C08_X_1_256TH));
 	r600_store_context_reg(&rs->buffer, R_028DFC_PA_SU_POLY_OFFSET_CLAMP, fui(state->offset_clamp));
-	r600_store_context_reg(&rs->buffer, R_028814_PA_SU_SC_MODE_CNTL,
-			       S_028814_PROVOKING_VTX_LAST(!state->flatshade_first) |
-			       S_028814_CULL_FRONT(state->cull_face & PIPE_FACE_FRONT ? 1 : 0) |
-			       S_028814_CULL_BACK(state->cull_face & PIPE_FACE_BACK ? 1 : 0) |
-			       S_028814_FACE(!state->front_ccw) |
-			       S_028814_POLY_OFFSET_FRONT_ENABLE(state->offset_tri) |
-			       S_028814_POLY_OFFSET_BACK_ENABLE(state->offset_tri) |
-			       S_028814_POLY_OFFSET_PARA_ENABLE(state->offset_tri) |
-			       S_028814_POLY_MODE(state->fill_front != PIPE_POLYGON_MODE_FILL ||
-						  state->fill_back != PIPE_POLYGON_MODE_FILL) |
-			       S_028814_POLYMODE_FRONT_PTYPE(r600_translate_fill(state->fill_front)) |
-			       S_028814_POLYMODE_BACK_PTYPE(r600_translate_fill(state->fill_back)));
-	r600_store_context_reg(&rs->buffer, R_028350_SX_MISC, S_028350_MULTIPASS(state->rasterizer_discard));
+
+	rs->pa_su_sc_mode_cntl = S_028814_PROVOKING_VTX_LAST(!state->flatshade_first) |
+				 S_028814_CULL_FRONT(state->cull_face & PIPE_FACE_FRONT ? 1 : 0) |
+				 S_028814_CULL_BACK(state->cull_face & PIPE_FACE_BACK ? 1 : 0) |
+				 S_028814_FACE(!state->front_ccw) |
+				 S_028814_POLY_OFFSET_FRONT_ENABLE(state->offset_tri) |
+				 S_028814_POLY_OFFSET_BACK_ENABLE(state->offset_tri) |
+				 S_028814_POLY_OFFSET_PARA_ENABLE(state->offset_tri) |
+				 S_028814_POLY_MODE(state->fill_front != PIPE_POLYGON_MODE_FILL ||
+									 state->fill_back != PIPE_POLYGON_MODE_FILL) |
+				 S_028814_POLYMODE_FRONT_PTYPE(r600_translate_fill(state->fill_front)) |
+				 S_028814_POLYMODE_BACK_PTYPE(r600_translate_fill(state->fill_back));
+	if (rctx->b.chip_class == R700) {
+		r600_store_context_reg(&rs->buffer, R_028814_PA_SU_SC_MODE_CNTL, rs->pa_su_sc_mode_cntl);
+	}
+	if (rctx->b.chip_class == R600) {
+		r600_store_context_reg(&rs->buffer, R_028350_SX_MISC,
+				       S_028350_MULTIPASS(state->rasterizer_discard));
+	}
 	return rs;
 }
 
@@ -2052,6 +2062,11 @@ static void r600_emit_db_misc_state(struct r600_context *rctx, struct r600_atom
 		db_render_control |= S_028D0C_DEPTH_CLEAR_ENABLE(1);
 	}
 
+	/* RV770 workaround for a hang with 8x MSAA. */
+	if (rctx->b.family == CHIP_RV770 && a->log_samples == 3) {
+		db_render_override |= S_028D10_MAX_TILES_IN_DTT(6);
+	}
+
 	r600_write_context_reg_seq(cs, R_028D0C_DB_RENDER_CONTROL, 2);
 	radeon_emit(cs, db_render_control); /* R_028D0C_DB_RENDER_CONTROL */
 	radeon_emit(cs, db_render_override); /* R_028D10_DB_RENDER_OVERRIDE */
@@ -2817,8 +2832,11 @@ void r600_init_atom_start_cs(struct r600_context *rctx)
 
 	r600_store_context_reg(cb, R_0288A4_SQ_PGM_RESOURCES_FS, 0);
 
+	if (rctx->b.chip_class == R700)
+		r600_store_context_reg(cb, R_028350_SX_MISC, 0);
 	if (rctx->b.chip_class == R700 && rctx->screen->b.has_streamout)
 		r600_store_context_reg(cb, R_028354_SX_SURFACE_SYNC, S_028354_SURFACE_SYNC_MASK(0xf));
+
 	r600_store_context_reg(cb, R_028800_DB_DEPTH_CONTROL, 0);
 	if (rctx->screen->b.has_streamout) {
 		r600_store_context_reg(cb, R_028B28_VGT_STRMOUT_DRAW_OPAQUE_OFFSET, 0);
@@ -2996,31 +3014,9 @@ void r600_update_vs_state(struct pipe_context *ctx, struct r600_pipe_shader *sha
 		S_02881C_VS_OUT_CCDIST0_VEC_ENA((rshader->clip_dist_write & 0x0F) != 0) |
 		S_02881C_VS_OUT_CCDIST1_VEC_ENA((rshader->clip_dist_write & 0xF0) != 0) |
 		S_02881C_VS_OUT_MISC_VEC_ENA(rshader->vs_out_misc_write) |
-		S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size);
-}
-
-static unsigned r600_conv_prim_to_gs_out(unsigned mode)
-{
-	static const int prim_conv[] = {
-		V_028A6C_OUTPRIM_TYPE_POINTLIST,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_LINESTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP,
-		V_028A6C_OUTPRIM_TYPE_TRISTRIP
-	};
-	assert(mode < Elements(prim_conv));
-
-	return prim_conv[mode];
+		S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size) |
+		S_02881C_USE_VTX_EDGE_FLAG(rshader->vs_out_edgeflag) |
+		S_02881C_USE_VTX_RENDER_TARGET_INDX(rshader->vs_out_layer);
 }
 
 void r600_update_gs_state(struct pipe_context *ctx, struct r600_pipe_shader *shader)

src/gallium/drivers/r600/r600_state_common.c

@@ -1408,6 +1408,25 @@ static void r600_draw_vbo(struct pipe_context *ctx, const struct pipe_draw_info
 		r600_emit_atom(rctx, rctx->atoms[i]);
 	}
 
+	/* On R6xx, CULL_FRONT=1 culls all points, lines, and rectangles,
+	 * even though it should have no effect on those. */
+	if (rctx->b.chip_class == R600 && rctx->rasterizer) {
+		unsigned su_sc_mode_cntl = rctx->rasterizer->pa_su_sc_mode_cntl;
+		unsigned prim = info.mode;
+
+		if (rctx->gs_shader) {
+			prim = rctx->gs_shader->current->shader.gs_output_prim;
+		}
+		prim = r600_conv_prim_to_gs_out(prim); /* decrease the number of types to 3 */
+
+		if (prim == V_028A6C_OUTPRIM_TYPE_POINTLIST ||
+		    prim == V_028A6C_OUTPRIM_TYPE_LINESTRIP ||
+		    info.mode == R600_PRIM_RECTANGLE_LIST) {
+			su_sc_mode_cntl &= C_028814_CULL_FRONT;
+		}
+		r600_write_context_reg(cs, R_028814_PA_SU_SC_MODE_CNTL, su_sc_mode_cntl);
+	}
+
 	/* Update start instance. */
 	if (rctx->last_start_instance != info.start_instance) {
 		r600_write_ctl_const(cs, R_03CFF4_SQ_VTX_START_INST_LOC, info.start_instance);

src/gallium/drivers/r600/r600d.h

@@ -543,6 +543,9 @@
 #define   S_028810_VTX_KILL_OR(x)                      (((x) & 0x1) << 21)
 #define   G_028810_VTX_KILL_OR(x)                      (((x) >> 21) & 0x1)
 #define   C_028810_VTX_KILL_OR                         0xFFDFFFFF
+#define   S_028810_DX_RASTERIZATION_KILL(x)            (((x) & 0x1) << 22) /* R700 only? */
+#define   G_028810_DX_RASTERIZATION_KILL(x)            (((x) >> 22) & 0x1)
+#define   C_028810_DX_RASTERIZATION_KILL               0xFFBFFFFF
 #define   S_028810_DX_LINEAR_ATTR_CLIP_ENA(x)          (((x) & 0x1) << 24)
 #define   G_028810_DX_LINEAR_ATTR_CLIP_ENA(x)          (((x) >> 24) & 0x1)
 #define   C_028810_DX_LINEAR_ATTR_CLIP_ENA             0xFEFFFFFF
@@ -818,6 +821,9 @@
 #define   S_028D10_IGNORE_SC_ZRANGE(x)                 (((x) & 0x1) << 17)
 #define   G_028D10_IGNORE_SC_ZRANGE(x)                 (((x) >> 17) & 0x1)
 #define   C_028D10_IGNORE_SC_ZRANGE                    0xFFFDFFFF
+#define   S_028D10_MAX_TILES_IN_DTT(x)                 (((x) & 0x1F) << 21)
+#define   G_028D10_MAX_TILES_IN_DTT(x)                 (((x) >> 21) & 0x1F)
+#define   C_028D10_MAX_TILES_IN_DTT                    0xFC1FFFFF
 #define R_02880C_DB_SHADER_CONTROL                    0x02880C
 #define   S_02880C_Z_EXPORT_ENABLE(x)                  (((x) & 0x1) << 0)
 #define   G_02880C_Z_EXPORT_ENABLE(x)                  (((x) >> 0) & 0x1)

src/gallium/drivers/radeon/r600_pipe_common.h

@@ -390,6 +390,15 @@ r600_resource_reference(struct r600_resource **ptr, struct r600_resource *res)
 				(struct pipe_resource *)res);
 }
 
+static inline unsigned r600_tex_aniso_filter(unsigned filter)
+{
+	if (filter <= 1)   return 0;
+	if (filter <= 2)   return 1;
+	if (filter <= 4)   return 2;
+	if (filter <= 8)   return 3;
+	 /* else */        return 4;
+}
+
 #define R600_ERR(fmt, args...) \
 	fprintf(stderr, "EE %s:%d %s - "fmt, __FILE__, __LINE__, __func__, ##args)
 

src/gallium/drivers/radeon/r600_texture.c

@@ -296,6 +296,12 @@ void r600_texture_get_fmask_info(struct r600_common_screen *rscreen,
 	fmask.nsamples = 1;
 	fmask.flags |= RADEON_SURF_FMASK;
 
+	/* Force 2D tiling if it wasn't set. This may occur when creating
+	 * FMASK for MSAA resolve on R6xx. On R6xx, the single-sample
+	 * destination buffer must have an FMASK too. */
+	fmask.flags = RADEON_SURF_CLR(fmask.flags, MODE);
+	fmask.flags |= RADEON_SURF_SET(RADEON_SURF_MODE_2D, MODE);
+
 	if (rscreen->chip_class >= SI) {
 		fmask.flags |= RADEON_SURF_HAS_TILE_MODE_INDEX;
 	}

src/gallium/drivers/radeonsi/si_descriptors.c

@@ -151,7 +151,7 @@ static void si_update_descriptors(struct si_context *sctx,
 			7 + /* copy */
 			(4 + desc->element_dw_size) * util_bitcount(desc->dirty_mask) + /* update */
 			4; /* pointer update */
-#if HAVE_LLVM >= 0x0305
+#if LLVM_SUPPORTS_GEOM_SHADERS
 		if (desc->shader_userdata_reg >= R_00B130_SPI_SHADER_USER_DATA_VS_0 &&
 		    desc->shader_userdata_reg < R_00B230_SPI_SHADER_USER_DATA_GS_0)
 			desc->atom.num_dw += 4; /* second pointer update */
@@ -176,7 +176,7 @@ static void si_emit_shader_pointer(struct si_context *sctx,
 	radeon_emit(cs, va);
 	radeon_emit(cs, va >> 32);
 
-#if HAVE_LLVM >= 0x0305
+#if LLVM_SUPPORTS_GEOM_SHADERS
 	if (desc->shader_userdata_reg >= R_00B130_SPI_SHADER_USER_DATA_VS_0 &&
 	    desc->shader_userdata_reg < R_00B230_SPI_SHADER_USER_DATA_GS_0) {
 		radeon_emit(cs, PKT3(PKT3_SET_SH_REG, 2, 0));

src/gallium/drivers/radeonsi/si_pipe.c

@@ -269,7 +269,7 @@ static int si_get_param(struct pipe_screen* pscreen, enum pipe_cap param)
 		return 256;
 
 	case PIPE_CAP_GLSL_FEATURE_LEVEL:
-		return HAVE_LLVM >= 0x0305 ? 330 : 140;
+		return (LLVM_SUPPORTS_GEOM_SHADERS) ? 330 : 140;
 
 	case PIPE_CAP_TEXTURE_BUFFER_OFFSET_ALIGNMENT:
 		return 1;
@@ -307,11 +307,14 @@ static int si_get_param(struct pipe_screen* pscreen, enum pipe_cap param)
 
 	/* Texturing. */
 	case PIPE_CAP_MAX_TEXTURE_2D_LEVELS:
-	case PIPE_CAP_MAX_TEXTURE_3D_LEVELS:
 	case PIPE_CAP_MAX_TEXTURE_CUBE_LEVELS:
-			return 15;
+		return 15; /* 16384 */
+	case PIPE_CAP_MAX_TEXTURE_3D_LEVELS:
+		/* textures support 8192, but layered rendering supports 2048 */
+		return 12;
 	case PIPE_CAP_MAX_TEXTURE_ARRAY_LAYERS:
-		return 16384;
+		/* textures support 8192, but layered rendering supports 2048 */
+		return 2048;
 	case PIPE_CAP_MAX_COMBINED_SAMPLERS:
 		return HAVE_LLVM >= 0x0305 ? 48 : 32;
 
@@ -346,7 +349,7 @@ static int si_get_shader_param(struct pipe_screen* pscreen, unsigned shader, enu
 	case PIPE_SHADER_VERTEX:
 		break;
 	case PIPE_SHADER_GEOMETRY:
-#if HAVE_LLVM < 0x0305
+#if !(LLVM_SUPPORTS_GEOM_SHADERS)
 		return 0;
 #endif
 		break;

src/gallium/drivers/radeonsi/si_pipe.h

@@ -39,6 +39,10 @@
 
 #define SI_MAX_DRAW_CS_DWORDS 18
 
+#define LLVM_SUPPORTS_GEOM_SHADERS \
+	((HAVE_LLVM >= 0x0305) || \
+	(HAVE_LLVM == 0x0304 && LLVM_VERSION_PATCH >= 1))
+
 struct si_pipe_compute;
 
 struct si_screen {

src/gallium/drivers/radeonsi/si_state.c

@@ -2307,7 +2307,7 @@ static void *si_create_fs_state(struct pipe_context *ctx,
 	return si_create_shader_state(ctx, state, PIPE_SHADER_FRAGMENT);
 }
 
-#if HAVE_LLVM >= 0x0305
+#if LLVM_SUPPORTS_GEOM_SHADERS
 
 static void *si_create_gs_state(struct pipe_context *ctx,
 				const struct pipe_shader_state *state)
@@ -2337,7 +2337,7 @@ static void si_bind_vs_shader(struct pipe_context *ctx, void *state)
 	sctx->vs_shader = sel;
 }
 
-#if HAVE_LLVM >= 0x0305
+#if LLVM_SUPPORTS_GEOM_SHADERS
 
 static void si_bind_gs_shader(struct pipe_context *ctx, void *state)
 {
@@ -2396,7 +2396,7 @@ static void si_delete_vs_shader(struct pipe_context *ctx, void *state)
 	si_delete_shader_selector(ctx, sel);
 }
 
-#if HAVE_LLVM >= 0x0305
+#if LLVM_SUPPORTS_GEOM_SHADERS
 
 static void si_delete_gs_shader(struct pipe_context *ctx, void *state)
 {
@@ -2723,16 +2723,15 @@ static void *si_create_sampler_state(struct pipe_context *ctx,
 	rstate->val[0] = (S_008F30_CLAMP_X(si_tex_wrap(state->wrap_s)) |
 			  S_008F30_CLAMP_Y(si_tex_wrap(state->wrap_t)) |
 			  S_008F30_CLAMP_Z(si_tex_wrap(state->wrap_r)) |
-			  (state->max_anisotropy & 0x7) << 9 | /* XXX */
+			  r600_tex_aniso_filter(state->max_anisotropy) << 9 |
 			  S_008F30_DEPTH_COMPARE_FUNC(si_tex_compare(state->compare_func)) |
 			  S_008F30_FORCE_UNNORMALIZED(!state->normalized_coords) |
-			  aniso_flag_offset << 16 | /* XXX */
 			  S_008F30_DISABLE_CUBE_WRAP(!state->seamless_cube_map));
 	rstate->val[1] = (S_008F34_MIN_LOD(S_FIXED(CLAMP(state->min_lod, 0, 15), 8)) |
 			  S_008F34_MAX_LOD(S_FIXED(CLAMP(state->max_lod, 0, 15), 8)));
 	rstate->val[2] = (S_008F38_LOD_BIAS(S_FIXED(CLAMP(state->lod_bias, -16, 16), 8)) |
-			  S_008F38_XY_MAG_FILTER(si_tex_filter(state->mag_img_filter)) |
-			  S_008F38_XY_MIN_FILTER(si_tex_filter(state->min_img_filter)) |
+			  S_008F38_XY_MAG_FILTER(si_tex_filter(state->mag_img_filter) | aniso_flag_offset) |
+			  S_008F38_XY_MIN_FILTER(si_tex_filter(state->min_img_filter) | aniso_flag_offset) |
 			  S_008F38_MIP_FILTER(si_tex_mipfilter(state->min_mip_filter)));
 	rstate->val[3] = S_008F3C_BORDER_COLOR_TYPE(border_color_type);
 
@@ -2890,7 +2889,7 @@ static void si_bind_vs_sampler_states(struct pipe_context *ctx, unsigned count,
 	si_set_sampler_states(sctx, pm4, count, states,
 			      &sctx->samplers[PIPE_SHADER_VERTEX],
 			      R_00B130_SPI_SHADER_USER_DATA_VS_0);
-#if HAVE_LLVM >= 0x0305
+#if LLVM_SUPPORTS_GEOM_SHADERS
 	si_set_sampler_states(sctx, pm4, count, states,
 			      &sctx->samplers[PIPE_SHADER_VERTEX],
 			      R_00B330_SPI_SHADER_USER_DATA_ES_0);
@@ -3166,7 +3165,7 @@ void si_init_state_functions(struct si_context *sctx)
 	sctx->b.b.bind_fs_state = si_bind_ps_shader;
 	sctx->b.b.delete_vs_state = si_delete_vs_shader;
 	sctx->b.b.delete_fs_state = si_delete_ps_shader;
-#if HAVE_LLVM >= 0x0305
+#if LLVM_SUPPORTS_GEOM_SHADERS
 	sctx->b.b.create_gs_state = si_create_gs_state;
 	sctx->b.b.bind_gs_state = si_bind_gs_shader;
 	sctx->b.b.delete_gs_state = si_delete_gs_shader;

src/gallium/drivers/softpipe/sp_texture.c

@@ -60,7 +60,7 @@ softpipe_resource_layout(struct pipe_screen *screen,
    unsigned width = pt->width0;
    unsigned height = pt->height0;
    unsigned depth = pt->depth0;
-   unsigned buffer_size = 0;
+   uint64_t buffer_size = 0;
 
    for (level = 0; level <= pt->last_level; level++) {
       unsigned slices;
@@ -76,8 +76,8 @@ softpipe_resource_layout(struct pipe_screen *screen,
 
       spr->level_offset[level] = buffer_size;
 
-      buffer_size += (util_format_get_nblocksy(pt->format, height) *
-                      slices * spr->stride[level]);
+      buffer_size += (uint64_t) util_format_get_nblocksy(pt->format, height) *
+                     slices * spr->stride[level];
 
       width  = u_minify(width, 1);
       height = u_minify(height, 1);

src/gallium/drivers/svga/svga_context.c

@@ -92,6 +92,8 @@ struct pipe_context *svga_context_create( struct pipe_screen *screen,
    if (svga == NULL)
       goto no_svga;
 
+   LIST_INITHEAD(&svga->dirty_buffers);
+
    svga->pipe.screen = screen;
    svga->pipe.priv = priv;
    svga->pipe.destroy = svga_destroy;
@@ -154,8 +156,6 @@ struct pipe_context *svga_context_create( struct pipe_screen *screen,
 
    svga->dirty = ~0;
 
-   LIST_INITHEAD(&svga->dirty_buffers);
-
    return &svga->pipe;
 
 no_state:

src/gallium/drivers/svga/svga_state_fs.c

@@ -235,15 +235,43 @@ make_fs_key(const struct svga_context *svga,
    if (svga->curr.blend->need_white_fragments) {
       key->white_fragments = 1;
    }
-   
+
+#ifdef DEBUG
+   /*
+    * We expect a consistent set of samplers and sampler views.
+    * Do some debug checks/warnings here.
+    */
+   {
+      static boolean warned = FALSE;
+      unsigned i, n = MAX2(svga->curr.num_sampler_views,
+                           svga->curr.num_samplers);
+      /* Only warn once to prevent too much debug output */
+      if (!warned) {
+         if (svga->curr.num_sampler_views != svga->curr.num_samplers) {
+            debug_printf("svga: mismatched number of sampler views (%u) "
+                         "vs. samplers (%u)\n",
+                         svga->curr.num_sampler_views,
+                         svga->curr.num_samplers);
+         }
+         for (i = 0; i < n; i++) {
+            if ((svga->curr.sampler_views[i] == NULL) !=
+                (svga->curr.sampler[i] == NULL))
+               debug_printf("sampler_view[%u] = %p but sampler[%u] = %p\n",
+                            i, svga->curr.sampler_views[i],
+                            i, svga->curr.sampler[i]);
+         }
+         warned = TRUE;
+      }
+   }
+#endif
+
    /* XXX: want to limit this to the textures that the shader actually
     * refers to.
     *
     * SVGA_NEW_TEXTURE_BINDING | SVGA_NEW_SAMPLER
     */
    for (i = 0; i < svga->curr.num_sampler_views; i++) {
-      if (svga->curr.sampler_views[i]) {
-         assert(svga->curr.sampler[i]);
+      if (svga->curr.sampler_views[i] && svga->curr.sampler[i]) {
          assert(svga->curr.sampler_views[i]->texture);
          key->tex[i].texture_target = svga->curr.sampler_views[i]->texture->target;
          if (!svga->curr.sampler[i]->normalized_coords) {
@@ -262,7 +290,7 @@ make_fs_key(const struct svga_context *svga,
 
    idx = 0;
    for (i = 0; i < svga->curr.num_samplers; ++i) {
-      if (svga->curr.sampler_views[i]) {
+      if (svga->curr.sampler_views[i] && svga->curr.sampler[i]) {
          struct pipe_resource *tex = svga->curr.sampler_views[i]->texture;
          struct svga_texture *stex = svga_texture(tex);
          SVGA3dSurfaceFormat format = stex->key.format;

src/gallium/drivers/svga/svga_state_tss.c

@@ -85,7 +85,7 @@ update_tss_binding(struct svga_context *svga,
       struct pipe_sampler_view *sv = svga->curr.sampler_views[i];
 
       /* get min max lod */
-      if (sv) {
+      if (sv && s) {
          min_lod = MAX2(0, (s->view_min_lod + sv->u.tex.first_level));
          max_lod = MIN2(s->view_max_lod + sv->u.tex.first_level,
                         sv->texture->last_level);

src/gallium/include/pipe/p_config.h

@@ -157,6 +157,16 @@
 # define PIPE_ARCH_BIG_ENDIAN
 #endif
 
+#elif defined(__OpenBSD__)
+#include <sys/types.h>
+#include <machine/endian.h>
+
+#if _BYTE_ORDER == _LITTLE_ENDIAN
+# define PIPE_ARCH_LITTLE_ENDIAN
+#elif _BYTE_ORDER == _BIG_ENDIAN
+# define PIPE_ARCH_BIG_ENDIAN
+#endif
+
 #else
 
 #if defined(PIPE_ARCH_X86) || defined(PIPE_ARCH_X86_64) || defined(PIPE_ARCH_ARM) || defined(PIPE_ARCH_AARCH64)

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

@@ -112,12 +112,13 @@ namespace {
 #endif
 
    llvm::Module *
-   compile(const std::string &source, const std::string &name,
-           const std::string &triple, const std::string &processor,
-           const std::string &opts, clang::LangAS::Map& address_spaces) {
+   compile(llvm::LLVMContext &llvm_ctx, const std::string &source,
+           const std::string &name, const std::string &triple,
+           const std::string &processor, const std::string &opts,
+           clang::LangAS::Map& address_spaces) {
 
       clang::CompilerInstance c;
-      clang::EmitLLVMOnlyAction act(&llvm::getGlobalContext());
+      clang::EmitLLVMOnlyAction act(&llvm_ctx);
       std::string log;
       llvm::raw_string_ostream s_log(log);
       std::string libclc_path = LIBCLC_LIBEXECDIR + processor + "-"
@@ -182,6 +183,11 @@ namespace {
       c.getLangOpts().NoBuiltin = true;
       c.getTargetOpts().Triple = triple;
       c.getTargetOpts().CPU = processor;
+
+      // This is a workaround for a Clang bug which causes the number
+      // of warnings and errors to be printed to stderr.
+      // http://www.llvm.org/bugs/show_bug.cgi?id=19735
+      c.getDiagnosticOpts().ShowCarets = false;
 #if HAVE_LLVM <= 0x0301
       c.getInvocation().setLangDefaults(clang::IK_OpenCL);
 #else
@@ -379,10 +385,12 @@ clover::compile_program_llvm(const compat::string &source,
                       target.size() - processor_str_len - 1);
    clang::LangAS::Map address_spaces;
 
+   llvm::LLVMContext llvm_ctx;
+
    // The input file name must have the .cl extension in order for the
    // CompilerInvocation class to recognize it as an OpenCL source file.
-   llvm::Module *mod = compile(source, "input.cl", triple, processor, opts,
-                                                                address_spaces);
+   llvm::Module *mod = compile(llvm_ctx, source, "input.cl", triple, processor,
+                               opts, address_spaces);
 
    find_kernels(mod, kernels);
 

src/gallium/state_trackers/osmesa/osmesa.c

@@ -480,12 +480,13 @@ osmesa_create_buffer(enum pipe_format color_format,
 
 
 /**
- * Search linked list for a buffer with matching pixel formats.
+ * Search linked list for a buffer with matching pixel formats and size.
  */
 static struct osmesa_buffer *
 osmesa_find_buffer(enum pipe_format color_format,
                    enum pipe_format ds_format,
-                   enum pipe_format accum_format)
+                   enum pipe_format accum_format,
+                   GLsizei width, GLsizei height)
 {
    struct osmesa_buffer *b;
 
@@ -493,7 +494,9 @@ osmesa_find_buffer(enum pipe_format color_format,
    for (b = BufferList; b; b = b->next) {
       if (b->visual.color_format == color_format &&
           b->visual.depth_stencil_format == ds_format &&
-          b->visual.accum_format == accum_format) {
+          b->visual.accum_format == accum_format &&
+          b->width == width &&
+          b->height == height) {
          return b;
       }
    }
@@ -673,7 +676,7 @@ OSMesaMakeCurrent(OSMesaContext osmesa, void *buffer, GLenum type,
    /* See if we already have a buffer that uses these pixel formats */
    osbuffer = osmesa_find_buffer(color_format,
                                  osmesa->depth_stencil_format,
-                                 osmesa->accum_format);
+                                 osmesa->accum_format, width, height);
    if (!osbuffer) {
       /* Existing buffer found, create new buffer */
       osbuffer = osmesa_create_buffer(color_format,

src/gallium/state_trackers/xa/xa_context.c

@@ -78,6 +78,8 @@ xa_context_destroy(struct xa_context *r)
     }
 
     xa_ctx_sampler_views_destroy(r);
+    if (r->srf)
+        pipe_surface_reference(&r->srf, NULL);
 
     if (r->cso) {
 	cso_release_all(r->cso);
@@ -185,8 +187,15 @@ xa_ctx_srf_create(struct xa_context *ctx, struct xa_surface *dst)
     struct pipe_screen *screen = ctx->pipe->screen;
     struct pipe_surface srf_templ;
 
-    if (ctx->srf)
-	return -XA_ERR_INVAL;
+    /*
+     * Cache surfaces unless we change render target
+     */
+    if (ctx->srf) {
+        if (ctx->srf->texture == dst->tex)
+            return XA_ERR_NONE;
+
+        pipe_surface_reference(&ctx->srf, NULL);
+    }
 
     if (!screen->is_format_supported(screen,  dst->tex->format,
 				     PIPE_TEXTURE_2D, 0,
@@ -204,14 +213,17 @@ xa_ctx_srf_create(struct xa_context *ctx, struct xa_surface *dst)
 void
 xa_ctx_srf_destroy(struct xa_context *ctx)
 {
-    pipe_surface_reference(&ctx->srf, NULL);
+    /*
+     * Cache surfaces unless we change render target.
+     * Final destruction on context destroy.
+     */
 }
 
 XA_EXPORT int
 xa_copy_prepare(struct xa_context *ctx,
 		struct xa_surface *dst, struct xa_surface *src)
 {
-    if (src == dst || ctx->srf != NULL)
+    if (src == dst)
 	return -XA_ERR_INVAL;
 
     if (src->tex->format != dst->tex->format) {

src/gallium/state_trackers/xa/xa_renderer.c

@@ -406,6 +406,9 @@ renderer_copy_prepare(struct xa_context *r,
 				       PIPE_BIND_RENDER_TARGET));
     (void)screen;
 
+    renderer_bind_destination(r, dst_surface,
+			      dst_surface->width, dst_surface->height);
+
     /* set misc state we care about */
     {
 	struct pipe_blend_state blend;
@@ -422,6 +425,7 @@ renderer_copy_prepare(struct xa_context *r,
     /* sampler */
     {
 	struct pipe_sampler_state sampler;
+        const struct pipe_sampler_state *p_sampler = &sampler;
 
 	memset(&sampler, 0, sizeof(sampler));
 	sampler.wrap_s = PIPE_TEX_WRAP_CLAMP_TO_EDGE;
@@ -431,13 +435,10 @@ renderer_copy_prepare(struct xa_context *r,
 	sampler.min_img_filter = PIPE_TEX_FILTER_NEAREST;
 	sampler.mag_img_filter = PIPE_TEX_FILTER_NEAREST;
 	sampler.normalized_coords = 1;
-	cso_single_sampler(r->cso, PIPE_SHADER_FRAGMENT, 0, &sampler);
-	cso_single_sampler_done(r->cso, PIPE_SHADER_FRAGMENT);
+        cso_set_samplers(r->cso, PIPE_SHADER_FRAGMENT, 1, &p_sampler);
+        r->num_bound_samplers = 1;
     }
 
-    renderer_bind_destination(r, dst_surface,
-			      dst_surface->width, dst_surface->height);
-
     /* texture/sampler view */
     {
 	struct pipe_sampler_view templ;

src/gallium/targets/haiku-softpipe/GalliumContext.cpp

@@ -44,9 +44,15 @@ extern "C" {
 
 
 static void
-hgl_viewport(struct gl_context* glContext, GLint x, GLint y,
-	GLsizei width, GLsizei height)
+hgl_viewport(struct gl_context* glContext)
 {
+	// TODO: We should try to eliminate this function
+
+	GLint x = glContext->ViewportArray[0].X;
+	GLint y = glContext->ViewportArray[0].Y;
+	GLint width = glContext->ViewportArray[0].Width;
+	GLint height = glContext->ViewportArray[0].Height;
+
 	TRACE("%s(glContext: %p, x: %d, y: %d, w: %d, h: %d\n", __func__,
 		glContext, x, y, width, height);
 
@@ -525,7 +531,7 @@ GalliumContext::ResizeViewport(int32 width, int32 height)
 	for (context_id i = 0; i < CONTEXT_MAX; i++) {
 		if (fContext[i] && fContext[i]->st) {
 			struct st_context *stContext = (struct st_context*)fContext[i]->st;
-			_mesa_set_viewport(stContext->ctx, 0, 0, width, height);
+			_mesa_set_viewport(stContext->ctx, 0, 0, 0, width, height);
         		st_manager_validate_framebuffers(stContext);
 		}
 	}

src/gallium/winsys/radeon/drm/radeon_drm_winsys.c

@@ -337,7 +337,8 @@ static boolean do_winsys_init(struct radeon_drm_winsys *ws)
 
     /* Check for dma */
     ws->info.r600_has_dma = FALSE;
-    if (ws->info.chip_class >= R700 && ws->info.drm_minor >= 27) {
+    /* DMA is disabled on R700. There is IB corruption and hangs. */
+    if (ws->info.chip_class >= EVERGREEN && ws->info.drm_minor >= 27) {
         ws->info.r600_has_dma = TRUE;
     }
 

src/gallium/winsys/svga/drm/vmw_screen.h

@@ -74,6 +74,7 @@ struct vmw_winsys_screen
       struct vmw_cap_3d *cap_3d;
       uint64_t max_mob_memory;
       uint64_t max_surface_memory;
+      boolean have_drm_2_6;
    } ioctl;
 
    struct {
@@ -138,10 +139,11 @@ vmw_ioctl_gb_surface_create(struct vmw_winsys_screen *vws,
 
 int
 vmw_ioctl_gb_surface_ref(struct vmw_winsys_screen *vws,
-                         uint32_t handle,
+                         const struct winsys_handle *whandle,
                          SVGA3dSurfaceFlags *flags,
                          SVGA3dSurfaceFormat *format,
                          uint32_t *numMipLevels,
+                         uint32_t *handle,
                          struct vmw_region **p_region);
 
 void

src/gallium/winsys/svga/drm/vmw_screen_dri.c

@@ -183,10 +183,11 @@ vmw_drm_gb_surface_from_handle(struct svga_winsys_screen *sws,
     struct vmw_buffer_desc desc;
     struct pb_manager *provider = vws->pools.gmr;
     struct pb_buffer *pb_buf;
+    uint32_t handle;
     int ret;
 
-    ret = vmw_ioctl_gb_surface_ref(vws, whandle->handle, &flags, format,
-                                   &mip_levels, &desc.region);
+    ret = vmw_ioctl_gb_surface_ref(vws, whandle, &flags, format,
+                                   &mip_levels, &handle, &desc.region);
 
     if (ret) {
 	fprintf(stderr, "Failed referencing shared surface. SID %d.\n"
@@ -209,7 +210,7 @@ vmw_drm_gb_surface_from_handle(struct svga_winsys_screen *sws,
     pipe_reference_init(&vsrf->refcnt, 1);
     p_atomic_set(&vsrf->validated, 0);
     vsrf->screen = vws;
-    vsrf->sid = whandle->handle;
+    vsrf->sid = handle;
     vsrf->size = vmw_region_size(desc.region);
 
     /*

src/gallium/winsys/svga/drm/vmw_screen_ioctl.c

@@ -37,6 +37,7 @@
 #include "util/u_memory.h"
 #include "util/u_math.h"
 #include "svgadump/svga_dump.h"
+#include "state_tracker/drm_driver.h"
 #include "vmw_screen.h"
 #include "vmw_context.h"
 #include "vmw_fence.h"
@@ -250,6 +251,63 @@ out_fail_create:
    return SVGA3D_INVALID_ID;
 }
 
+/**
+ * vmw_ioctl_surface_req - Fill in a struct surface_req
+ *
+ * @vws: Winsys screen
+ * @whandle: Surface handle
+ * @req: The struct surface req to fill in
+ * @needs_unref: This call takes a kernel surface reference that needs to
+ * be unreferenced.
+ *
+ * Returns 0 on success, negative error type otherwise.
+ * Fills in the surface_req structure according to handle type and kernel
+ * capabilities.
+ */
+static int
+vmw_ioctl_surface_req(const struct vmw_winsys_screen *vws,
+                      const struct winsys_handle *whandle,
+                      struct drm_vmw_surface_arg *req,
+                      boolean *needs_unref)
+{
+   int ret;
+
+   switch(whandle->type) {
+   case DRM_API_HANDLE_TYPE_SHARED:
+   case DRM_API_HANDLE_TYPE_KMS:
+      *needs_unref = FALSE;
+      req->handle_type = DRM_VMW_HANDLE_LEGACY;
+      req->sid = whandle->handle;
+      break;
+   case DRM_API_HANDLE_TYPE_FD:
+      if (!vws->ioctl.have_drm_2_6) {
+         uint32_t handle;
+
+         ret = drmPrimeFDToHandle(vws->ioctl.drm_fd, whandle->handle, &handle);
+         if (ret) {
+            vmw_error("Failed to get handle from prime fd %d.\n",
+                      (int) whandle->handle);
+            return -EINVAL;
+         }
+
+         *needs_unref = TRUE;
+         req->handle_type = DRM_VMW_HANDLE_LEGACY;
+         req->sid = handle;
+      } else {
+         *needs_unref = FALSE;
+         req->handle_type = DRM_VMW_HANDLE_PRIME;
+         req->sid = whandle->handle;
+      }
+      break;
+   default:
+      vmw_error("Attempt to import unsupported handle type %d.\n",
+                whandle->type);
+      return -EINVAL;
+   }
+
+   return 0;
+}
+
 /**
  * vmw_ioctl_gb_surface_ref - Put a reference on a guest-backed surface and
  * get surface information
@@ -266,16 +324,18 @@ out_fail_create:
  */
 int
 vmw_ioctl_gb_surface_ref(struct vmw_winsys_screen *vws,
-                         uint32_t handle,
+                         const struct winsys_handle *whandle,
                          SVGA3dSurfaceFlags *flags,
                          SVGA3dSurfaceFormat *format,
                          uint32_t *numMipLevels,
+                         uint32_t *handle,
                          struct vmw_region **p_region)
 {
    union drm_vmw_gb_surface_reference_arg s_arg;
    struct drm_vmw_surface_arg *req = &s_arg.req;
    struct drm_vmw_gb_surface_ref_rep *rep = &s_arg.rep;
    struct vmw_region *region = NULL;
+   boolean needs_unref = FALSE;
    int ret;
 
    vmw_printf("%s flags %d format %d\n", __FUNCTION__, flags, format);
@@ -286,8 +346,11 @@ vmw_ioctl_gb_surface_ref(struct vmw_winsys_screen *vws,
       return -ENOMEM;
 
    memset(&s_arg, 0, sizeof(s_arg));
-   req->sid = handle;
+   ret = vmw_ioctl_surface_req(vws, whandle, req, &needs_unref);
+   if (ret)
+      goto out_fail_req;
 
+   *handle = req->sid;
    ret = drmCommandWriteRead(vws->ioctl.drm_fd, DRM_VMW_GB_SURFACE_REF,
 			     &s_arg, sizeof(s_arg));
 
@@ -300,12 +363,19 @@ vmw_ioctl_gb_surface_ref(struct vmw_winsys_screen *vws,
    region->size = rep->crep.backup_size;
    *p_region = region;
 
+   *handle = rep->crep.handle;
    *flags = rep->creq.svga3d_flags;
    *format = rep->creq.format;
    *numMipLevels = rep->creq.mip_levels;
 
+   if (needs_unref)
+      vmw_ioctl_surface_destroy(vws, *handle);
+
    return 0;
 out_fail_ref:
+   if (needs_unref)
+      vmw_ioctl_surface_destroy(vws, *handle);
+out_fail_req:
    if (region)
       FREE(region);
    return ret;
@@ -772,6 +842,8 @@ vmw_ioctl_init(struct vmw_winsys_screen *vws)
 
    have_drm_2_5 = version->version_major > 2 ||
       (version->version_major == 2 && version->version_minor > 4);
+   vws->ioctl.have_drm_2_6 = version->version_major > 2 ||
+      (version->version_major == 2 && version->version_minor > 5);
 
    memset(&gp_arg, 0, sizeof(gp_arg));
    gp_arg.param = DRM_VMW_PARAM_3D;

src/gallium/winsys/svga/drm/vmw_screen_pools.c

@@ -76,15 +76,23 @@ vmw_pools_cleanup(struct vmw_winsys_screen *vws)
  *
  * Typically this pool should be created on demand when we
  * detect that the app will be using queries. There's nothing
- * special with this pool other than the backing kernel buffer size,
- * which is limited to 8192.
+ * special with this pool other than the backing kernel buffer sizes,
+ * which are limited to 8192.
+ * If there is a performance issue with allocation and freeing of the
+ * query slabs, it should be easily fixable by allocating them out
+ * of a buffer cache.
  */
 boolean
 vmw_query_pools_init(struct vmw_winsys_screen *vws)
 {
-   vws->pools.query_mm = mm_bufmgr_create(vws->pools.gmr,
-					  VMW_QUERY_POOL_SIZE,
-					  3 /* 8 alignment */);
+   struct pb_desc desc;
+
+   desc.alignment = 16;
+   desc.usage = ~(VMW_BUFFER_USAGE_SHARED | VMW_BUFFER_USAGE_SYNC);
+
+   vws->pools.query_mm = pb_slab_range_manager_create(vws->pools.gmr, 16, 128,
+                                                      VMW_QUERY_POOL_SIZE,
+                                                      &desc);
    if (!vws->pools.query_mm)
       return FALSE;
 

src/gallium/winsys/svga/drm/vmwgfx_drm.h

@@ -87,6 +87,16 @@
 #define DRM_VMW_PARAM_MAX_SURF_MEMORY  7
 #define DRM_VMW_PARAM_3D_CAPS_SIZE     8
 #define DRM_VMW_PARAM_MAX_MOB_MEMORY   9
+#define DRM_VMW_PARAM_MAX_MOB_SIZE     10
+
+/**
+ * enum drm_vmw_handle_type - handle type for ref ioctls
+ *
+ */
+enum drm_vmw_handle_type {
+	DRM_VMW_HANDLE_LEGACY = 0,
+	DRM_VMW_HANDLE_PRIME = 1
+};
 
 /**
  * struct drm_vmw_getparam_arg
@@ -176,6 +186,7 @@ struct drm_vmw_surface_create_req {
  * struct drm_wmv_surface_arg
  *
  * @sid: Surface id of created surface or surface to destroy or reference.
+ * @handle_type: Handle type for DRM_VMW_REF_SURFACE Ioctl.
  *
  * Output data from the DRM_VMW_CREATE_SURFACE Ioctl.
  * Input argument to the DRM_VMW_UNREF_SURFACE Ioctl.
@@ -184,7 +195,7 @@ struct drm_vmw_surface_create_req {
 
 struct drm_vmw_surface_arg {
 	int32_t sid;
-	uint32_t pad64;
+	enum drm_vmw_handle_type handle_type;
 };
 
 /**

src/gbm/backends/dri/gbm_dri.c

@@ -684,6 +684,8 @@ dri_device_create(int fd)
    int ret;
 
    dri = calloc(1, sizeof *dri);
+   if (!dri)
+      return NULL;
 
    dri->base.base.fd = fd;
    dri->base.base.bo_create = gbm_dri_bo_create;

src/glsl/ast_to_hir.cpp

@@ -119,6 +119,11 @@ _mesa_ast_to_hir(exec_list *instructions, struct _mesa_glsl_parse_state *state)
       instructions->push_head(var);
    }
 
+   /* Figure out if gl_FragCoord is actually used in fragment shader */
+   ir_variable *const var = state->symbols->get_variable("gl_FragCoord");
+   if (var != NULL)
+      state->fs_uses_gl_fragcoord = var->data.used;
+
    /* From section 7.1 (Built-In Language Variables) of the GLSL 4.10 spec:
     *
     *     If multiple shaders using members of a built-in block belonging to
@@ -2203,6 +2208,34 @@ validate_explicit_location(const struct ast_type_qualifier *qual,
    return;
 }
 
+static inline const char*
+get_layout_qualifier_string(bool origin_upper_left, bool pixel_center_integer)
+{
+   if (origin_upper_left && pixel_center_integer)
+      return "origin_upper_left, pixel_center_integer";
+   else if (origin_upper_left)
+      return "origin_upper_left";
+   else if (pixel_center_integer)
+      return "pixel_center_integer";
+   else
+      return " ";
+}
+
+static inline bool
+is_conflicting_fragcoord_redeclaration(struct _mesa_glsl_parse_state *state,
+                                       const struct ast_type_qualifier *qual)
+{
+   /* If gl_FragCoord was previously declared, and the qualifiers were
+    * different in any way, return true.
+    */
+   if (state->fs_redeclares_gl_fragcoord) {
+      return (state->fs_pixel_center_integer != qual->flags.q.pixel_center_integer
+         || state->fs_origin_upper_left != qual->flags.q.origin_upper_left);
+   }
+
+   return false;
+}
+
 static void
 apply_type_qualifier_to_variable(const struct ast_type_qualifier *qual,
 				 ir_variable *var,
@@ -2357,6 +2390,53 @@ apply_type_qualifier_to_variable(const struct ast_type_qualifier *qual,
 		       qual_string);
    }
 
+   if (var->name != NULL && strcmp(var->name, "gl_FragCoord") == 0) {
+
+      /* Section 4.3.8.1, page 39 of GLSL 1.50 spec says:
+       *
+       *    "Within any shader, the first redeclarations of gl_FragCoord
+       *     must appear before any use of gl_FragCoord."
+       *
+       * Generate a compiler error if above condition is not met by the
+       * fragment shader.
+       */
+      ir_variable *earlier = state->symbols->get_variable("gl_FragCoord");
+      if (earlier != NULL &&
+          earlier->data.used &&
+          !state->fs_redeclares_gl_fragcoord) {
+         _mesa_glsl_error(loc, state,
+                          "gl_FragCoord used before its first redeclaration "
+                          "in fragment shader");
+      }
+
+      /* Make sure all gl_FragCoord redeclarations specify the same layout
+       * qualifiers.
+       */
+      if (is_conflicting_fragcoord_redeclaration(state, qual)) {
+         const char *const qual_string =
+            get_layout_qualifier_string(qual->flags.q.origin_upper_left,
+                                        qual->flags.q.pixel_center_integer);
+
+         const char *const state_string =
+            get_layout_qualifier_string(state->fs_origin_upper_left,
+                                        state->fs_pixel_center_integer);
+
+         _mesa_glsl_error(loc, state,
+                          "gl_FragCoord redeclared with different layout "
+                          "qualifiers (%s) and (%s) ",
+                          state_string,
+                          qual_string);
+      }
+      state->fs_origin_upper_left = qual->flags.q.origin_upper_left;
+      state->fs_pixel_center_integer = qual->flags.q.pixel_center_integer;
+      state->fs_redeclares_gl_fragcoord_with_no_layout_qualifiers =
+         !qual->flags.q.origin_upper_left && !qual->flags.q.pixel_center_integer;
+      state->fs_redeclares_gl_fragcoord =
+         state->fs_origin_upper_left ||
+         state->fs_pixel_center_integer ||
+         state->fs_redeclares_gl_fragcoord_with_no_layout_qualifiers;
+   }
+
    if (qual->flags.q.explicit_location) {
       validate_explicit_location(qual, var, state, loc);
    } else if (qual->flags.q.explicit_index) {
@@ -2908,16 +2988,10 @@ ast_declarator_list::hir(exec_list *instructions,
 	    _mesa_glsl_error(& loc, state,
 			     "undeclared variable `%s' cannot be marked "
 			     "invariant", decl->identifier);
-	 } else if ((state->stage == MESA_SHADER_VERTEX)
-	       && (earlier->data.mode != ir_var_shader_out)) {
-	    _mesa_glsl_error(& loc, state,
-			     "`%s' cannot be marked invariant, vertex shader "
-			     "outputs only", decl->identifier);
-	 } else if ((state->stage == MESA_SHADER_FRAGMENT)
-	       && (earlier->data.mode != ir_var_shader_in)) {
-	    _mesa_glsl_error(& loc, state,
-			     "`%s' cannot be marked invariant, fragment shader "
-			     "inputs only", decl->identifier);
+	 } else if (!is_varying_var(earlier, state->stage)) {
+	    _mesa_glsl_error(&loc, state,
+			     "`%s' cannot be marked invariant; interfaces between "
+			     "shader stages only.", decl->identifier);
 	 } else if (earlier->data.used) {
 	    _mesa_glsl_error(& loc, state,
 			     "variable `%s' may not be redeclared "
@@ -3092,20 +3166,11 @@ ast_declarator_list::hir(exec_list *instructions,
 				       & loc, false);
 
       if (this->type->qualifier.flags.q.invariant) {
-	 if ((state->stage == MESA_SHADER_VERTEX) &&
-             var->data.mode != ir_var_shader_out) {
-	    _mesa_glsl_error(& loc, state,
-			     "`%s' cannot be marked invariant, vertex shader "
-			     "outputs only", var->name);
-	 } else if ((state->stage == MESA_SHADER_FRAGMENT) &&
-		    var->data.mode != ir_var_shader_in) {
-	    /* FINISHME: Note that this doesn't work for invariant on
-	     * a function signature inval
-	     */
-	    _mesa_glsl_error(& loc, state,
-			     "`%s' cannot be marked invariant, fragment shader "
-			     "inputs only", var->name);
-	 }
+         if (!is_varying_var(var, state->stage)) {
+            _mesa_glsl_error(&loc, state,
+                             "`%s' cannot be marked invariant; interfaces between "
+                             "shader stages only", var->name);
+         }
       }
 
       if (state->current_function != NULL) {
@@ -3429,11 +3494,15 @@ ast_declarator_list::hir(exec_list *instructions,
        * instruction stream.
        */
       exec_list initializer_instructions;
+
+      /* Examine var name here since var may get deleted in the next call */
+      bool var_is_gl_id = (strncmp(var->name, "gl_", 3) == 0);
+
       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 &&
+         if (var_is_gl_id &&
              earlier->data.how_declared == ir_var_declared_in_block) {
             _mesa_glsl_error(&loc, state,
                              "`%s' has already been redeclared using "
@@ -5143,6 +5212,13 @@ ast_interface_block::hir(exec_list *instructions,
          earlier->reinit_interface_type(block_type);
          delete var;
       } else {
+         /* Propagate the "binding" keyword into this UBO's fields;
+          * the UBO declaration itself doesn't get an ir_variable unless it
+          * has an instance name.  This is ugly.
+          */
+         var->data.explicit_binding = this->layout.flags.q.explicit_binding;
+         var->data.binding = this->layout.binding;
+
          state->symbols->add_variable(var);
          instructions->push_tail(var);
       }

src/glsl/glsl_parser.yy

@@ -1242,6 +1242,13 @@ layout_qualifier_id:
       if (match_layout_qualifier("location", $1, state) == 0) {
          $$.flags.q.explicit_location = 1;
 
+         if ($$.flags.q.attribute == 1 &&
+             state->ARB_explicit_attrib_location_warn) {
+            _mesa_glsl_warning(& @1, state,
+                               "GL_ARB_explicit_attrib_location layout "
+                               "identifier `%s' used", $1);
+         }
+
          if ($3 >= 0) {
             $$.location = $3;
          } else {
@@ -1298,10 +1305,6 @@ layout_qualifier_id:
          _mesa_glsl_error(& @1, state, "unrecognized layout identifier "
                           "`%s'", $1);
          YYERROR;
-      } else if (state->ARB_explicit_attrib_location_warn) {
-         _mesa_glsl_warning(& @1, state,
-                            "GL_ARB_explicit_attrib_location layout "
-                            "identifier `%s' used", $1);
       }
    }
    | interface_block_layout_qualifier

src/glsl/glsl_parser_extras.cpp

@@ -186,6 +186,12 @@ _mesa_glsl_parse_state::_mesa_glsl_parse_state(struct gl_context *_ctx,
    this->default_uniform_qualifier->flags.q.shared = 1;
    this->default_uniform_qualifier->flags.q.column_major = 1;
 
+   this->fs_uses_gl_fragcoord = false;
+   this->fs_redeclares_gl_fragcoord = false;
+   this->fs_origin_upper_left = false;
+   this->fs_pixel_center_integer = false;
+   this->fs_redeclares_gl_fragcoord_with_no_layout_qualifiers = false;
+
    this->gs_input_prim_type_specified = false;
    this->gs_input_prim_type = GL_POINTS;
    this->gs_input_size = 0;
@@ -1332,23 +1338,47 @@ set_shader_inout_layout(struct gl_shader *shader,
       /* Should have been prevented by the parser. */
       assert(!state->gs_input_prim_type_specified);
       assert(!state->out_qualifier->flags.i);
-      return;
    }
 
-   shader->Geom.VerticesOut = 0;
-   if (state->out_qualifier->flags.q.max_vertices)
-      shader->Geom.VerticesOut = state->out_qualifier->max_vertices;
-
-   if (state->gs_input_prim_type_specified) {
-      shader->Geom.InputType = state->gs_input_prim_type;
-   } else {
-      shader->Geom.InputType = PRIM_UNKNOWN;
+   if (shader->Stage != MESA_SHADER_FRAGMENT) {
+      /* Should have been prevented by the parser. */
+      assert(!state->fs_uses_gl_fragcoord);
+      assert(!state->fs_redeclares_gl_fragcoord);
+      assert(!state->fs_pixel_center_integer);
+      assert(!state->fs_origin_upper_left);
    }
 
-   if (state->out_qualifier->flags.q.prim_type) {
-      shader->Geom.OutputType = state->out_qualifier->prim_type;
-   } else {
-      shader->Geom.OutputType = PRIM_UNKNOWN;
+   switch(shader->Stage) {
+   case MESA_SHADER_GEOMETRY:
+      shader->Geom.VerticesOut = 0;
+      if (state->out_qualifier->flags.q.max_vertices)
+         shader->Geom.VerticesOut = state->out_qualifier->max_vertices;
+
+      if (state->gs_input_prim_type_specified) {
+         shader->Geom.InputType = state->gs_input_prim_type;
+      } else {
+         shader->Geom.InputType = PRIM_UNKNOWN;
+      }
+
+      if (state->out_qualifier->flags.q.prim_type) {
+         shader->Geom.OutputType = state->out_qualifier->prim_type;
+      } else {
+         shader->Geom.OutputType = PRIM_UNKNOWN;
+      }
+      break;
+
+   case MESA_SHADER_FRAGMENT:
+      shader->redeclares_gl_fragcoord = state->fs_redeclares_gl_fragcoord;
+      shader->uses_gl_fragcoord = state->fs_uses_gl_fragcoord;
+      shader->pixel_center_integer = state->fs_pixel_center_integer;
+      shader->origin_upper_left = state->fs_origin_upper_left;
+      shader->ARB_fragment_coord_conventions_enable =
+         state->ARB_fragment_coord_conventions_enable;
+      break;
+
+    default:
+      /* Nothing to do. */
+      break;
    }
 }
 

src/glsl/glsl_parser_extras.h

@@ -181,6 +181,18 @@ struct _mesa_glsl_parse_state {
     */
    struct ast_type_qualifier *default_uniform_qualifier;
 
+   /**
+    * Variables to track different cases if a fragment shader redeclares
+    * built-in variable gl_FragCoord.
+    *
+    * Note: These values are computed at ast_to_hir time rather than at parse
+    * time.
+    */
+   bool fs_redeclares_gl_fragcoord;
+   bool fs_origin_upper_left;
+   bool fs_pixel_center_integer;
+   bool fs_redeclares_gl_fragcoord_with_no_layout_qualifiers;
+
    /**
     * True if a geometry shader input primitive type was specified using a
     * layout directive.
@@ -362,6 +374,7 @@ struct _mesa_glsl_parse_state {
    const struct gl_extensions *extensions;
 
    bool uses_builtin_functions;
+   bool fs_uses_gl_fragcoord;
 
    /**
     * For geometry shaders, size of the most recently seen input declaration

src/glsl/link_uniform_block_active_visitor.cpp

@@ -46,6 +46,14 @@ process_block(void *mem_ctx, struct hash_table *ht, ir_variable *var)
       b->type = block_type;
       b->has_instance_name = var->is_interface_instance();
 
+      if (var->data.explicit_binding) {
+         b->has_binding = true;
+         b->binding = var->data.binding;
+      } else {
+         b->has_binding = false;
+         b->binding = 0;
+      }
+
       _mesa_hash_table_insert(ht, h, var->get_interface_type()->name,
 			      (void *) b);
       return b;

src/glsl/link_uniform_block_active_visitor.h

@@ -36,7 +36,10 @@ struct link_uniform_block_active {
    unsigned *array_elements;
    unsigned num_array_elements;
 
+   unsigned binding;
+
    bool has_instance_name;
+   bool has_binding;
 };
 
 class link_uniform_block_active_visitor : public ir_hierarchical_visitor {

src/glsl/link_uniform_blocks.cpp

@@ -251,7 +251,17 @@ link_uniform_blocks(void *mem_ctx,
             blocks[i].Name = ralloc_asprintf(blocks, "%s[%u]", name,
                                              b->array_elements[j]);
             blocks[i].Uniforms = &variables[parcel.index];
-            blocks[i].Binding = 0;
+
+            /* The GL_ARB_shading_language_420pack spec says:
+             *
+             *     "If the binding identifier is used with a uniform block
+             *     instanced as an array then the first element of the array
+             *     takes the specified block binding and each subsequent
+             *     element takes the next consecutive uniform block binding
+             *     point."
+             */
+            blocks[i].Binding = (b->has_binding) ? b->binding + j : 0;
+
             blocks[i].UniformBufferSize = 0;
             blocks[i]._Packing =
                gl_uniform_block_packing(block_type->interface_packing);
@@ -269,7 +279,7 @@ link_uniform_blocks(void *mem_ctx,
       } else {
          blocks[i].Name = ralloc_strdup(blocks, block_type->name);
          blocks[i].Uniforms = &variables[parcel.index];
-         blocks[i].Binding = 0;
+         blocks[i].Binding = (b->has_binding) ? b->binding : 0;
          blocks[i].UniformBufferSize = 0;
          blocks[i]._Packing =
             gl_uniform_block_packing(block_type->interface_packing);

src/glsl/link_uniform_initializers.cpp

@@ -46,6 +46,18 @@ get_storage(gl_uniform_storage *storage, unsigned num_storage,
    return NULL;
 }
 
+static unsigned
+get_uniform_block_index(const gl_shader_program *shProg,
+                        const char *uniformBlockName)
+{
+   for (unsigned i = 0; i < shProg->NumUniformBlocks; i++) {
+      if (!strcmp(shProg->UniformBlocks[i].Name, uniformBlockName))
+	 return i;
+   }
+
+   return GL_INVALID_INDEX;
+}
+
 void
 copy_constant_to_storage(union gl_constant_value *storage,
 			 const ir_constant *val,
@@ -83,8 +95,7 @@ copy_constant_to_storage(union gl_constant_value *storage,
 }
 
 void
-set_uniform_binding(void *mem_ctx, gl_shader_program *prog,
-                    const char *name, const glsl_type *type, int binding)
+set_sampler_binding(gl_shader_program *prog, const char *name, int binding)
 {
    struct gl_uniform_storage *const storage =
       get_storage(prog->UniformStorage, prog->NumUserUniformStorage, name);
@@ -94,42 +105,53 @@ set_uniform_binding(void *mem_ctx, gl_shader_program *prog,
       return;
    }
 
-   if (storage->type->is_sampler()) {
-      unsigned elements = MAX2(storage->array_elements, 1);
+   const unsigned elements = MAX2(storage->array_elements, 1);
 
-      /* From section 4.4.4 of the GLSL 4.20 specification:
-       * "If the binding identifier is used with an array, the first element
-       *  of the array takes the specified unit and each subsequent element
-       *  takes the next consecutive unit."
-       */
-      for (unsigned int i = 0; i < elements; i++) {
-         storage->storage[i].i = binding + i;
-      }
+   /* Section 4.4.4 (Opaque-Uniform Layout Qualifiers) of the GLSL 4.20 spec
+    * says:
+    *
+    *     "If the binding identifier is used with an array, the first element
+    *     of the array takes the specified unit and each subsequent element
+    *     takes the next consecutive unit."
+    */
+   for (unsigned int i = 0; i < elements; i++) {
+      storage->storage[i].i = binding + i;
+   }
 
-      for (int sh = 0; sh < MESA_SHADER_STAGES; sh++) {
-         gl_shader *shader = prog->_LinkedShaders[sh];
+   for (int sh = 0; sh < MESA_SHADER_STAGES; sh++) {
+      gl_shader *shader = prog->_LinkedShaders[sh];
 
-         if (shader && storage->sampler[sh].active) {
-            for (unsigned i = 0; i < elements; i++) {
-               unsigned index = storage->sampler[sh].index + i;
+      if (shader && storage->sampler[sh].active) {
+         for (unsigned i = 0; i < elements; i++) {
+            unsigned index = storage->sampler[sh].index + i;
 
-               shader->SamplerUnits[index] = storage->storage[i].i;
-            }
+            shader->SamplerUnits[index] = storage->storage[i].i;
          }
       }
-   } else if (storage->block_index != -1) {
+   }
+
+   storage->initialized = true;
+}
+
+void
+set_block_binding(gl_shader_program *prog, const char *block_name, int binding)
+{
+   const unsigned block_index = get_uniform_block_index(prog, block_name);
+
+   if (block_index == GL_INVALID_INDEX) {
+      assert(block_index != GL_INVALID_INDEX);
+      return;
+   }
+
       /* This is a field of a UBO.  val is the binding index. */
       for (int i = 0; i < MESA_SHADER_STAGES; i++) {
-         int stage_index = prog->UniformBlockStageIndex[i][storage->block_index];
+         int stage_index = prog->UniformBlockStageIndex[i][block_index];
 
          if (stage_index != -1) {
             struct gl_shader *sh = prog->_LinkedShaders[i];
             sh->UniformBlocks[stage_index].Binding = binding;
          }
       }
-   }
-
-   storage->initialized = true;
 }
 
 void
@@ -231,8 +253,53 @@ link_set_uniform_initializers(struct gl_shader_program *prog)
 	    mem_ctx = ralloc_context(NULL);
 
          if (var->data.explicit_binding) {
-            linker::set_uniform_binding(mem_ctx, prog, var->name,
-                                        var->type, var->data.binding);
+            const glsl_type *const type = var->type;
+
+            if (type->is_sampler()
+                || (type->is_array() && type->fields.array->is_sampler())) {
+               linker::set_sampler_binding(prog, var->name, var->data.binding);
+            } else if (var->is_in_uniform_block()) {
+               const glsl_type *const iface_type = var->get_interface_type();
+
+               /* If the variable is an array and it is an interface instance,
+                * we need to set the binding for each array element.  Just
+                * checking that the variable is an array is not sufficient.
+                * The variable could be an array element of a uniform block
+                * that lacks an instance name.  For example:
+                *
+                *     uniform U {
+                *         float f[4];
+                *     };
+                *
+                * In this case "f" would pass is_in_uniform_block (above) and
+                * type->is_array(), but it will fail is_interface_instance().
+                */
+               if (var->is_interface_instance() && var->type->is_array()) {
+                  for (unsigned i = 0; i < var->type->length; i++) {
+                     const char *name =
+                        ralloc_asprintf(mem_ctx, "%s[%u]", iface_type->name, i);
+
+                     /* Section 4.4.3 (Uniform Block Layout Qualifiers) of the
+                      * GLSL 4.20 spec says:
+                      *
+                      *     "If the binding identifier is used with a uniform
+                      *     block instanced as an array then the first element
+                      *     of the array takes the specified block binding and
+                      *     each subsequent element takes the next consecutive
+                      *     uniform block binding point."
+                      */
+                     linker::set_block_binding(prog, name,
+                                               var->data.binding + i);
+                  }
+               } else {
+                  linker::set_block_binding(prog, iface_type->name,
+                                            var->data.binding);
+               }
+            } else if (type->contains_atomic()) {
+               /* we don't actually need to do anything. */
+            } else {
+               assert(!"Explicit binding not on a sampler, UBO or atomic.");
+            }
          } else if (var->constant_value) {
             linker::set_uniform_initializer(mem_ctx, prog, var->name,
                                             var->type, var->constant_value);

src/glsl/linker.cpp

@@ -1194,6 +1194,83 @@ private:
    hash_table *unnamed_interfaces;
 };
 
+/**
+ * Performs the cross-validation of layout qualifiers specified in
+ * redeclaration of gl_FragCoord for the attached fragment shaders,
+ * and propagates them to the linked FS and linked shader program.
+ */
+static void
+link_fs_input_layout_qualifiers(struct gl_shader_program *prog,
+	                        struct gl_shader *linked_shader,
+	                        struct gl_shader **shader_list,
+	                        unsigned num_shaders)
+{
+   linked_shader->redeclares_gl_fragcoord = false;
+   linked_shader->uses_gl_fragcoord = false;
+   linked_shader->origin_upper_left = false;
+   linked_shader->pixel_center_integer = false;
+
+   if (linked_shader->Stage != MESA_SHADER_FRAGMENT ||
+       (prog->Version < 150 && !prog->ARB_fragment_coord_conventions_enable))
+      return;
+
+   for (unsigned i = 0; i < num_shaders; i++) {
+      struct gl_shader *shader = shader_list[i];
+      /* From the GLSL 1.50 spec, page 39:
+       *
+       *   "If gl_FragCoord is redeclared in any fragment shader in a program,
+       *    it must be redeclared in all the fragment shaders in that program
+       *    that have a static use gl_FragCoord."
+       *
+       * Exclude the case when one of the 'linked_shader' or 'shader' redeclares
+       * gl_FragCoord with no layout qualifiers but the other one doesn't
+       * redeclare it. If we strictly follow GLSL 1.50 spec's language, it
+       * should be a link error. But, generating link error for this case will
+       * be a wrong behaviour which spec didn't intend to do and it could also
+       * break some applications.
+       */
+      if ((linked_shader->redeclares_gl_fragcoord
+           && !shader->redeclares_gl_fragcoord
+           && shader->uses_gl_fragcoord
+           && (linked_shader->origin_upper_left
+               || linked_shader->pixel_center_integer))
+          || (shader->redeclares_gl_fragcoord
+              && !linked_shader->redeclares_gl_fragcoord
+              && linked_shader->uses_gl_fragcoord
+              && (shader->origin_upper_left
+                  || shader->pixel_center_integer))) {
+             linker_error(prog, "fragment shader defined with conflicting "
+                         "layout qualifiers for gl_FragCoord\n");
+      }
+
+      /* From the GLSL 1.50 spec, page 39:
+       *
+       *   "All redeclarations of gl_FragCoord in all fragment shaders in a
+       *    single program must have the same set of qualifiers."
+       */
+      if (linked_shader->redeclares_gl_fragcoord && shader->redeclares_gl_fragcoord
+          && (shader->origin_upper_left != linked_shader->origin_upper_left
+          || shader->pixel_center_integer != linked_shader->pixel_center_integer)) {
+         linker_error(prog, "fragment shader defined with conflicting "
+                      "layout qualifiers for gl_FragCoord\n");
+      }
+
+      /* Update the linked shader state.  Note that uses_gl_fragcoord should
+       * accumulate the results.  The other values should replace.  If there
+       * are multiple redeclarations, all the fields except uses_gl_fragcoord
+       * are already known to be the same.
+       */
+      if (shader->redeclares_gl_fragcoord || shader->uses_gl_fragcoord) {
+         linked_shader->redeclares_gl_fragcoord =
+            shader->redeclares_gl_fragcoord;
+         linked_shader->uses_gl_fragcoord = linked_shader->uses_gl_fragcoord
+            || shader->uses_gl_fragcoord;
+         linked_shader->origin_upper_left = shader->origin_upper_left;
+         linked_shader->pixel_center_integer = shader->pixel_center_integer;
+      }
+   }
+}
+
 /**
  * Performs the cross-validation of geometry shader max_vertices and
  * primitive type layout qualifiers for the attached geometry shaders,
@@ -1390,6 +1467,7 @@ link_intrastage_shaders(void *mem_ctx,
    linked->NumUniformBlocks = num_uniform_blocks;
    ralloc_steal(linked, linked->UniformBlocks);
 
+   link_fs_input_layout_qualifiers(prog, linked, shader_list, num_shaders);
    link_gs_inout_layout_qualifiers(prog, linked, shader_list, num_shaders);
 
    populate_symbol_table(linked);
@@ -1716,10 +1794,12 @@ assign_attribute_or_color_locations(gl_shader_program *prog,
 	     *     active attribute array, both of which require multiple
 	     *     contiguous generic attributes."
 	     *
-	     * Previous versions of the spec contain similar language but omit
-	     * the bit about attribute arrays.
+	     * I think above text prohibits the aliasing of explicit and
+	     * automatic assignments. But, aliasing is allowed in manual
+	     * assignments of attribute locations. See below comments for
+	     * the details.
 	     *
-	     * Page 61 of the OpenGL 4.0 spec also says:
+	     * From OpenGL 4.0 spec, page 61:
 	     *
 	     *     "It is possible for an application to bind more than one
 	     *     attribute name to the same location. This is referred to as
@@ -1732,29 +1812,84 @@ assign_attribute_or_color_locations(gl_shader_program *prog,
 	     *     but implementations are not required to generate an error
 	     *     in this case."
 	     *
-	     * These two paragraphs are either somewhat contradictory, or I
-	     * don't fully understand one or both of them.
-	     */
-	    /* FINISHME: The code as currently written does not support
-	     * FINISHME: attribute location aliasing (see comment above).
+	     * From GLSL 4.30 spec, page 54:
+	     *
+	     *    "A program will fail to link if any two non-vertex shader
+	     *     input variables are assigned to the same location. For
+	     *     vertex shaders, multiple input variables may be assigned
+	     *     to the same location using either layout qualifiers or via
+	     *     the OpenGL API. However, such aliasing is intended only to
+	     *     support vertex shaders where each execution path accesses
+	     *     at most one input per each location. Implementations are
+	     *     permitted, but not required, to generate link-time errors
+	     *     if they detect that every path through the vertex shader
+	     *     executable accesses multiple inputs assigned to any single
+	     *     location. For all shader types, a program will fail to link
+	     *     if explicit location assignments leave the linker unable
+	     *     to find space for other variables without explicit
+	     *     assignments."
+	     *
+	     * From OpenGL ES 3.0 spec, page 56:
+	     *
+	     *    "Binding more than one attribute name to the same location
+	     *     is referred to as aliasing, and is not permitted in OpenGL
+	     *     ES Shading Language 3.00 vertex shaders. LinkProgram will
+	     *     fail when this condition exists. However, aliasing is
+	     *     possible in OpenGL ES Shading Language 1.00 vertex shaders.
+	     *     This will only work if only one of the aliased attributes
+	     *     is active in the executable program, or if no path through
+	     *     the shader consumes more than one attribute of a set of
+	     *     attributes aliased to the same location. A link error can
+	     *     occur if the linker determines that every path through the
+	     *     shader consumes multiple aliased attributes, but implemen-
+	     *     tations are not required to generate an error in this case."
+	     *
+	     * After looking at above references from OpenGL, OpenGL ES and
+	     * GLSL specifications, we allow aliasing of vertex input variables
+	     * in: OpenGL 2.0 (and above) and OpenGL ES 2.0.
+	     *
+	     * NOTE: This is not required by the spec but its worth mentioning
+	     * here that we're not doing anything to make sure that no path
+	     * through the vertex shader executable accesses multiple inputs
+	     * assigned to any single location.
 	     */
+
 	    /* Mask representing the contiguous slots that will be used by
 	     * this attribute.
 	     */
 	    const unsigned attr = var->data.location - generic_base;
 	    const unsigned use_mask = (1 << slots) - 1;
+            const char *const string = (target_index == MESA_SHADER_VERTEX)
+               ? "vertex shader input" : "fragment shader output";
+
+            /* Generate a link error if the requested locations for this
+             * attribute exceed the maximum allowed attribute location.
+             */
+            if (attr + slots > max_index) {
+               linker_error(prog,
+                           "insufficient contiguous locations "
+                           "available for %s `%s' %d %d %d", string,
+                           var->name, used_locations, use_mask, attr);
+               return false;
+            }
 
 	    /* Generate a link error if the set of bits requested for this
 	     * attribute overlaps any previously allocated bits.
 	     */
 	    if ((~(use_mask << attr) & used_locations) != used_locations) {
-	       const char *const string = (target_index == MESA_SHADER_VERTEX)
-		  ? "vertex shader input" : "fragment shader output";
-	       linker_error(prog,
-			    "insufficient contiguous locations "
-			    "available for %s `%s' %d %d %d", string,
-			    var->name, used_locations, use_mask, attr);
-	       return false;
+               if (target_index == MESA_SHADER_FRAGMENT ||
+                   (prog->IsES && prog->Version >= 300)) {
+                  linker_error(prog,
+                               "overlapping location is assigned "
+                               "to %s `%s' %d %d %d\n", string,
+                               var->name, used_locations, use_mask, attr);
+                  return false;
+               } else {
+                  linker_warning(prog,
+                                 "overlapping location is assigned "
+                                 "to %s `%s' %d %d %d\n", string,
+                                 var->name, used_locations, use_mask, attr);
+               }
 	    }
 
 	    used_locations |= (use_mask << attr);
@@ -1993,6 +2128,7 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
    ralloc_free(prog->AtomicBuffers);
    prog->AtomicBuffers = NULL;
    prog->NumAtomicBuffers = 0;
+   prog->ARB_fragment_coord_conventions_enable = false;
 
    /* Separate the shaders into groups based on their type.
     */
@@ -2019,6 +2155,9 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
 	 goto done;
       }
 
+      prog->ARB_fragment_coord_conventions_enable |=
+         prog->Shaders[i]->ARB_fragment_coord_conventions_enable;
+
       gl_shader_stage shader_type = prog->Shaders[i]->Stage;
       shader_list[shader_type][num_shaders[shader_type]] = prog->Shaders[i];
       num_shaders[shader_type]++;

src/glsl/lower_ubo_reference.cpp

@@ -194,12 +194,16 @@ lower_ubo_reference_visitor::handle_rvalue(ir_rvalue **rvalue)
 	    array_stride = glsl_align(array_stride, 16);
 	 }
 
-	 ir_constant *const_index = deref_array->array_index->as_constant();
+         ir_rvalue *array_index = deref_array->array_index;
+         if (array_index->type->base_type == GLSL_TYPE_INT)
+            array_index = i2u(array_index);
+
+	 ir_constant *const_index = array_index->as_constant();
 	 if (const_index) {
-	    const_offset += array_stride * const_index->value.i[0];
+	    const_offset += array_stride * const_index->value.u[0];
 	 } else {
 	    offset = add(offset,
-			 mul(deref_array->array_index,
+			 mul(array_index,
 			     new(mem_ctx) ir_constant(array_stride)));
 	 }
 	 deref = deref_array->array->as_dereference();

src/glsl/opt_vectorize.cpp

@@ -260,6 +260,7 @@ ir_vectorize_visitor::visit_enter(ir_assignment *ir)
    if (ir->condition ||
        this->channels >= 4 ||
        !single_channel_write_mask(ir->write_mask) ||
+       this->assignment[write_mask_to_swizzle(ir->write_mask)] != NULL ||
        (lhs && !ir->lhs->equals(lhs)) ||
        (rhs && !ir->rhs->equals(rhs, ir_type_swizzle))) {
       try_vectorize();

src/glx/apple/apple_visual.c

@@ -73,11 +73,15 @@ apple_visual_create_pfobj(CGLPixelFormatObj * pfobj, const struct glx_config * m
    GLint vsref = 0;
    CGLError error = 0;
 
-   /* Request an OpenGL 3.2 profile if one is available */
-   if(apple_cgl.version_major > 1 || (apple_cgl.version_major == 1 && apple_cgl.version_minor >= 3)) {
-      attr[numattr++] = kCGLPFAOpenGLProfile;
-      attr[numattr++] = kCGLOGLPVersion_3_2_Core;
-   }
+   /* Request an OpenGL 3.2 profile if one is available and supported */
+   attr[numattr++] = kCGLPFAOpenGLProfile;
+   attr[numattr++] = kCGLOGLPVersion_3_2_Core;
+
+   /* Test for kCGLPFAOpenGLProfile support at runtime and roll it out if not supported */
+   attr[numattr] = 0;
+   error = apple_cgl.choose_pixel_format(attr, pfobj, &vsref);
+   if (error == kCGLBadAttribute)
+      numattr -= 2;
 
    if (offscreen) {
       apple_glx_diagnostic

src/glx/dri2.c

@@ -132,10 +132,14 @@ DRI2WireToEvent(Display *dpy, XEvent *event, xEvent *wire)
       aevent->msc = ((CARD64)awire->msc_hi << 32) | awire->msc_lo;
 
       glxDraw = GetGLXDrawable(dpy, pdraw->drawable);
-      if (awire->sbc < glxDraw->lastEventSbc)
-	 glxDraw->eventSbcWrap += 0x100000000;
-      glxDraw->lastEventSbc = awire->sbc;
-      aevent->sbc = awire->sbc + glxDraw->eventSbcWrap;
+      if (glxDraw != NULL) {
+         if (awire->sbc < glxDraw->lastEventSbc)
+            glxDraw->eventSbcWrap += 0x100000000;
+         glxDraw->lastEventSbc = awire->sbc;
+         aevent->sbc = awire->sbc + glxDraw->eventSbcWrap;
+      } else {
+         aevent->sbc = awire->sbc;
+      }
 
       return True;
    }

src/glx/dri2.h

@@ -95,4 +95,12 @@ _X_HIDDEN int
 dri2_query_renderer_string(struct glx_screen *base, int attribute,
                            const char **value);
 
+_X_HIDDEN int
+dri3_query_renderer_integer(struct glx_screen *base, int attribute,
+                            unsigned int *value);
+
+_X_HIDDEN int
+dri3_query_renderer_string(struct glx_screen *base, int attribute,
+                           const char **value);
+
 #endif

src/glx/dri2_glx.c

@@ -1201,6 +1201,16 @@ dri2CreateScreen(int screen, struct glx_display * priv)
       goto handle_error;
    }
 
+   if (drmGetMagic(psc->fd, &magic)) {
+      ErrorMessageF("failed to get magic\n");
+      goto handle_error;
+   }
+
+   if (!DRI2Authenticate(priv->dpy, RootWindow(priv->dpy, screen), magic)) {
+      ErrorMessageF("failed to authenticate magic %d\n", magic);
+      goto handle_error;
+   }
+
    /* If Mesa knows about the appropriate driver for this fd, then trust it.
     * Otherwise, default to the server's value.
     */
@@ -1232,16 +1242,6 @@ dri2CreateScreen(int screen, struct glx_display * priv)
       goto handle_error;
    }
 
-   if (drmGetMagic(psc->fd, &magic)) {
-      ErrorMessageF("failed to get magic\n");
-      goto handle_error;
-   }
-
-   if (!DRI2Authenticate(priv->dpy, RootWindow(priv->dpy, screen), magic)) {
-      ErrorMessageF("failed to authenticate magic %d\n", magic);
-      goto handle_error;
-   }
-
    if (psc->dri2->base.version >= 4) {
       psc->driScreen =
          psc->dri2->createNewScreen2(screen, psc->fd,

src/glx/dri2_query_renderer.c

@@ -29,6 +29,9 @@
 #include "dri2.h"
 #include "dri_interface.h"
 #include "dri2_priv.h"
+#if defined(HAVE_DRI3)
+#include "dri3_priv.h"
+#endif
 
 static int
 dri2_convert_glx_query_renderer_attribs(int attribute)
@@ -99,4 +102,44 @@ dri2_query_renderer_string(struct glx_screen *base, int attribute,
    return psc->rendererQuery->queryString(psc->driScreen, dri_attribute, value);
 }
 
+#if defined(HAVE_DRI3)
+_X_HIDDEN int
+dri3_query_renderer_integer(struct glx_screen *base, int attribute,
+                            unsigned int *value)
+{
+   struct dri3_screen *const psc = (struct dri3_screen *) base;
+
+   /* Even though there are invalid values (and
+    * dri2_convert_glx_query_renderer_attribs may return -1), the higher level
+    * GLX code is required to perform the filtering.  Assume that we got a
+    * good value.
+    */
+   const int dri_attribute = dri2_convert_glx_query_renderer_attribs(attribute);
+
+   if (psc->rendererQuery == NULL)
+      return -1;
+
+   return psc->rendererQuery->queryInteger(psc->driScreen, dri_attribute,
+                                           value);
+}
+
+_X_HIDDEN int
+dri3_query_renderer_string(struct glx_screen *base, int attribute,
+                           const char **value)
+{
+   struct dri3_screen *const psc = (struct dri3_screen *) base;
+
+   /* Even though queryString only accepts a subset of the possible GLX
+    * queries, the higher level GLX code is required to perform the filtering.
+    * Assume that we got a good value.
+    */
+   const int dri_attribute = dri2_convert_glx_query_renderer_attribs(attribute);
+
+   if (psc->rendererQuery == NULL)
+      return -1;
+
+   return psc->rendererQuery->queryString(psc->driScreen, dri_attribute, value);
+}
+#endif /* HAVE_DRI3 */
+
 #endif /* GLX_DIRECT_RENDERING */

src/glx/dri3_glx.c

@@ -79,6 +79,7 @@
 #include "dri_common.h"
 #include "dri3_priv.h"
 #include "loader.h"
+#include "dri2.h"
 
 static const struct glx_context_vtable dri3_context_vtable;
 
@@ -1580,12 +1581,19 @@ dri3_bind_extensions(struct dri3_screen *psc, struct glx_display * priv,
       if (strcmp(extensions[i]->name, __DRI2_ROBUSTNESS) == 0)
          __glXEnableDirectExtension(&psc->base,
                                     "GLX_ARB_create_context_robustness");
+
+      if (strcmp(extensions[i]->name, __DRI2_RENDERER_QUERY) == 0) {
+         psc->rendererQuery = (__DRI2rendererQueryExtension *) extensions[i];
+         __glXEnableDirectExtension(&psc->base, "GLX_MESA_query_renderer");
+      }
    }
 }
 
 static const struct glx_screen_vtable dri3_screen_vtable = {
    dri3_create_context,
-   dri3_create_context_attribs
+   dri3_create_context_attribs,
+   dri3_query_renderer_integer,
+   dri3_query_renderer_string,
 };
 
 /** dri3_create_screen

src/glx/dri3_priv.h

@@ -129,6 +129,7 @@ struct dri3_screen {
    const __DRI2flushExtension *f;
    const __DRI2configQueryExtension *config;
    const __DRItexBufferExtension *texBuffer;
+   const __DRI2rendererQueryExtension *rendererQuery;
    const __DRIconfig **driver_configs;
 
    void *driver;

src/glx/glxext.c

@@ -249,6 +249,10 @@ glx_display_free(struct glx_display *priv)
    if (priv->dri2Display)
       (*priv->dri2Display->destroyDisplay) (priv->dri2Display);
    priv->dri2Display = NULL;
+
+   if (priv->dri3Display)
+      (*priv->dri3Display->destroyDisplay) (priv->dri3Display);
+   priv->dri3Display = NULL;
 #endif
 
    free((char *) priv);
@@ -825,7 +829,6 @@ __glXInitialize(Display * dpy)
    dpyPriv->codes = XInitExtension(dpy, __glXExtensionName);
    if (!dpyPriv->codes) {
       free(dpyPriv);
-      _XUnlockMutex(_Xglobal_lock);
       return NULL;
    }
 
@@ -841,7 +844,6 @@ __glXInitialize(Display * dpy)
 		     &dpyPriv->majorVersion, &dpyPriv->minorVersion)
        || (dpyPriv->majorVersion == 1 && dpyPriv->minorVersion < 1)) {
       free(dpyPriv);
-      _XUnlockMutex(_Xglobal_lock);
       return NULL;
    }
 
@@ -906,7 +908,7 @@ __glXInitialize(Display * dpy)
    dpyPriv->next = glx_displays;
    glx_displays = dpyPriv;
 
-    _XUnlockMutex(_Xglobal_lock);
+   _XUnlockMutex(_Xglobal_lock);
 
    return dpyPriv;
 }

src/loader/Makefile.sources

@@ -1,2 +1,3 @@
 LOADER_C_FILES := \
-	loader.c
+	loader.c \
+	pci_id_driver_map.c

src/loader/loader.c

@@ -78,7 +78,7 @@
 #endif
 
 #define __IS_LOADER
-#include "pci_ids/pci_id_driver_map.h"
+#include "pci_id_driver_map.h"
 
 static void default_logger(int level, const char *fmt, ...)
 {
@@ -202,7 +202,7 @@ out:
    return (*chip_id >= 0);
 }
 
-#elif defined(ANDROID) && !defined(__NOT_HAVE_DRM_H)
+#elif !defined(__NOT_HAVE_DRM_H)
 
 /* for i915 */
 #include <i915_drm.h>
@@ -219,12 +219,12 @@ loader_get_pci_id_for_fd(int fd, int *vendor_id, int *chip_id)
    version = drmGetVersion(fd);
    if (!version) {
       log_(_LOADER_WARNING, "MESA-LOADER: invalid drm fd\n");
-      return FALSE;
+      return 0;
    }
    if (!version->name) {
       log_(_LOADER_WARNING, "MESA-LOADER: unable to determine the driver name\n");
       drmFreeVersion(version);
-      return FALSE;
+      return 0;
    }
 
    if (strcmp(version->name, "i915") == 0) {
@@ -352,6 +352,9 @@ loader_get_driver_for_fd(int fd, unsigned driver_types)
       if (!(driver_types & driver_map[i].driver_types))
          continue;
 
+      if (driver_map[i].predicate && !driver_map[i].predicate(fd))
+         continue;
+
       if (driver_map[i].num_chips_ids == -1) {
          driver = strdup(driver_map[i].driver);
          goto out;

src/loader/pci_id_driver_map.c

@@ -0,0 +1,55 @@
+/*
+ * Copyright 2014 Ilia Mirkin
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+int is_nouveau_vieux(int fd);
+
+#ifndef __NOT_HAVE_DRM_H
+
+#include <xf86drm.h>
+#include <nouveau_drm.h>
+
+static int
+nouveau_chipset(int fd)
+{
+   struct drm_nouveau_getparam gp = { NOUVEAU_GETPARAM_CHIPSET_ID, 0 };
+   int ret;
+
+   ret = drmCommandWriteRead(fd, DRM_NOUVEAU_GETPARAM, &gp, sizeof(gp));
+   if (ret)
+      return -1;
+
+   return gp.value;
+}
+
+int
+is_nouveau_vieux(int fd)
+{
+   int chipset = nouveau_chipset(fd);
+   return chipset > 0 && chipset < 0x30;
+}
+
+#else
+
+int is_nouveau_vieux(int fd) { return 0; }
+
+#endif

src/loader/pci_id_driver_map.h

@@ -59,12 +59,15 @@ static const int vmwgfx_chip_ids[] = {
 #undef CHIPSET
 };
 
+int is_nouveau_vieux(int fd);
+
 static const struct {
    int vendor_id;
    const char *driver;
    const int *chip_ids;
    int num_chips_ids;
    unsigned driver_types;
+   int (*predicate)(int fd);
 } driver_map[] = {
    { 0x8086, "i915", i915_chip_ids, ARRAY_SIZE(i915_chip_ids), _LOADER_DRI | _LOADER_GALLIUM },
    { 0x8086, "i965", i965_chip_ids, ARRAY_SIZE(i965_chip_ids), _LOADER_DRI | _LOADER_GALLIUM },
@@ -73,7 +76,8 @@ static const struct {
    { 0x1002, "r300", r300_chip_ids, ARRAY_SIZE(r300_chip_ids), _LOADER_GALLIUM },
    { 0x1002, "r600", r600_chip_ids, ARRAY_SIZE(r600_chip_ids), _LOADER_GALLIUM },
    { 0x1002, "radeonsi", radeonsi_chip_ids, ARRAY_SIZE(radeonsi_chip_ids), _LOADER_GALLIUM},
-   { 0x10de, "nouveau", NULL, -1,  _LOADER_GALLIUM  },
+   { 0x10de, "nouveau_vieux", NULL, -1,  _LOADER_DRI, is_nouveau_vieux },
+   { 0x10de, "nouveau", NULL, -1,  _LOADER_GALLIUM },
    { 0x15ad, "vmwgfx", vmwgfx_chip_ids, ARRAY_SIZE(vmwgfx_chip_ids), _LOADER_GALLIUM },
    { 0x0000, NULL, NULL, 0 },
 };