docs: Add sha256sums for the 10.5.9 release

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
Add release notes for the 10.5.9 release
2015-07-04 12:48:39 +01:00 · 2015-07-04 12:09:10 +01:00 · 2015-07-04 12:04:06 +01:00 · 2015-07-04 12:03:15 +01:00 · 2015-07-01 15:22:41 +01:00 · 2015-07-01 15:22:41 +01:00
147 changed files with 1857 additions and 423 deletions
--- a/Android.common.mk
+++ b/Android.common.mk
@@ -31,6 +31,7 @@ endif
 endif

 LOCAL_C_INCLUDES += \
+	$(MESA_TOP)/src \
 	$(MESA_TOP)/include

 MESA_VERSION=$(shell cat $(MESA_TOP)/VERSION)
@@ -41,6 +42,19 @@ LOCAL_CFLAGS += \
 	-DANDROID_VERSION=0x0$(MESA_ANDROID_MAJOR_VERSION)0$(MESA_ANDROID_MINOR_VERSION)

 LOCAL_CFLAGS += \
+	-DHAVE___BUILTIN_EXPECT \
+	-DHAVE___BUILTIN_FFS \
+	-DHAVE___BUILTIN_FFSLL \
+	-DHAVE_FUNC_ATTRIBUTE_FLATTEN \
+	-DHAVE_FUNC_ATTRIBUTE_UNUSED \
+	-DHAVE_FUNC_ATTRIBUTE_FORMAT \
+	-DHAVE_FUNC_ATTRIBUTE_PACKED \
+	-DHAVE___BUILTIN_CTZ \
+	-DHAVE___BUILTIN_POPCOUNT \
+	-DHAVE___BUILTIN_POPCOUNTLL \
+	-DHAVE___BUILTIN_CLZ \
+	-DHAVE___BUILTIN_CLZLL \
+	-DHAVE___BUILTIN_UNREACHABLE \
 	-DHAVE_PTHREAD=1 \
 	-fvisibility=hidden \
 	-Wno-sign-compare
--- a/2
+++ b/2
@@ -1 +1 @@
-10.5.3
+10.5.9
--- a/bin/.cherry-ignore
+++ b/bin/.cherry-ignore
@@ -24,3 +24,6 @@ fe5fddd7e2df74233a2a02ae021418485f39d11c

 # The optimisations mentioned are not available in 10.5
 627c68308683abbd6e563a09af6013a33938a790 i965/fs: in MAD optimizations, switch last argument to be immediate
+
+# 10.5 has the compat string implementation, which includes the <string>
+967825d053f71c5f5fc3ba31eabc0c6004fde4f1 clover: Build fix for FreeBSD.
--- a/bin/get-pick-list.sh
+++ b/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='^\([[:space:]]*NOTE: .*[Cc]andidate\|CC:.*10\.5.*mesa-stable\)' HEAD..origin/master |\
 while read sha
 do
 	# Check to see whether the patch is on the ignore list.
--- a/configure.ac
+++ b/configure.ac
@@ -609,6 +609,7 @@ if test "x$enable_asm" = xyes; then
 fi

 AC_CHECK_HEADER([xlocale.h], [DEFINES="$DEFINES -DHAVE_XLOCALE_H"])
+AC_CHECK_HEADER([sys/sysctl.h], [DEFINES="$DEFINES -DHAVE_SYS_SYSCTL_H"])
 AC_CHECK_FUNC([strtof], [DEFINES="$DEFINES -DHAVE_STRTOF"])

 dnl Check to see if dlopen is in default libraries (like Solaris, which
@@ -910,7 +911,7 @@ x*yes*yes*)
 esac

 # Building Xlib-GLX requires shared glapi to be disabled.
-if test "x$enable_xlib_glx" = xyes; then
+if test "x$enable_shared_glapi$enable_xlib_glx" = xyesyes; then
    AC_MSG_NOTICE([Shared GLAPI should not used with Xlib-GLX, disabling])
    enable_shared_glapi=no
 fi
@@ -1378,8 +1379,15 @@ if test "x$enable_egl" = xyes; then

    if test "$enable_static" != yes; then
        if test "x$enable_dri" = xyes; then
-	    HAVE_EGL_DRIVER_DRI2=1
-	fi
+            HAVE_EGL_DRIVER_DRI2=1
+            if test "x$enable_shared_glapi" = xno; then
+                AC_MSG_ERROR([egl_dri2 requires --enable-shared-glapi])
+            fi
+        else
+            # Avoid building an "empty" libEGL. Drop/update this
+            # when other backends (haiku?) come along.
+            AC_MSG_ERROR([egl requires --enable-dri])
+        fi

    fi
 fi
--- a/docs/relnotes/10.5.3.html
+++ b/docs/relnotes/10.5.3.html
@@ -31,7 +31,8 @@ because compatibility contexts are not supported.

 <h2>SHA256 checksums</h2>
 <pre>
-TBD
+2371b8e210ccd19f61dd94b6664d612e5a479ba7d431a074512d87633bd6aeb4  mesa-10.5.3.tar.gz
+8701ee1be4f5c03238f5e63c1a9bd4cc03a2f6c0155ed42a1ae7d58f18912ba2  mesa-10.5.3.tar.xz
 </pre>


--- a/docs/relnotes/10.5.4.html
+++ b/docs/relnotes/10.5.4.html
@@ -0,0 +1,125 @@
+<!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 10.5.4 Release Notes / April 24, 2015</h1>
+
+<p>
+Mesa 10.5.4 is a bug fix release which fixes bugs found since the 10.5.3 release.
+</p>
+<p>
+Mesa 10.5.4 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>SHA256 checksums</h2>
+<pre>
+e1089567fc7bf8d9b2d8badcc9f2fc3b758701c8c0ccfe7af1805549fea53f11  mesa-10.5.4.tar.gz
+b51e723f3a20d842c88a92d809435b229fc4744ca0dbec0317d9d4a3ac4c6803  mesa-10.5.4.tar.xz
+</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=69226">Bug 69226</a> - Cannot enable basic shaders with Second Life aborts attempt</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=71591">Bug 71591</a> - Second Life shaders fail to compile (extension declared in middle of shader)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=81025">Bug 81025</a> - [IVB/BYT Bisected]Piglit spec_ARB_draw_indirect_arb_draw_indirect-draw-elements-prim-restart-ugly fails</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=89457">Bug 89457</a> - [BSW Bisected]ogles3conform ES3-CTS.gtf.GL3Tests.shadow.shadow_execution_vert fails</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=89957">Bug 89957</a> - vm protection faults in piglit lest: texsubimage cube_map_array pbo</li>
+
+</ul>
+
+
+<h2>Changes</h2>
+
+<p>Brian Paul (1):</p>
+<ul>
+  <li>glsl: rewrite glsl_type::record_key_hash() to avoid buffer overflow</li>
+</ul>
+
+<p>Dave Airlie (2):</p>
+<ul>
+  <li>st/mesa: convert sub image for cube map arrays to 2d arrays for upload</li>
+  <li>st/mesa: align cube map arrays layers</li>
+</ul>
+
+<p>Emil Velikov (11):</p>
+<ul>
+  <li>docs: Add 256 sums for the 10.5.3 release</li>
+  <li>radeonsi: remove unused si_dump_key()</li>
+  <li>android: use LOCAL_SHARED_LIBRARIES over TARGET_OUT_HEADERS</li>
+  <li>android: add $(mesa_top)/src include to the whole of mesa</li>
+  <li>android: egl: add libsync_cflags to the build</li>
+  <li>android: dri/common: conditionally include drm_cflags/set __NOT_HAVE_DRM_H</li>
+  <li>android: add HAVE__BUILTIN_* and HAVE_FUNC_ATTRIBUTE_* defines</li>
+  <li>android: add $(mesa_top)/src/mesa/main to the includes list</li>
+  <li>android: dri: link against libmesa_util</li>
+  <li>android: mesa: fix the path of the SSE4_1 optimisations</li>
+  <li>Update version to 10.5.4</li>
+</ul>
+
+<p>Ian Romanick (1):</p>
+<ul>
+  <li>nir: Fix typo in "ushr by 0" algebraic replacement</li>
+</ul>
+
+<p>Kenneth Graunke (2):</p>
+<ul>
+  <li>i965: Fix software primitive restart with indirect draws.</li>
+  <li>drirc: Add "Second Life" quirk (allow_glsl_extension_directive_midshader).</li>
+</ul>
+
+<p>Kristian Høgsberg (1):</p>
+<ul>
+  <li>i965: Rewrite ir_tex to ir_txl with lod 0 for vertex shaders</li>
+</ul>
+
+<p>Marek Olšák (2):</p>
+<ul>
+  <li>glsl_to_tgsi: fix out-of-bounds constant access and crash for uniforms</li>
+  <li>glsl_to_tgsi: don't use a potentially-undefined immediate for ir_query_levels</li>
+</ul>
+
+<p>Mathias Froehlich (1):</p>
+<ul>
+  <li>i965: Flush batchbuffer containing the query on glQueryCounter.</li>
+</ul>
+
+<p>Mauro Rossi (2):</p>
+<ul>
+  <li>android: mesa: generate the format_{un,}pack.[ch] sources</li>
+  <li>android: add inital NIR build</li>
+</ul>
+
+
+</div>
+</body>
+</html>
--- a/docs/relnotes/10.5.5.html
+++ b/docs/relnotes/10.5.5.html
@@ -0,0 +1,95 @@
+<!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 10.5.5 Release Notes / May 11, 2015</h1>
+
+<p>
+Mesa 10.5.5 is a bug fix release which fixes bugs found since the 10.5.4 release.
+</p>
+<p>
+Mesa 10.5.5 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>
+c10f00fd792b8290dd51ebcc48a9016c4cafab19ec205423c6fcadfd7f3a59f2  mesa-10.5.5.tar.gz
+4ac4e4ea3414f1cadb1467f2f173f9e56170d31e8674f7953a46f0549d319f28  mesa-10.5.5.tar.xz
+</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=88521">Bug 88521</a> - GLBenchmark 2.7 TRex renders with artifacts on Gen8 with !UXA</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=89455">Bug 89455</a> - [NVC0/Gallium] Unigine Heaven black and white boxes</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=89689">Bug 89689</a> - [Regression] Weston on DRM backend won't start with new version of mesa</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90130">Bug 90130</a> - gl_PrimitiveId seems to reset at 340</li>
+
+</ul>
+
+
+<h2>Changes</h2>
+
+<p>Boyan Ding (1):</p>
+<ul>
+  <li>i965: Add XRGB8888 format to intel_screen_make_configs</li>
+</ul>
+
+<p>Emil Velikov (3):</p>
+<ul>
+  <li>docs: Add sha256 sums for the 10.5.4 release</li>
+  <li>r300: do not link against libdrm_intel</li>
+  <li>Update version to 10.5.5</li>
+</ul>
+
+<p>Ilia Mirkin (4):</p>
+<ul>
+  <li>nvc0/ir: flush denorms to zero in non-compute shaders</li>
+  <li>gk110/ir: fix set with a register dest to not auto-set the abs flag</li>
+  <li>nvc0/ir: fix predicated PFETCH emission</li>
+  <li>nv50/ir: fix asFlow() const helper for OP_JOIN</li>
+</ul>
+
+<p>Kenneth Graunke (2):</p>
+<ul>
+  <li>i965: Make intel_emit_linear_blit handle Gen8+ alignment restrictions.</li>
+  <li>i965: Disallow linear blits that are not cacheline aligned.</li>
+</ul>
+
+<p>Roland Scheidegger (1):</p>
+<ul>
+  <li>draw: fix prim ids when there's no gs</li>
+</ul>
+
+
+</div>
+</body>
+</html>
--- a/docs/relnotes/10.5.6.html
+++ b/docs/relnotes/10.5.6.html
@@ -0,0 +1,147 @@
+<!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 10.5.6 Release Notes / May 23, 2015</h1>
+
+<p>
+Mesa 10.5.6 is a bug fix release which fixes bugs found since the 10.5.5 release.
+</p>
+<p>
+Mesa 10.5.6 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>
+89ff9cb08d0f6e3f34154864c3071253057cd21020759457c8ae27e0f70985d3  mesa-10.5.6.tar.gz
+66017853bde5f7a6647db3eede30512a091a3491daa1708e0ad8027c328ba595  mesa-10.5.6.tar.xz
+</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=86792">Bug 86792</a> - [NVC0] Portal 2 Crashes in Wine</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90147">Bug 90147</a> - swrast: build error undeclared _SC_PHYS_PAGES on osx</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90350">Bug 90350</a> - [G96] Portal's portal are incorrectly rendered</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90363">Bug 90363</a> - [nv50] HW state is not reset correctly when using a new GL context</li>
+
+</ul>
+
+
+<h2>Changes</h2>
+
+<p>Alex Deucher (1):</p>
+<ul>
+  <li>radeonsi: add new bonaire pci id</li>
+</ul>
+
+<p>Axel Davy (2):</p>
+<ul>
+  <li>egl/wayland: properly destroy wayland objects</li>
+  <li>glx/dri3: Add additional check for gpu offloading case</li>
+</ul>
+
+<p>Emil Velikov (4):</p>
+<ul>
+  <li>docs: Add sha256 sums for the 10.5.5 release</li>
+  <li>egl/main: fix EGL_KHR_get_all_proc_addresses</li>
+  <li>targets/osmesa: drop the -module tag from LDFLAGS</li>
+  <li>Update version to 10.5.6</li>
+</ul>
+
+<p>Francisco Jerez (4):</p>
+<ul>
+  <li>clover: Refactor event::trigger and ::abort to prevent deadlock and reentrancy issues.</li>
+  <li>clover: Wrap event::_status in a method to prevent unlocked access.</li>
+  <li>clover: Implement locking of the wait_count, _chain and _status members of event.</li>
+  <li>i965: Fix PBO cache coherency issue after _mesa_meta_pbo_GetTexSubImage().</li>
+</ul>
+
+<p>Fredrik Höglund (2):</p>
+<ul>
+  <li>main: Require that the texture exists in framebuffer_texture</li>
+  <li>mesa: Generate GL_INVALID_VALUE in framebuffer_texture when layer &lt; 0</li>
+</ul>
+
+<p>Ilia Mirkin (7):</p>
+<ul>
+  <li>nv50/ir: only propagate saturate up if some actual folding took place</li>
+  <li>nv50: keep track of PGRAPH state in nv50_screen</li>
+  <li>nvc0: keep track of PGRAPH state in nvc0_screen</li>
+  <li>nvc0: reset the instanced elements state when doing blit using 3d engine</li>
+  <li>nv50/ir: only enable mul saturate on G200+</li>
+  <li>st/mesa: make sure to create a "clean" bool when doing i2b</li>
+  <li>nvc0: switch mechanism for shader eviction to be a while loop</li>
+</ul>
+
+<p>Jeremy Huddleston Sequoia (2):</p>
+<ul>
+  <li>swrast: Build fix for darwin</li>
+  <li>darwin: Fix install name of libOSMesa</li>
+</ul>
+
+<p>Laura Ekstrand (2):</p>
+<ul>
+  <li>main: Fix an error generated by FramebufferTexture</li>
+  <li>main: Complete error conditions for glInvalidate*Framebuffer.</li>
+</ul>
+
+<p>Marta Lofstedt (1):</p>
+<ul>
+  <li>main: glGetIntegeri_v fails for GL_VERTEX_BINDING_STRIDE</li>
+</ul>
+
+<p>Rob Clark (2):</p>
+<ul>
+  <li>freedreno: enable a306</li>
+  <li>freedreno: fix bug in tile/slot calculation</li>
+</ul>
+
+<p>Roland Scheidegger (1):</p>
+<ul>
+  <li>draw: (trivial) fix out-of-bounds vector initialization</li>
+</ul>
+
+<p>Tim Rowley (1):</p>
+<ul>
+  <li>mesa: fix shininess check for ffvertex_prog v2</li>
+</ul>
+
+<p>Tom Stellard (2):</p>
+<ul>
+  <li>clover: Add a mutex to guard queue::queued_events</li>
+  <li>clover: Fix a bug with multi-threaded events v2</li>
+</ul>
+
+
+</div>
+</body>
+</html>
--- a/docs/relnotes/10.5.7.html
+++ b/docs/relnotes/10.5.7.html
@@ -0,0 +1,103 @@
+<!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 10.5.7 Release Notes / June 07, 2015</h1>
+
+<p>
+Mesa 10.5.7 is a bug fix release which fixes bugs found since the 10.5.6 release.
+</p>
+<p>
+Mesa 10.5.7 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>
+8f865ce497435fdf25d4e35f3b5551b2bcd5f9bc6570561183be82af20d18b82  mesa-10.5.7.tar.gz
+04d06890cd69af8089d6ca76f40e46dcf9cacfe4a9788b32be620574d4638818  mesa-10.5.7.tar.xz
+</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=89131">Bug 89131</a> - [Bisected] Graphical corruption in Weston,  shows old framebuffer pieces</li>
+
+</ul>
+
+
+<h2>Changes</h2>
+
+<p>Ben Widawsky (1):</p>
+<ul>
+  <li>i965: Emit 3DSTATE_MULTISAMPLE before WM_HZ_OP (gen8+)</li>
+</ul>
+
+<p>Emil Velikov (4):</p>
+<ul>
+  <li>docs: Add sha256sums for the 10.5.6 release</li>
+  <li>get-pick-list.sh: Require explicit "10.5" for nominating stable patches</li>
+  <li>cherry-ignore: add clover build fix not applicable for 10.5</li>
+  <li>Update version to 10.5.7</li>
+</ul>
+
+<p>Ilia Mirkin (18):</p>
+<ul>
+  <li>nvc0/ir: set ftz when sources are floats, not just destinations</li>
+  <li>nv50/ir: guess that the constant offset is the starting slot of array</li>
+  <li>nvc0/ir: LOAD's can't be used for shader inputs</li>
+  <li>nvc0: a geometry shader can have up to 1024 vertices output</li>
+  <li>nv50/ir: avoid messing up arg1 of PFETCH</li>
+  <li>nv30: don't leak fragprog consts</li>
+  <li>nv30: avoid leaking render state and draw shaders</li>
+  <li>nv30: fix clip plane uploads and enable changes</li>
+  <li>nv30/draw: avoid leaving stale pointers in draw state</li>
+  <li>nv30/draw: draw expects constbuf size in bytes, not vec4 units</li>
+  <li>st/mesa: don't leak glsl_to_tgsi object on link failure</li>
+  <li>glsl: avoid leaking linked gl_shader when there's a late linker error</li>
+  <li>nv30/draw: fix indexed draws with swtnl path and a resource index buffer</li>
+  <li>nv30/draw: only use the DMA1 object (GART) if the bo is not in VRAM</li>
+  <li>nv30/draw: allocate vertex buffers in gart</li>
+  <li>nv30/draw: switch varying hookup logic to know about texcoords</li>
+  <li>nv30: falling back to draw path for edgeflag does no good</li>
+  <li>nv30: avoid doing extra work on clear and hitting unexpected states</li>
+</ul>
+
+<p>Jason Ekstrand (1):</p>
+<ul>
+  <li>i965/fs: Fix implied_mrf_writes for scratch writes</li>
+</ul>
+
+<p>Marek Olšák (1):</p>
+<ul>
+  <li>st/dri: fix postprocessing crash when there's no depth buffer</li>
+</ul>
+
+
+</div>
+</body>
+</html>
--- a/docs/relnotes/10.5.8.html
+++ b/docs/relnotes/10.5.8.html
@@ -0,0 +1,112 @@
+<!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 10.5.8 Release Notes / June 20, 2015</h1>
+
+<p>
+Mesa 10.5.8 is a bug fix release which fixes bugs found since the 10.5.7 release.
+</p>
+<p>
+Mesa 10.5.8 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>
+611ddcfa3c1bf13f7e6ccac785c8749c3b74c9a78452bac70f8372cf6b209aa0  mesa-10.5.8.tar.gz
+2866b855c5299a4aed066338c77ff6467c389b2c30ada7647be8758663da2b54  mesa-10.5.8.tar.xz
+</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=90310">Bug 90310</a> - Fails to build gallium_dri.so at linking stage with clang because of multiple redefinitions</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90347">Bug 90347</a> - [NVE0+] Failure to insert texbar under some circumstances (causing bad colors in Terasology)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90520">Bug 90520</a> - Register spilling clobbers registers used elsewhere in the shader</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90905">Bug 90905</a> - mesa: Finish subdir-objects transition</li>
+
+</ul>
+
+
+<h2>Changes</h2>
+
+<p>Ben Widawsky (1):</p>
+<ul>
+  <li>i965: Disable compaction for EOT send messages</li>
+</ul>
+
+<p>Boyan Ding (1):</p>
+<ul>
+  <li>egl/x11: Set version of swrastLoader to 2</li>
+</ul>
+
+<p>Emil Velikov (2):</p>
+<ul>
+  <li>docs: Add sha256sums for the 10.5.7 release</li>
+  <li>Update version to 10.5.8</li>
+</ul>
+
+<p>Erik Faye-Lund (1):</p>
+<ul>
+  <li>mesa: build xmlconfig to a separate static library</li>
+</ul>
+
+<p>Francisco Jerez (1):</p>
+<ul>
+  <li>i965: Don't compact instructions with unmapped bits.</li>
+</ul>
+
+<p>Ilia Mirkin (3):</p>
+<ul>
+  <li>nvc0/ir: fix collection of first uses for texture barrier insertion</li>
+  <li>nv50,nvc0: clamp uniform size to 64k</li>
+  <li>nvc0/ir: can't have a join on a load with an indirect source</li>
+</ul>
+
+<p>Jason Ekstrand (1):</p>
+<ul>
+  <li>i965/fs: Don't let the EOT send message interfere with the MRF hack</li>
+</ul>
+
+<p>Marek Olšák (1):</p>
+<ul>
+  <li>egl: fix setting context flags</li>
+</ul>
+
+<p>Roland Scheidegger (1):</p>
+<ul>
+  <li>draw: (trivial) fix NULL pointer dereference</li>
+</ul>
+
+
+</div>
+</body>
+</html>
--- a/docs/relnotes/10.5.9.html
+++ b/docs/relnotes/10.5.9.html
@@ -0,0 +1,140 @@
+<!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 10.5.9 Release Notes / July 04, 2015</h1>
+
+<p>
+Mesa 10.5.9 is a bug fix release which fixes bugs found since the 10.5.8 release.
+</p>
+<p>
+Mesa 10.5.9 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>
+0c081b59572ee9732e7438d34adc3817fe8cc8d4b58abc0e71fd4b4c904945cb  mesa-10.5.9.tar.gz
+71c69f31d3dbc35cfa79950e58a01d27030378d8c7ef1259a0b31d4d0487f4ec  mesa-10.5.9.tar.xz
+</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=84225">Bug 84225</a> - Allow constant-index-expression sampler array indexing with GLSL-ES &lt; 300</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=88999">Bug 88999</a> - [SKL] Compiz crashes after opening unity dash</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=89118">Bug 89118</a> - [SKL Bisected]many Ogles3conform cases core dumped</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90537">Bug 90537</a> - radeonsi bo/va conflict on RADEON_GEM_VA (rscreen-&gt;ws-&gt;buffer_from_handle returns NULL)</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90839">Bug 90839</a> - [10.5.5/10.6 regression, bisected] PBO glDrawPixels no longer using blit fastpath</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=90873">Bug 90873</a> - Kernel hang, TearFree On, Mate desktop environment</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=91056">Bug 91056</a> - The Bard's Tale (2005, native)  has rendering issues</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=91117">Bug 91117</a> - Nimbus (running in wine) has rendering issues, objects are semi-transparent</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=91124">Bug 91124</a> - Civilization V (in Wine) has rendering issues: text missing, menu bar corrupted</li>
+
+</ul>
+
+
+<h2>Changes</h2>
+
+<p>Ben Widawsky (2):</p>
+<ul>
+  <li>i965/gen9: Implement Push Constant Buffer workaround</li>
+  <li>i965/skl: Use 1 register for uniform pull constant payload</li>
+</ul>
+
+<p>Boyan Ding (1):</p>
+<ul>
+  <li>egl/x11: Remove duplicate call to dri2_x11_add_configs_for_visuals</li>
+</ul>
+
+<p>Chris Wilson (3):</p>
+<ul>
+  <li>i965: Fix HW blitter pitch limits</li>
+  <li>i915: Blit RGBX&lt;-&gt;RGBA drawpixels</li>
+  <li>i965: Export format comparison for blitting between miptrees</li>
+</ul>
+
+<p>Emil Velikov (6):</p>
+<ul>
+  <li>docs: Add sha256sums for the 10.5.8 release</li>
+  <li>configure: warn about shared_glapi &amp; xlib-glx only when both are set</li>
+  <li>configure: error out when building backend-less libEGL</li>
+  <li>configure: error out when building libEGL without shared-glapi</li>
+  <li>gbm: do not (over)link against libglapi.so</li>
+  <li>Update version to 10.5.9</li>
+</ul>
+
+<p>Frank Henigman (1):</p>
+<ul>
+  <li>gbm: dlopen libglapi so gbm_create_device works</li>
+</ul>
+
+<p>Ilia Mirkin (8):</p>
+<ul>
+  <li>glsl: add version checks to conditionals for builtin variable enablement</li>
+  <li>mesa: add GL_PROGRAM_PIPELINE support in KHR_debug calls</li>
+  <li>glsl: binding point is a texture unit, which is a combined space</li>
+  <li>nvc0: always put all tfb bufs into bufctx</li>
+  <li>nv50,nvc0: make sure to pushbuf_refn before putting bo into pushbuf_data</li>
+  <li>nv50/ir: propagate modifier to right arg when const-folding mad</li>
+  <li>nv50/ir: fix emission of address reg in 3rd source</li>
+  <li>nv50/ir: copy joinAt when splitting both before and after</li>
+</ul>
+
+<p>Mario Kleiner (2):</p>
+<ul>
+  <li>nouveau: Use dup fd as key in drm-winsys hash table to fix ZaphodHeads.</li>
+  <li>winsys/radeon: Use dup fd as key in drm-winsys hash table to fix ZaphodHeads.</li>
+</ul>
+
+<p>Michel Dänzer (1):</p>
+<ul>
+  <li>winsys/radeon: Unmap GPU VM address range when destroying BO</li>
+</ul>
+
+<p>Tapani Pälli (6):</p>
+<ul>
+  <li>glsl: Allow dynamic sampler array indexing with GLSL ES &lt; 3.00</li>
+  <li>mesa/glsl: new compiler option EmitNoIndirectSampler</li>
+  <li>i915: use EmitNoIndirectSampler</li>
+  <li>mesa/st: use EmitNoIndirectSampler if !ARB_gpu_shader5</li>
+  <li>i965: use EmitNoIndirectSampler for gen &lt; 7</li>
+  <li>glsl: validate sampler array indexing for 'constant-index-expression'</li>
+</ul>
+
+
+</div>
+</body>
+</html>
--- a/include/pci_ids/radeonsi_pci_ids.h
+++ b/include/pci_ids/radeonsi_pci_ids.h
@@ -85,6 +85,7 @@ CHIPSET(0x6651, BONAIRE_6651, BONAIRE)
 CHIPSET(0x6658, BONAIRE_6658, BONAIRE)
 CHIPSET(0x665C, BONAIRE_665C, BONAIRE)
 CHIPSET(0x665D, BONAIRE_665D, BONAIRE)
+CHIPSET(0x665F, BONAIRE_665F, BONAIRE)

 CHIPSET(0x9830, KABINI_9830, KABINI)
 CHIPSET(0x9831, KABINI_9831, KABINI)
--- a/src/egl/drivers/dri2/Android.mk
+++ b/src/egl/drivers/dri2/Android.mk
@@ -40,12 +40,18 @@ LOCAL_C_INCLUDES := \
 	$(MESA_TOP)/src/mapi \
 	$(MESA_TOP)/src/egl/main \
 	$(MESA_TOP)/src/loader \
-	$(TARGET_OUT_HEADERS)/libdrm \
 	$(DRM_GRALLOC_TOP)

 LOCAL_STATIC_LIBRARIES := \
 	libmesa_loader

+LOCAL_SHARED_LIBRARIES := libdrm
+
+ifeq ($(shell echo "$(MESA_ANDROID_VERSION) >= 4.2" | bc),1)
+LOCAL_SHARED_LIBRARIES += \
+	libsync
+endif
+
 LOCAL_MODULE := libmesa_egl_dri2

 include $(MESA_COMMON_MK)
--- a/src/egl/drivers/dri2/egl_dri2.c
+++ b/src/egl/drivers/dri2/egl_dri2.c
@@ -703,6 +703,8 @@ dri2_terminate(_EGLDriver *drv, _EGLDisplay *disp)
 #ifdef HAVE_WAYLAND_PLATFORM
   case _EGL_PLATFORM_WAYLAND:
      wl_drm_destroy(dri2_dpy->wl_drm);
+      wl_registry_destroy(dri2_dpy->wl_registry);
+      wl_event_queue_destroy(dri2_dpy->wl_queue);
      if (dri2_dpy->own_device) {
         wl_display_disconnect(dri2_dpy->wl_dpy);
      }
--- a/src/egl/drivers/dri2/platform_wayland.c
+++ b/src/egl/drivers/dri2/platform_wayland.c
@@ -1028,7 +1028,7 @@ dri2_initialize_wayland(_EGLDriver *drv, _EGLDisplay *disp)
   wl_registry_add_listener(dri2_dpy->wl_registry,
                            &registry_listener, dri2_dpy);
   if (roundtrip(dri2_dpy) < 0 || dri2_dpy->wl_drm == NULL)
-      goto cleanup_dpy;
+      goto cleanup_registry;

   if (roundtrip(dri2_dpy) < 0 || dri2_dpy->fd == -1)
      goto cleanup_drm;
@@ -1111,6 +1111,9 @@ dri2_initialize_wayland(_EGLDriver *drv, _EGLDisplay *disp)
 cleanup_drm:
   free(dri2_dpy->device_name);
   wl_drm_destroy(dri2_dpy->wl_drm);
+ cleanup_registry:
+   wl_registry_destroy(dri2_dpy->wl_registry);
+   wl_event_queue_destroy(dri2_dpy->wl_queue);
 cleanup_dpy:
   free(dri2_dpy);
   
--- a/src/egl/drivers/dri2/platform_x11.c
+++ b/src/egl/drivers/dri2/platform_x11.c
@@ -1120,7 +1120,7 @@ dri2_initialize_x11_swrast(_EGLDriver *drv, _EGLDisplay *disp)
      goto cleanup_conn;

   dri2_dpy->swrast_loader_extension.base.name = __DRI_SWRAST_LOADER;
-   dri2_dpy->swrast_loader_extension.base.version = __DRI_SWRAST_LOADER_VERSION;
+   dri2_dpy->swrast_loader_extension.base.version = 2;
   dri2_dpy->swrast_loader_extension.getDrawableInfo = swrastGetDrawableInfo;
   dri2_dpy->swrast_loader_extension.putImage = swrastPutImage;
   dri2_dpy->swrast_loader_extension.getImage = swrastGetImage;
@@ -1291,11 +1291,6 @@ dri2_initialize_x11_dri2(_EGLDriver *drv, _EGLDisplay *disp)

   dri2_x11_setup_swap_interval(dri2_dpy);

-   if (dri2_dpy->conn) {
-      if (!dri2_x11_add_configs_for_visuals(dri2_dpy, disp))
-	 goto cleanup_configs;
-   }
-
   disp->Extensions.KHR_image_pixmap = EGL_TRUE;
   disp->Extensions.NOK_swap_region = EGL_TRUE;
   disp->Extensions.NOK_texture_from_pixmap = EGL_TRUE;
--- a/src/egl/main/eglapi.c
+++ b/src/egl/main/eglapi.c
@@ -1029,8 +1029,9 @@ eglGetProcAddress(const char *procname)
      const char *name;
      _EGLProc function;
   } egl_functions[] = {
-      /* core functions should not be queryable, but, well... */
-#ifdef _EGL_GET_CORE_ADDRESSES
+      /* core functions queryable in the presence of
+       * EGL_KHR_get_all_proc_addresses or EGL 1.5
+       */
      /* alphabetical order */
      { "eglBindAPI", (_EGLProc) eglBindAPI },
      { "eglBindTexImage", (_EGLProc) eglBindTexImage },
@@ -1066,7 +1067,6 @@ eglGetProcAddress(const char *procname)
      { "eglWaitClient", (_EGLProc) eglWaitClient },
      { "eglWaitGL", (_EGLProc) eglWaitGL },
      { "eglWaitNative", (_EGLProc) eglWaitNative },
-#endif /* _EGL_GET_CORE_ADDRESSES */
 #ifdef EGL_MESA_screen_surface
      { "eglChooseModeMESA", (_EGLProc) eglChooseModeMESA },
      { "eglGetModesMESA", (_EGLProc) eglGetModesMESA },
--- a/src/egl/main/eglcontext.c
+++ b/src/egl/main/eglcontext.c
@@ -131,7 +131,7 @@ _eglParseContextAttribList(_EGLContext *ctx, _EGLDisplay *dpy,
            break;
         }

-         ctx->Flags = val;
+         ctx->Flags |= val;
         break;

      case EGL_CONTEXT_OPENGL_PROFILE_MASK_KHR:
@@ -194,7 +194,8 @@ _eglParseContextAttribList(_EGLContext *ctx, _EGLDisplay *dpy,
            break;
         }

-         ctx->Flags = EGL_CONTEXT_OPENGL_ROBUST_ACCESS_BIT_KHR;
+         if (val == EGL_TRUE)
+            ctx->Flags |= EGL_CONTEXT_OPENGL_ROBUST_ACCESS_BIT_KHR;
         break;

      default:
--- a/src/gallium/auxiliary/Android.mk
+++ b/src/gallium/auxiliary/Android.mk
@@ -33,8 +33,7 @@ LOCAL_SRC_FILES := \
 	$(VL_STUB_SOURCES)

 LOCAL_C_INCLUDES := \
-	$(GALLIUM_TOP)/auxiliary/util \
-	$(MESA_TOP)/src
+	$(GALLIUM_TOP)/auxiliary/util

 LOCAL_MODULE := libmesa_gallium

--- a/src/gallium/auxiliary/draw/draw_context.c
+++ b/src/gallium/auxiliary/draw/draw_context.c
@@ -182,6 +182,7 @@ boolean draw_init(struct draw_context *draw)
 void draw_new_instance(struct draw_context *draw)
 {
   draw_geometry_shader_new_instance(draw->gs.geometry_shader);
+   draw_prim_assembler_new_instance(draw->ia);
 }


--- a/src/gallium/auxiliary/draw/draw_llvm.c
+++ b/src/gallium/auxiliary/draw/draw_llvm.c
@@ -1966,7 +1966,7 @@ draw_llvm_set_sampler_state(struct draw_context *draw,
      for (i = 0; i < draw->num_samplers[PIPE_SHADER_VERTEX]; i++) {
         struct draw_jit_sampler *jit_sam = &draw->llvm->jit_context.samplers[i];

-         if (draw->samplers[i]) {
+         if (draw->samplers[PIPE_SHADER_VERTEX][i]) {
            const struct pipe_sampler_state *s
               = draw->samplers[PIPE_SHADER_VERTEX][i];
            jit_sam->min_lod = s->min_lod;
@@ -1979,7 +1979,7 @@ draw_llvm_set_sampler_state(struct draw_context *draw,
      for (i = 0; i < draw->num_samplers[PIPE_SHADER_GEOMETRY]; i++) {
         struct draw_jit_sampler *jit_sam = &draw->llvm->gs_jit_context.samplers[i];

-         if (draw->samplers[i]) {
+         if (draw->samplers[PIPE_SHADER_GEOMETRY][i]) {
            const struct pipe_sampler_state *s
               = draw->samplers[PIPE_SHADER_GEOMETRY][i];
            jit_sam->min_lod = s->min_lod;
@@ -2049,7 +2049,7 @@ generate_mask_value(struct draw_gs_llvm_variant *variant,

   num_prims = lp_build_broadcast(gallivm, lp_build_vec_type(gallivm, mask_type),
                                  variant->num_prims);
-   for (i = 0; i <= gs_type.length; i++) {
+   for (i = 0; i < gs_type.length; i++) {
      LLVMValueRef idx = lp_build_const_int32(gallivm, i);
      mask_val = LLVMBuildInsertElement(builder, mask_val, idx, idx, "");
   }
--- a/src/gallium/auxiliary/draw/draw_prim_assembler.c
+++ b/src/gallium/auxiliary/draw/draw_prim_assembler.c
@@ -189,7 +189,6 @@ draw_prim_assembler_prepare_outputs(struct draw_assembler *ia)
   } else {
      ia->primid_slot = -1;
   }
-   ia->primid = 0;
 }


@@ -233,7 +232,6 @@ draw_prim_assembler_run(struct draw_context *draw,
   asmblr->input_prims = input_prims;
   asmblr->input_verts = input_verts;
   asmblr->needs_primid = needs_primid(asmblr->draw);
-   asmblr->primid = 0;
   asmblr->num_prims = 0;

   output_prims->linear = TRUE;
@@ -284,3 +282,14 @@ draw_prim_assembler_destroy(struct draw_assembler *ia)
 {
   FREE(ia);
 }
+
+
+/*
+ * Called at the very begin of the draw call with a new instance
+ * Used to reset state that should persist between primitive restart.
+ */
+void
+draw_prim_assembler_new_instance(struct draw_assembler *asmblr)
+{
+   asmblr->primid = 0;
+}
--- a/src/gallium/auxiliary/draw/draw_prim_assembler.h
+++ b/src/gallium/auxiliary/draw/draw_prim_assembler.h
@@ -70,5 +70,8 @@ draw_prim_assembler_run(struct draw_context *draw,
 void
 draw_prim_assembler_prepare_outputs(struct draw_assembler *ia);

+void
+draw_prim_assembler_new_instance(struct draw_assembler *ia);
+

 #endif
--- a/src/gallium/drivers/freedreno/Android.mk
+++ b/src/gallium/drivers/freedreno/Android.mk
@@ -34,10 +34,9 @@ LOCAL_CFLAGS := \
 	-Wno-packed-bitfield-compat

 LOCAL_C_INCLUDES := \
-	$(LOCAL_PATH)/ir3 \
-	$(TARGET_OUT_HEADERS)/libdrm \
-	$(TARGET_OUT_HEADERS)/freedreno
+	$(LOCAL_PATH)/ir3

+LOCAL_SHARED_LIBRARIES := libdrm libdrm_freedreno
 LOCAL_MODULE := libmesa_pipe_freedreno

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/drivers/freedreno/freedreno_context.h
+++ b/src/gallium/drivers/freedreno/freedreno_context.h
@@ -293,7 +293,7 @@ struct fd_context {
 	 */
 	struct fd_gmem_stateobj gmem;
 	struct fd_vsc_pipe      pipe[8];
-	struct fd_tile          tile[64];
+	struct fd_tile          tile[256];

 	/* which state objects need to be re-emit'd: */
 	enum {
--- a/src/gallium/drivers/freedreno/freedreno_gmem.c
+++ b/src/gallium/drivers/freedreno/freedreno_gmem.c
@@ -91,6 +91,7 @@ calculate_tiles(struct fd_context *ctx)
 	uint32_t i, j, t, xoff, yoff;
 	uint32_t tpp_x, tpp_y;
 	bool has_zs = !!(ctx->resolve & (FD_BUFFER_DEPTH | FD_BUFFER_STENCIL));
+	int tile_n[ARRAY_SIZE(ctx->pipe)];

 	if (pfb->cbufs[0])
 		cpp = util_format_get_blocksize(pfb->cbufs[0]->format);
@@ -213,6 +214,7 @@ calculate_tiles(struct fd_context *ctx)
 	/* configure tiles: */
 	t = 0;
 	yoff = miny;
+	memset(tile_n, 0, sizeof(tile_n));
 	for (i = 0; i < nbins_y; i++) {
 		uint32_t bw, bh;

@@ -223,20 +225,17 @@ calculate_tiles(struct fd_context *ctx)

 		for (j = 0; j < nbins_x; j++) {
 			struct fd_tile *tile = &ctx->tile[t];
-			uint32_t n, p;
+			uint32_t p;

 			assert(t < ARRAY_SIZE(ctx->tile));

 			/* pipe number: */
 			p = ((i / tpp_y) * div_round_up(nbins_x, tpp_x)) + (j / tpp_x);

-			/* slot number: */
-			n = ((i % tpp_y) * tpp_x) + (j % tpp_x);
-
 			/* clip bin width: */
 			bw = MIN2(bin_w, minx + width - xoff);

-			tile->n = n;
+			tile->n = tile_n[p]++;
 			tile->p = p;
 			tile->bin_w = bw;
 			tile->bin_h = bh;
--- a/src/gallium/drivers/freedreno/freedreno_screen.c
+++ b/src/gallium/drivers/freedreno/freedreno_screen.c
@@ -531,6 +531,7 @@ fd_screen_create(struct fd_device *dev)
 	case 220:
 		fd2_screen_init(pscreen);
 		break;
+	case 307:
 	case 320:
 	case 330:
 		fd3_screen_init(pscreen);
--- a/src/gallium/drivers/nouveau/Android.mk
+++ b/src/gallium/drivers/nouveau/Android.mk
@@ -36,9 +36,7 @@ LOCAL_SRC_FILES := \
 	$(NVC0_CODEGEN_SOURCES) \
 	$(NVC0_C_SOURCES)

-LOCAL_C_INCLUDES := \
-	$(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_nouveau
 LOCAL_MODULE := libmesa_pipe_nouveau

 include external/stlport/libstlport.mk
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_bb.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_bb.cpp
@@ -332,6 +332,9 @@ BasicBlock::splitBefore(Instruction *insn, bool attach)
   BasicBlock *bb = new BasicBlock(func);
   assert(!insn || insn->op != OP_PHI);

+   bb->joinAt = joinAt;
+   joinAt = NULL;
+
   splitCommon(insn, bb, attach);
   return bb;
 }
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_gk110.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_gk110.cpp
@@ -903,7 +903,7 @@ CodeEmitterGK110::emitSET(const CmpInstruction *i)
      code[0] |= 0x1c;
   } else {
      switch (i->sType) {
-      case TYPE_F32: op2 = 0x000; op1 = 0x820; break;
+      case TYPE_F32: op2 = 0x000; op1 = 0x800; break;
      case TYPE_F64: op2 = 0x080; op1 = 0x900; break;
      default:
         op2 = 0x1a8;
@@ -1265,8 +1265,10 @@ CodeEmitterGK110::emitPFETCH(const Instruction *i)

   emitPredicate(i);

+   const int src1 = (i->predSrc == 1) ? 2 : 1; // if predSrc == 1, !srcExists(2)
+
   defId(i->def(0), 2);
-   srcId(i->src(1), 10);
+   srcId(i, src1, 10);
 }

 void
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp
@@ -499,10 +499,14 @@ CodeEmitterNV50::emitForm_MAD(const Instruction *i)
   setSrc(i, 2, 2);

   if (i->getIndirect(0, 0)) {
-      assert(!i->getIndirect(1, 0));
+      assert(!i->srcExists(1) || !i->getIndirect(1, 0));
+      assert(!i->srcExists(2) || !i->getIndirect(2, 0));
      setAReg16(i, 0);
-   } else {
+   } else if (i->srcExists(1) && i->getIndirect(1, 0)) {
+      assert(!i->srcExists(2) || !i->getIndirect(2, 0));
      setAReg16(i, 1);
+   } else {
+      setAReg16(i, 2);
   }
 }

--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nvc0.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nvc0.cpp
@@ -1441,8 +1441,10 @@ CodeEmitterNVC0::emitPFETCH(const Instruction *i)

   emitPredicate(i);

+   const int src1 = (i->predSrc == 1) ? 2 : 1; // if predSrc == 1, !srcExists(2)
+
   defId(i->def(0), 14);
-   srcId(i->src(1), 20);
+   srcId(i, src1, 20);
 }

 void
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_from_tgsi.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_from_tgsi.cpp
@@ -1270,18 +1270,20 @@ Converter::srcToSym(tgsi::Instruction::SrcRegister src, int c)
 {
   const int swz = src.getSwizzle(c);

+   /* TODO: Use Array ID when it's available for the index */
   return makeSym(src.getFile(),
                  src.is2D() ? src.getIndex(1) : 0,
-                  src.isIndirect(0) ? -1 : src.getIndex(0), swz,
+                  src.getIndex(0), swz,
                  src.getIndex(0) * 16 + swz * 4);
 }

 Symbol *
 Converter::dstToSym(tgsi::Instruction::DstRegister dst, int c)
 {
+   /* TODO: Use Array ID when it's available for the index */
   return makeSym(dst.getFile(),
                  dst.is2D() ? dst.getIndex(1) : 0,
-                  dst.isIndirect(0) ? -1 : dst.getIndex(0), c,
+                  dst.getIndex(0), c,
                  dst.getIndex(0) * 16 + c * 4);
 }

@@ -1610,6 +1612,7 @@ Converter::insertConvergenceOps(BasicBlock *conv, BasicBlock *fork)
   join->fixed = 1;
   conv->insertHead(join);

+   assert(!fork->joinAt);
   fork->joinAt = new_FlowInstruction(func, OP_JOINAT, conv);
   fork->insertBefore(fork->getExit(), fork->joinAt);
 }
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_inlines.h
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_inlines.h
@@ -302,7 +302,7 @@ FlowInstruction *Instruction::asFlow()

 const FlowInstruction *Instruction::asFlow() const
 {
-   if (op >= OP_BRA && op <= OP_JOINAT)
+   if (op >= OP_BRA && op <= OP_JOIN)
      return static_cast<const FlowInstruction *>(this);
   return NULL;
 }
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_gm107.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_gm107.cpp
@@ -240,6 +240,7 @@ GM107LoweringPass::visit(Instruction *i)
            Value *ptr = bld.mkOp2v(OP_SHL, TYPE_U32, bld.getSSA(),
                                    i->getIndirect(0, 0), bld.mkImm(4));
            i->setIndirect(0, 0, ptr);
+            i->op = OP_VFETCH;
         } else {
            i->op = OP_VFETCH;
            assert(prog->getType() != Program::TYPE_FRAGMENT); // INTERP
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
@@ -871,6 +871,7 @@ NV50LoweringPreSSA::handleTXL(TexInstruction *i)
   BasicBlock *joinBB = i->bb->splitAfter(i);

   bld.setPosition(currBB, true);
+   assert(!currBB->joinAt);
   currBB->joinAt = bld.mkFlow(OP_JOINAT, joinBB, CC_ALWAYS, NULL);

   for (int l = 0; l <= 3; ++l) {
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp
@@ -73,6 +73,25 @@ NVC0LegalizeSSA::handleRCPRSQ(Instruction *i)
   // TODO
 }

+void
+NVC0LegalizeSSA::handleFTZ(Instruction *i)
+{
+   // Only want to flush float inputs
+   assert(i->sType == TYPE_F32);
+
+   // If we're already flushing denorms (and NaN's) to zero, no need for this.
+   if (i->dnz)
+      return;
+
+   // Only certain classes of operations can flush
+   OpClass cls = prog->getTarget()->getOpClass(i->op);
+   if (cls != OPCLASS_ARITH && cls != OPCLASS_COMPARE &&
+       cls != OPCLASS_CONVERT)
+      return;
+
+   i->ftz = true;
+}
+
 bool
 NVC0LegalizeSSA::visit(Function *fn)
 {
@@ -86,8 +105,11 @@ NVC0LegalizeSSA::visit(BasicBlock *bb)
   Instruction *next;
   for (Instruction *i = bb->getEntry(); i; i = next) {
      next = i->next;
-      if (i->dType == TYPE_F32)
+      if (i->sType == TYPE_F32) {
+         if (prog->getType() != Program::TYPE_COMPUTE)
+            handleFTZ(i);
         continue;
+      }
      switch (i->op) {
      case OP_DIV:
      case OP_MOD:
@@ -123,7 +145,7 @@ NVC0LegalizePostRA::insnDominatedBy(const Instruction *later,

 void
 NVC0LegalizePostRA::addTexUse(std::list<TexUse> &uses,
-                              Instruction *usei, const Instruction *insn)
+                              Instruction *usei, const Instruction *texi)
 {
   bool add = true;
   for (std::list<TexUse>::iterator it = uses.begin();
@@ -138,7 +160,7 @@ NVC0LegalizePostRA::addTexUse(std::list<TexUse> &uses,
         ++it;
   }
   if (add)
-      uses.push_back(TexUse(usei, insn));
+      uses.push_back(TexUse(usei, texi));
 }

 void
@@ -150,7 +172,8 @@ NVC0LegalizePostRA::findOverwritingDefs(const Instruction *texi,
   while (insn->op == OP_MOV && insn->getDef(0)->equals(insn->getSrc(0)))
      insn = insn->getSrc(0)->getUniqueInsn();

-   if (!insn->bb->reachableBy(texi->bb, term))
+   // NOTE: the tex itself is, of course, not an overwriting definition
+   if (insn == texi || !insn->bb->reachableBy(texi->bb, term))
      return;

   switch (insn->op) {
@@ -198,7 +221,12 @@ NVC0LegalizePostRA::findFirstUses(
         visited.insert(usei);

         if (usei->op == OP_PHI || usei->op == OP_UNION) {
-            // need a barrier before WAW cases
+            // need a barrier before WAW cases, like:
+            //   %r0 = tex
+            //   if ...
+            //     texbar <- is required or tex might replace x again
+            //     %r1 = x <- overwriting def
+            //   %r2 = phi %r0, %r1
            for (int s = 0; usei->srcExists(s); ++s) {
               Instruction *defi = usei->getSrc(s)->getUniqueInsn();
               if (defi && &usei->src(s) != *u)
@@ -217,7 +245,7 @@ NVC0LegalizePostRA::findFirstUses(
             usei->subOp != NV50_IR_SUBOP_MOV_FINAL) {
            findFirstUses(texi, usei, uses, visited);
         } else {
-            addTexUse(uses, usei, insn);
+            addTexUse(uses, usei, texi);
         }
      }
   }
@@ -1692,6 +1720,7 @@ NVC0LoweringPass::visit(Instruction *i)
            Value *ptr = bld.mkOp2v(OP_SHL, TYPE_U32, bld.getSSA(),
                                    i->getIndirect(0, 0), bld.mkImm(4));
            i->setIndirect(0, 0, ptr);
+            i->op = OP_VFETCH;
         } else {
            i->op = OP_VFETCH;
            assert(prog->getType() != Program::TYPE_FRAGMENT); // INTERP
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.h
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.h
@@ -36,6 +36,7 @@ private:
   // we want to insert calls to the builtin library only after optimization
   void handleDIV(Instruction *); // integer division, modulus
   void handleRCPRSQ(Instruction *); // double precision float recip/rsqrt
+   void handleFTZ(Instruction *);

 private:
   BuildUtil bld;
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp
@@ -207,6 +207,9 @@ LoadPropagation::visit(BasicBlock *bb)
      if (i->op == OP_CALL) // calls have args as sources, they must be in regs
         continue;

+      if (i->op == OP_PFETCH) // pfetch expects arg1 to be a reg
+         continue;
+
      if (i->srcExists(1))
         checkSwapSrc01(i);

@@ -545,6 +548,11 @@ ConstantFolding::expr(Instruction *i,
   case OP_POPCNT:
      res.data.u32 = util_bitcount(a->data.u32 & b->data.u32);
      break;
+   case OP_PFETCH:
+      // The two arguments to pfetch are logically added together. Normally
+      // the second argument will not be constant, but that can happen.
+      res.data.u32 = a->data.u32 + b->data.u32;
+      break;
   default:
      return;
   }
@@ -559,12 +567,17 @@ ConstantFolding::expr(Instruction *i,

   i->getSrc(0)->reg.data = res.data;

-   if (i->op == OP_MAD || i->op == OP_FMA) {
+   switch (i->op) {
+   case OP_MAD:
+   case OP_FMA: {
      i->op = OP_ADD;

+      /* Move the immediate to the second arg, otherwise the ADD operation
+       * won't be emittable
+       */
      i->setSrc(1, i->getSrc(0));
-      i->src(1).mod = i->src(2).mod;
      i->setSrc(0, i->getSrc(2));
+      i->src(0).mod = i->src(2).mod;
      i->setSrc(2, NULL);

      ImmediateValue src0;
@@ -574,8 +587,14 @@ ConstantFolding::expr(Instruction *i,
         bld.setPosition(i, false);
         i->setSrc(1, bld.loadImm(NULL, res.data.u32));
      }
-   } else {
+      break;
+   }
+   case OP_PFETCH:
+      // Leave PFETCH alone... we just folded its 2 args into 1.
+      break;
+   default:
      i->op = i->saturate ? OP_SAT : OP_MOV; /* SAT handled by unary() */
+      break;
   }
   i->subOp = 0;
 }
@@ -676,6 +695,7 @@ ConstantFolding::tryCollapseChainedMULs(Instruction *mul2,
            mul1->setSrc(s1, bld.loadImm(NULL, f * imm1.reg.data.f32));
            mul1->src(s1).mod = Modifier(0);
            mul2->def(0).replace(mul1->getDef(0), false);
+            mul1->saturate = mul2->saturate;
         } else
         if (prog->getTarget()->isPostMultiplySupported(OP_MUL, f, e)) {
            // c = mul a, b
@@ -684,8 +704,8 @@ ConstantFolding::tryCollapseChainedMULs(Instruction *mul2,
            mul2->def(0).replace(mul1->getDef(0), false);
            if (f < 0)
               mul1->src(0).mod *= Modifier(NV50_IR_MOD_NEG);
+            mul1->saturate = mul2->saturate;
         }
-         mul1->saturate = mul2->saturate;
         return;
      }
   }
@@ -2186,7 +2206,7 @@ FlatteningPass::visit(BasicBlock *bb)
             insn->op != OP_LINTERP && // probably just nve4
             insn->op != OP_PINTERP && // probably just nve4
             ((insn->op != OP_LOAD && insn->op != OP_STORE) ||
-              typeSizeof(insn->dType) <= 4) &&
+              (typeSizeof(insn->dType) <= 4 && !insn->src(0).isIndirect(0))) &&
             !insn->isNop()) {
            insn->join = 1;
            bb->remove(bb->getExit());
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_target_nv50.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_target_nv50.cpp
@@ -84,7 +84,7 @@ static const struct opProperties _initProps[] =
   //           neg  abs  not  sat  c[]  s[], a[], imm
   { OP_ADD,    0x3, 0x0, 0x0, 0x8, 0x2, 0x1, 0x1, 0x2 },
   { OP_SUB,    0x3, 0x0, 0x0, 0x8, 0x2, 0x1, 0x1, 0x2 },
-   { OP_MUL,    0x3, 0x0, 0x0, 0x8, 0x2, 0x1, 0x1, 0x2 },
+   { OP_MUL,    0x3, 0x0, 0x0, 0x0, 0x2, 0x1, 0x1, 0x2 },
   { OP_MAX,    0x3, 0x3, 0x0, 0x0, 0x2, 0x1, 0x1, 0x0 },
   { OP_MIN,    0x3, 0x3, 0x0, 0x0, 0x2, 0x1, 0x1, 0x0 },
   { OP_MAD,    0x7, 0x0, 0x0, 0x8, 0x6, 0x1, 0x1, 0x0 }, // special constraint
@@ -188,6 +188,9 @@ void TargetNV50::initOpInfo()
      if (prop->mSat & 8)
         opInfo[prop->op].dstMods = NV50_IR_MOD_SAT;
   }
+
+   if (chipset >= 0xa0)
+      opInfo[OP_MUL].dstMods = NV50_IR_MOD_SAT;
 }

 unsigned int
--- a/src/gallium/drivers/nouveau/nv30/nv30_clear.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_clear.c
@@ -58,7 +58,7 @@ nv30_clear(struct pipe_context *pipe, unsigned buffers,
   struct pipe_framebuffer_state *fb = &nv30->framebuffer;
   uint32_t colr = 0, zeta = 0, mode = 0;

-   if (!nv30_state_validate(nv30, TRUE))
+   if (!nv30_state_validate(nv30, NV30_NEW_FRAMEBUFFER | NV30_NEW_SCISSOR, TRUE))
      return;

   if (buffers & PIPE_CLEAR_COLOR && fb->nr_cbufs) {
--- a/src/gallium/drivers/nouveau/nv30/nv30_context.h
+++ b/src/gallium/drivers/nouveau/nv30/nv30_context.h
@@ -204,7 +204,7 @@ void
 nv30_render_vbo(struct pipe_context *pipe, const struct pipe_draw_info *info);

 boolean
-nv30_state_validate(struct nv30_context *nv30, boolean hwtnl);
+nv30_state_validate(struct nv30_context *nv30, uint32_t mask, boolean hwtnl);

 void
 nv30_state_release(struct nv30_context *nv30);
--- a/src/gallium/drivers/nouveau/nv30/nv30_draw.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_draw.c
@@ -71,12 +71,12 @@ nv30_render_allocate_vertices(struct vbuf_render *render,
   struct nv30_render *r = nv30_render(render);
   struct nv30_context *nv30 = r->nv30;

-   r->length = vertex_size * nr_vertices;
+   r->length = (uint32_t)vertex_size * (uint32_t)nr_vertices;

   if (r->offset + r->length >= render->max_vertex_buffer_bytes) {
      pipe_resource_reference(&r->buffer, NULL);
      r->buffer = pipe_buffer_create(&nv30->screen->base.base,
-                                     PIPE_BIND_VERTEX_BUFFER, 0,
+                                     PIPE_BIND_VERTEX_BUFFER, PIPE_USAGE_STREAM,
                                     render->max_vertex_buffer_bytes);
      if (!r->buffer)
         return FALSE;
@@ -91,10 +91,14 @@ static void *
 nv30_render_map_vertices(struct vbuf_render *render)
 {
   struct nv30_render *r = nv30_render(render);
-   char *map = pipe_buffer_map(&r->nv30->base.pipe, r->buffer,
-                               PIPE_TRANSFER_WRITE |
-                               PIPE_TRANSFER_UNSYNCHRONIZED, &r->transfer);
-   return map + r->offset;
+   char *map = pipe_buffer_map_range(
+         &r->nv30->base.pipe, r->buffer,
+         r->offset, r->length,
+         PIPE_TRANSFER_WRITE |
+         PIPE_TRANSFER_DISCARD_RANGE,
+         &r->transfer);
+   assert(map);
+   return map;
 }

 static void
@@ -103,6 +107,7 @@ nv30_render_unmap_vertices(struct vbuf_render *render,
 {
   struct nv30_render *r = nv30_render(render);
   pipe_buffer_unmap(&r->nv30->base.pipe, r->transfer);
+   r->transfer = NULL;
 }

 static void
@@ -126,10 +131,10 @@ nv30_render_draw_elements(struct vbuf_render *render,
   for (i = 0; i < r->vertex_info.num_attribs; i++) {
      PUSH_RESRC(push, NV30_3D(VTXBUF(i)), BUFCTX_VTXTMP,
                       nv04_resource(r->buffer), r->offset + r->vtxptr[i],
-                       NOUVEAU_BO_LOW | NOUVEAU_BO_RD, 0, 0);
+                       NOUVEAU_BO_LOW | NOUVEAU_BO_RD, 0, NV30_3D_VTXBUF_DMA1);
   }

-   if (!nv30_state_validate(nv30, FALSE))
+   if (!nv30_state_validate(nv30, ~0, FALSE))
      return;

   BEGIN_NV04(push, NV30_3D(VERTEX_BEGIN_END), 1);
@@ -171,10 +176,10 @@ nv30_render_draw_arrays(struct vbuf_render *render, unsigned start, uint nr)
   for (i = 0; i < r->vertex_info.num_attribs; i++) {
      PUSH_RESRC(push, NV30_3D(VTXBUF(i)), BUFCTX_VTXTMP,
                       nv04_resource(r->buffer), r->offset + r->vtxptr[i],
-                       NOUVEAU_BO_LOW | NOUVEAU_BO_RD, 0, 0);
+                       NOUVEAU_BO_LOW | NOUVEAU_BO_RD, 0, NV30_3D_VTXBUF_DMA1);
   }

-   if (!nv30_state_validate(nv30, FALSE))
+   if (!nv30_state_validate(nv30, ~0, FALSE))
      return;

   BEGIN_NV04(push, NV30_3D(VERTEX_BEGIN_END), 1);
@@ -213,22 +218,24 @@ static const struct {
   [TGSI_SEMANTIC_BCOLOR  ] = { EMIT_4F, INTERP_LINEAR     , 1, 3, 0x00000004 },
   [TGSI_SEMANTIC_FOG     ] = { EMIT_4F, INTERP_PERSPECTIVE, 5, 5, 0x00000010 },
   [TGSI_SEMANTIC_PSIZE   ] = { EMIT_1F_PSIZE, INTERP_POS  , 6, 6, 0x00000020 },
-   [TGSI_SEMANTIC_GENERIC ] = { EMIT_4F, INTERP_PERSPECTIVE, 8, 7, 0x00004000 }
+   [TGSI_SEMANTIC_TEXCOORD] = { EMIT_4F, INTERP_PERSPECTIVE, 8, 7, 0x00004000 },
 };

 static boolean
 vroute_add(struct nv30_render *r, uint attrib, uint sem, uint *idx)
 {
-   struct pipe_screen *pscreen = &r->nv30->screen->base.base;
+   struct nv30_screen *screen = r->nv30->screen;
   struct nv30_fragprog *fp = r->nv30->fragprog.program;
   struct vertex_info *vinfo = &r->vertex_info;
   enum pipe_format format;
   uint emit = EMIT_OMIT;
   uint result = *idx;

-   if (sem == TGSI_SEMANTIC_GENERIC && result >= 8) {
-      for (result = 0; result < 8; result++) {
-         if (fp->texcoord[result] == *idx) {
+   if (sem == TGSI_SEMANTIC_GENERIC) {
+      uint num_texcoords = (screen->eng3d->oclass < NV40_3D_CLASS) ? 8 : 10;
+      for (result = 0; result < num_texcoords; result++) {
+         if (fp->texcoord[result] == *idx + 8) {
+            sem = TGSI_SEMANTIC_TEXCOORD;
            emit = vroute[sem].emit;
            break;
         }
@@ -243,11 +250,11 @@ vroute_add(struct nv30_render *r, uint attrib, uint sem, uint *idx)
   draw_emit_vertex_attr(vinfo, emit, vroute[sem].interp, attrib);
   format = draw_translate_vinfo_format(emit);

-   r->vtxfmt[attrib] = nv30_vtxfmt(pscreen, format)->hw;
-   r->vtxptr[attrib] = vinfo->size | NV30_3D_VTXBUF_DMA1;
+   r->vtxfmt[attrib] = nv30_vtxfmt(&screen->base.base, format)->hw;
+   r->vtxptr[attrib] = vinfo->size;
   vinfo->size += draw_translate_vinfo_size(emit);

-   if (nv30_screen(pscreen)->eng3d->oclass < NV40_3D_CLASS) {
+   if (screen->eng3d->oclass < NV40_3D_CLASS) {
      r->vtxprog[attrib][0] = 0x001f38d8;
      r->vtxprog[attrib][1] = 0x0080001b | (attrib << 9);
      r->vtxprog[attrib][2] = 0x0836106c;
@@ -259,7 +266,12 @@ vroute_add(struct nv30_render *r, uint attrib, uint sem, uint *idx)
      r->vtxprog[attrib][3] = 0x6041ff80 | (result + vroute[sem].vp40) << 2;
   }

-   *idx = vroute[sem].ow40 << result;
+   if (result < 8)
+      *idx = vroute[sem].ow40 << result;
+   else {
+      assert(sem == TGSI_SEMANTIC_TEXCOORD);
+      *idx = 0x00001000 << (result - 8);
+   }
   return TRUE;
 }

@@ -313,7 +325,7 @@ nv30_render_validate(struct nv30_context *nv30)

   while (pntc && attrib < 16) {
      uint index = ffs(pntc) - 1; pntc &= ~(1 << index);
-      if (vroute_add(r, attrib, TGSI_SEMANTIC_GENERIC, &index)) {
+      if (vroute_add(r, attrib, TGSI_SEMANTIC_TEXCOORD, &index)) {
         vp_attribs |= (1 << attrib++);
         vp_results |= index;
      }
@@ -398,17 +410,17 @@ nv30_render_vbo(struct pipe_context *pipe, const struct pipe_draw_info *info)
      if (nv30->vertprog.constbuf) {
         void *map = nv04_resource(nv30->vertprog.constbuf)->data;
         draw_set_mapped_constant_buffer(draw, PIPE_SHADER_VERTEX, 0,
-                                         map, nv30->vertprog.constbuf_nr);
+                                         map, nv30->vertprog.constbuf_nr * 16);
+      } else {
+         draw_set_mapped_constant_buffer(draw, PIPE_SHADER_VERTEX, 0, NULL, 0);
      }
   }

   for (i = 0; i < nv30->num_vtxbufs; i++) {
      const void *map = nv30->vtxbuf[i].user_buffer;
      if (!map) {
-         if (!nv30->vtxbuf[i].buffer) {
-            continue;
-         }
-         map = pipe_buffer_map(pipe, nv30->vtxbuf[i].buffer,
+         if (nv30->vtxbuf[i].buffer)
+            map = pipe_buffer_map(pipe, nv30->vtxbuf[i].buffer,
                                  PIPE_TRANSFER_UNSYNCHRONIZED |
                                  PIPE_TRANSFER_READ, &transfer[i]);
      }
@@ -418,9 +430,9 @@ nv30_render_vbo(struct pipe_context *pipe, const struct pipe_draw_info *info)
   if (info->indexed) {
      const void *map = nv30->idxbuf.user_buffer;
      if (!map)
-         pipe_buffer_map(pipe, nv30->idxbuf.buffer,
-                                  PIPE_TRANSFER_UNSYNCHRONIZED |
-                                  PIPE_TRANSFER_READ, &transferi);
+         map = pipe_buffer_map(pipe, nv30->idxbuf.buffer,
+                               PIPE_TRANSFER_UNSYNCHRONIZED |
+                               PIPE_TRANSFER_READ, &transferi);
      draw_set_indexes(draw,
                       (ubyte *) map + nv30->idxbuf.offset,
                       nv30->idxbuf.index_size, ~0);
@@ -444,6 +456,12 @@ nv30_render_vbo(struct pipe_context *pipe, const struct pipe_draw_info *info)
 static void
 nv30_render_destroy(struct vbuf_render *render)
 {
+   struct nv30_render *r = nv30_render(render);
+
+   if (r->transfer)
+      pipe_buffer_unmap(&r->nv30->base.pipe, r->transfer);
+   pipe_resource_reference(&r->buffer, NULL);
+   nouveau_heap_free(&r->vertprog);
   FREE(render);
 }

--- a/src/gallium/drivers/nouveau/nv30/nv30_fragprog.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_fragprog.c
@@ -23,6 +23,7 @@
 *
 */

+#include "draw/draw_context.h"
 #include "tgsi/tgsi_parse.h"

 #include "nv_object.xml.h"
@@ -147,8 +148,12 @@ nv30_fp_state_delete(struct pipe_context *pipe, void *hwcso)

   pipe_resource_reference(&fp->buffer, NULL);

+   if (fp->draw)
+      draw_delete_fragment_shader(nv30_context(pipe)->draw, fp->draw);
+
   FREE((void *)fp->pipe.tokens);
   FREE(fp->insn);
+   FREE(fp->consts);
   FREE(fp);
 }

--- a/src/gallium/drivers/nouveau/nv30/nv30_state_validate.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_state_validate.c
@@ -272,15 +272,13 @@ nv30_validate_clip(struct nv30_context *nv30)
   uint32_t clpd_enable = 0;

   for (i = 0; i < 6; i++) {
-      if (nv30->rast->pipe.clip_plane_enable & (1 << i)) {
-         if (nv30->dirty & NV30_NEW_CLIP) {
-            BEGIN_NV04(push, NV30_3D(VP_UPLOAD_CONST_ID), 5);
-            PUSH_DATA (push, i);
-            PUSH_DATAp(push, nv30->clip.ucp[i], 4);
-         }
-
-         clpd_enable |= 1 << (1 + 4*i);
+      if (nv30->dirty & NV30_NEW_CLIP) {
+         BEGIN_NV04(push, NV30_3D(VP_UPLOAD_CONST_ID), 5);
+         PUSH_DATA (push, i);
+         PUSH_DATAp(push, nv30->clip.ucp[i], 4);
      }
+      if (nv30->rast->pipe.clip_plane_enable & (1 << i))
+         clpd_enable |= 2 << (4*i);
   }

   BEGIN_NV04(push, NV30_3D(VP_CLIP_PLANES_ENABLE), 1);
@@ -389,7 +387,7 @@ static struct state_validate hwtnl_validate_list[] = {
    { nv30_validate_stipple,       NV30_NEW_STIPPLE },
    { nv30_validate_scissor,       NV30_NEW_SCISSOR | NV30_NEW_RASTERIZER },
    { nv30_validate_viewport,      NV30_NEW_VIEWPORT },
-    { nv30_validate_clip,          NV30_NEW_CLIP },
+    { nv30_validate_clip,          NV30_NEW_CLIP | NV30_NEW_RASTERIZER },
    { nv30_fragprog_validate,      NV30_NEW_FRAGPROG | NV30_NEW_FRAGCONST },
    { nv30_vertprog_validate,      NV30_NEW_VERTPROG | NV30_NEW_VERTCONST |
                                   NV30_NEW_FRAGPROG | NV30_NEW_RASTERIZER },
@@ -456,7 +454,7 @@ nv30_state_context_switch(struct nv30_context *nv30)
 }

 boolean
-nv30_state_validate(struct nv30_context *nv30, boolean hwtnl)
+nv30_state_validate(struct nv30_context *nv30, uint32_t mask, boolean hwtnl)
 {
   struct nouveau_screen *screen = &nv30->screen->base;
   struct nouveau_pushbuf *push = nv30->base.pushbuf;
@@ -481,14 +479,16 @@ nv30_state_validate(struct nv30_context *nv30, boolean hwtnl)
   else
      validate = swtnl_validate_list;

-   if (nv30->dirty) {
+   mask &= nv30->dirty;
+
+   if (mask) {
      while (validate->func) {
-         if (nv30->dirty & validate->mask)
+         if (mask & validate->mask)
            validate->func(nv30);
         validate++;
      }

-      nv30->dirty = 0;
+      nv30->dirty &= ~mask;
   }

   nouveau_pushbuf_bufctx(push, bctx);
--- a/src/gallium/drivers/nouveau/nv30/nv30_vbo.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_vbo.c
@@ -564,7 +564,7 @@ nv30_draw_vbo(struct pipe_context *pipe, const struct pipe_draw_info *info)
   if (nv30->vbo_user && !(nv30->dirty & (NV30_NEW_VERTEX | NV30_NEW_ARRAYS)))
      nv30_update_user_vbufs(nv30);

-   nv30_state_validate(nv30, TRUE);
+   nv30_state_validate(nv30, ~0, TRUE);
   if (nv30->draw_flags) {
      nv30_render_vbo(pipe, info);
      return;
--- a/src/gallium/drivers/nouveau/nv30/nv30_vertprog.c
+++ b/src/gallium/drivers/nouveau/nv30/nv30_vertprog.c
@@ -23,6 +23,7 @@
 *
 */

+#include "draw/draw_context.h"
 #include "util/u_dynarray.h"
 #include "tgsi/tgsi_parse.h"

@@ -237,6 +238,10 @@ nv30_vp_state_delete(struct pipe_context *pipe, void *hwcso)

   if (vp->translated)
      nv30_vertprog_destroy(vp);
+
+   if (vp->draw)
+      draw_delete_vertex_shader(nv30_context(pipe)->draw, vp->draw);
+
   FREE((void *)vp->pipe.tokens);
   FREE(vp);
 }
--- a/src/gallium/drivers/nouveau/nv30/nvfx_vertprog.c
+++ b/src/gallium/drivers/nouveau/nv30/nvfx_vertprog.c
@@ -873,9 +873,8 @@ nvfx_vertprog_parse_decl_output(struct nvfx_vpc *vpc,
      }
      break;
   case TGSI_SEMANTIC_EDGEFLAG:
-      /* not really an error just a fallback */
-      NOUVEAU_ERR("cannot handle edgeflag output\n");
-      return FALSE;
+      vpc->r_result[idx] = nvfx_reg(NVFXSR_NONE, 0);
+      return TRUE;
   default:
      NOUVEAU_ERR("bad output semantic\n");
      return FALSE;
--- a/src/gallium/drivers/nouveau/nv50/nv50_context.c
+++ b/src/gallium/drivers/nouveau/nv50/nv50_context.c
@@ -138,8 +138,11 @@ nv50_destroy(struct pipe_context *pipe)
 {
   struct nv50_context *nv50 = nv50_context(pipe);

-   if (nv50_context_screen(nv50)->cur_ctx == nv50)
-      nv50_context_screen(nv50)->cur_ctx = NULL;
+   if (nv50->screen->cur_ctx == nv50) {
+      nv50->screen->cur_ctx = NULL;
+      /* Save off the state in case another context gets created */
+      nv50->screen->save_state = nv50->state;
+   }
   nouveau_pushbuf_bufctx(nv50->base.pushbuf, NULL);
   nouveau_pushbuf_kick(nv50->base.pushbuf, nv50->base.pushbuf->channel);

@@ -290,6 +293,10 @@ nv50_create(struct pipe_screen *pscreen, void *priv)
   pipe->get_sample_position = nv50_context_get_sample_position;

   if (!screen->cur_ctx) {
+      /* Restore the last context's state here, normally handled during
+       * context switch
+       */
+      nv50->state = screen->save_state;
      screen->cur_ctx = nv50;
      nouveau_pushbuf_bufctx(screen->base.pushbuf, nv50->bufctx);
   }
--- a/src/gallium/drivers/nouveau/nv50/nv50_context.h
+++ b/src/gallium/drivers/nouveau/nv50/nv50_context.h
@@ -104,28 +104,7 @@ struct nv50_context {
   uint32_t dirty;
   boolean cb_dirty;

-   struct {
-      uint32_t instance_elts; /* bitmask of per-instance elements */
-      uint32_t instance_base;
-      uint32_t interpolant_ctrl;
-      uint32_t semantic_color;
-      uint32_t semantic_psize;
-      int32_t index_bias;
-      boolean uniform_buffer_bound[3];
-      boolean prim_restart;
-      boolean point_sprite;
-      boolean rt_serialize;
-      boolean flushed;
-      boolean rasterizer_discard;
-      uint8_t tls_required;
-      boolean new_tls_space;
-      uint8_t num_vtxbufs;
-      uint8_t num_vtxelts;
-      uint8_t num_textures[3];
-      uint8_t num_samplers[3];
-      uint8_t prim_size;
-      uint16_t scissor;
-   } state;
+   struct nv50_graph_state state;

   struct nv50_blend_stateobj *blend;
   struct nv50_rasterizer_stateobj *rast;
@@ -191,12 +170,6 @@ nv50_context(struct pipe_context *pipe)
   return (struct nv50_context *)pipe;
 }

-static INLINE struct nv50_screen *
-nv50_context_screen(struct nv50_context *nv50)
-{
-   return nv50_screen(&nv50->base.screen->base);
-}
-
 /* return index used in nv50_context arrays for a specific shader type */
 static INLINE unsigned
 nv50_context_shader_stage(unsigned pipe)
--- a/src/gallium/drivers/nouveau/nv50/nv50_query.c
+++ b/src/gallium/drivers/nouveau/nv50/nv50_query.c
@@ -406,6 +406,7 @@ nv50_query_pushbuf_submit(struct nouveau_pushbuf *push,
   /* XXX: does this exist ? */
 #define NV50_IB_ENTRY_1_NO_PREFETCH (0 << (31 - 8))

+   PUSH_REFN(push, q->bo, NOUVEAU_BO_RD | NOUVEAU_BO_GART);
   nouveau_pushbuf_space(push, 0, 0, 1);
   nouveau_pushbuf_data(push, q->bo, q->offset + result_offset, 4 |
                        NV50_IB_ENTRY_1_NO_PREFETCH);
--- a/src/gallium/drivers/nouveau/nv50/nv50_screen.h
+++ b/src/gallium/drivers/nouveau/nv50/nv50_screen.h
@@ -25,10 +25,34 @@ struct nv50_context;

 struct nv50_blitter;

+struct nv50_graph_state {
+   uint32_t instance_elts; /* bitmask of per-instance elements */
+   uint32_t instance_base;
+   uint32_t interpolant_ctrl;
+   uint32_t semantic_color;
+   uint32_t semantic_psize;
+   int32_t index_bias;
+   boolean uniform_buffer_bound[3];
+   boolean prim_restart;
+   boolean point_sprite;
+   boolean rt_serialize;
+   boolean flushed;
+   boolean rasterizer_discard;
+   uint8_t tls_required;
+   boolean new_tls_space;
+   uint8_t num_vtxbufs;
+   uint8_t num_vtxelts;
+   uint8_t num_textures[3];
+   uint8_t num_samplers[3];
+   uint8_t prim_size;
+   uint16_t scissor;
+};
+
 struct nv50_screen {
   struct nouveau_screen base;

   struct nv50_context *cur_ctx;
+   struct nv50_graph_state save_state;

   struct nouveau_bo *code;
   struct nouveau_bo *uniforms;
--- a/src/gallium/drivers/nouveau/nv50/nv50_state.c
+++ b/src/gallium/drivers/nouveau/nv50/nv50_state.c
@@ -811,12 +811,12 @@ nv50_set_constant_buffer(struct pipe_context *pipe, uint shader, uint index,
   nv50->constbuf[s][i].user = (cb && cb->user_buffer) ? TRUE : FALSE;
   if (nv50->constbuf[s][i].user) {
      nv50->constbuf[s][i].u.data = cb->user_buffer;
-      nv50->constbuf[s][i].size = cb->buffer_size;
+      nv50->constbuf[s][i].size = MIN2(cb->buffer_size, 0x10000);
      nv50->constbuf_valid[s] |= 1 << i;
   } else
   if (res) {
      nv50->constbuf[s][i].offset = cb->buffer_offset;
-      nv50->constbuf[s][i].size = align(cb->buffer_size, 0x100);
+      nv50->constbuf[s][i].size = MIN2(align(cb->buffer_size, 0x100), 0x10000);
      nv50->constbuf_valid[s] |= 1 << i;
   } else {
      nv50->constbuf_valid[s] &= ~(1 << i);
--- a/src/gallium/drivers/nouveau/nv50/nv50_state_validate.c
+++ b/src/gallium/drivers/nouveau/nv50/nv50_state_validate.c
@@ -394,6 +394,8 @@ nv50_switch_pipe_context(struct nv50_context *ctx_to)

   if (ctx_from)
      ctx_to->state = ctx_from->state;
+   else
+      ctx_to->state = ctx_to->screen->save_state;

   ctx_to->dirty = ~0;
   ctx_to->viewports_dirty = ~0;
--- a/src/gallium/drivers/nouveau/nv50/nv50_vbo.c
+++ b/src/gallium/drivers/nouveau/nv50/nv50_vbo.c
@@ -628,6 +628,7 @@ nv50_draw_elements(struct nv50_context *nv50, boolean shorten,
         BEGIN_NV04(push, NV50_3D(VERTEX_BEGIN_GL), 1);
         PUSH_DATA (push, prim);

+         PUSH_REFN(push, buf->bo, NOUVEAU_BO_RD | buf->domain);
         nouveau_pushbuf_space(push, 8, 0, 1);

         switch (index_size) {
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_context.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_context.c
@@ -139,8 +139,12 @@ nvc0_destroy(struct pipe_context *pipe)
 {
   struct nvc0_context *nvc0 = nvc0_context(pipe);

-   if (nvc0->screen->cur_ctx == nvc0)
+   if (nvc0->screen->cur_ctx == nvc0) {
      nvc0->screen->cur_ctx = NULL;
+      nvc0->screen->save_state = nvc0->state;
+      nvc0->screen->save_state.tfb = NULL;
+   }
+
   /* Unset bufctx, we don't want to revalidate any resources after the flush.
    * Other contexts will always set their bufctx again on action calls.
    */
@@ -303,6 +307,7 @@ nvc0_create(struct pipe_screen *pscreen, void *priv)
   pipe->get_sample_position = nvc0_context_get_sample_position;

   if (!screen->cur_ctx) {
+      nvc0->state = screen->save_state;
      screen->cur_ctx = nvc0;
      nouveau_pushbuf_bufctx(screen->base.pushbuf, nvc0->bufctx);
   }
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_context.h
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_context.h
@@ -113,29 +113,7 @@ struct nvc0_context {
   uint32_t dirty;
   uint32_t dirty_cp; /* dirty flags for compute state */

-   struct {
-      boolean flushed;
-      boolean rasterizer_discard;
-      boolean early_z_forced;
-      boolean prim_restart;
-      uint32_t instance_elts; /* bitmask of per-instance elements */
-      uint32_t instance_base;
-      uint32_t constant_vbos;
-      uint32_t constant_elts;
-      int32_t index_bias;
-      uint16_t scissor;
-      uint8_t vbo_mode; /* 0 = normal, 1 = translate, 3 = translate, forced */
-      uint8_t num_vtxbufs;
-      uint8_t num_vtxelts;
-      uint8_t num_textures[6];
-      uint8_t num_samplers[6];
-      uint8_t tls_required; /* bitmask of shader types using l[] */
-      uint8_t c14_bound; /* whether immediate array constbuf is bound */
-      uint8_t clip_enable;
-      uint32_t clip_mode;
-      uint32_t uniform_buffer_bound[5];
-      struct nvc0_transform_feedback_state *tfb;
-   } state;
+   struct nvc0_graph_state state;

   struct nvc0_blend_stateobj *blend;
   struct nvc0_rasterizer_stateobj *rast;
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
@@ -392,7 +392,7 @@ nvc0_gp_gen_header(struct nvc0_program *gp, struct nv50_ir_prog_info *info)
      break;
   }

-   gp->hdr[4] = info->prop.gp.maxVertices & 0x1ff;
+   gp->hdr[4] = MIN2(info->prop.gp.maxVertices, 1024);

   return nvc0_vtgp_gen_header(gp, info);
 }
@@ -683,11 +683,12 @@ nvc0_program_upload_code(struct nvc0_context *nvc0, struct nvc0_program *prog)
   ret = nouveau_heap_alloc(screen->text_heap, size, prog, &prog->mem);
   if (ret) {
      struct nouveau_heap *heap = screen->text_heap;
-      struct nouveau_heap *iter;
-      for (iter = heap; iter && iter->next != heap; iter = iter->next) {
-         struct nvc0_program *evict = iter->priv;
-         if (evict)
-            nouveau_heap_free(&evict->mem);
+      /* Note that the code library, which is allocated before anything else,
+       * does not have a priv pointer. We can stop once we hit it.
+       */
+      while (heap->next && heap->next->priv) {
+         struct nvc0_program *evict = heap->next->priv;
+         nouveau_heap_free(&evict->mem);
      }
      debug_printf("WARNING: out of code space, evicting all shaders.\n");
      ret = nouveau_heap_alloc(heap, size, prog, &prog->mem);
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_query.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_query.c
@@ -610,6 +610,7 @@ nvc0_query_pushbuf_submit(struct nouveau_pushbuf *push,

 #define NVC0_IB_ENTRY_1_NO_PREFETCH (1 << (31 - 8))

+   PUSH_REFN(push, q->bo, NOUVEAU_BO_RD | NOUVEAU_BO_GART);
   nouveau_pushbuf_space(push, 0, 0, 1);
   nouveau_pushbuf_data(push, q->bo, q->offset + result_offset, 4 |
                        NVC0_IB_ENTRY_1_NO_PREFETCH);
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_screen.h
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_screen.h
@@ -27,10 +27,35 @@ struct nvc0_context;

 struct nvc0_blitter;

+struct nvc0_graph_state {
+   boolean flushed;
+   boolean rasterizer_discard;
+   boolean early_z_forced;
+   boolean prim_restart;
+   uint32_t instance_elts; /* bitmask of per-instance elements */
+   uint32_t instance_base;
+   uint32_t constant_vbos;
+   uint32_t constant_elts;
+   int32_t index_bias;
+   uint16_t scissor;
+   uint8_t vbo_mode; /* 0 = normal, 1 = translate, 3 = translate, forced */
+   uint8_t num_vtxbufs;
+   uint8_t num_vtxelts;
+   uint8_t num_textures[6];
+   uint8_t num_samplers[6];
+   uint8_t tls_required; /* bitmask of shader types using l[] */
+   uint8_t c14_bound; /* whether immediate array constbuf is bound */
+   uint8_t clip_enable;
+   uint32_t clip_mode;
+   uint32_t uniform_buffer_bound[5];
+   struct nvc0_transform_feedback_state *tfb;
+};
+
 struct nvc0_screen {
   struct nouveau_screen base;

   struct nvc0_context *cur_ctx;
+   struct nvc0_graph_state save_state;

   int num_occlusion_queries_active;

--- a/src/gallium/drivers/nouveau/nvc0/nvc0_shader_state.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_shader_state.c
@@ -262,11 +262,13 @@ nvc0_tfb_validate(struct nvc0_context *nvc0)
      if (tfb)
         targ->stride = tfb->stride[b];

+      buf = nv04_resource(targ->pipe.buffer);
+
+      BCTX_REFN(nvc0->bufctx_3d, TFB, buf, WR);
+
      if (!(nvc0->tfbbuf_dirty & (1 << b)))
         continue;

-      buf = nv04_resource(targ->pipe.buffer);
-
      if (!targ->clean)
         nvc0_query_fifo_wait(push, targ->pq);
      BEGIN_NVC0(push, NVC0_3D(TFB_BUFFER_ENABLE(b)), 5);
@@ -280,7 +282,6 @@ nvc0_tfb_validate(struct nvc0_context *nvc0)
         PUSH_DATA(push, 0); /* TFB_BUFFER_OFFSET */
         targ->clean = FALSE;
      }
-      BCTX_REFN(nvc0->bufctx_3d, TFB, buf, WR);
   }
   for (; b < 4; ++b)
      IMMED_NVC0(push, NVC0_3D(TFB_BUFFER_ENABLE(b)), 0);
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_state.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_state.c
@@ -811,12 +811,12 @@ nvc0_set_constant_buffer(struct pipe_context *pipe, uint shader, uint index,
   nvc0->constbuf[s][i].user = (cb && cb->user_buffer) ? TRUE : FALSE;
   if (nvc0->constbuf[s][i].user) {
      nvc0->constbuf[s][i].u.data = cb->user_buffer;
-      nvc0->constbuf[s][i].size = cb->buffer_size;
+      nvc0->constbuf[s][i].size = MIN2(cb->buffer_size, 0x10000);
      nvc0->constbuf_valid[s] |= 1 << i;
   } else
   if (cb) {
      nvc0->constbuf[s][i].offset = cb->buffer_offset;
-      nvc0->constbuf[s][i].size = align(cb->buffer_size, 0x100);
+      nvc0->constbuf[s][i].size = MIN2(align(cb->buffer_size, 0x100), 0x10000);
      nvc0->constbuf_valid[s] |= 1 << i;
   }
   else {
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c
@@ -543,6 +543,8 @@ nvc0_switch_pipe_context(struct nvc0_context *ctx_to)

   if (ctx_from)
      ctx_to->state = ctx_from->state;
+   else
+      ctx_to->state = ctx_to->screen->save_state;

   ctx_to->dirty = ~0;
   ctx_to->viewports_dirty = ~0;
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_surface.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_surface.c
@@ -1152,6 +1152,12 @@ nvc0_blit_3d(struct nvc0_context *nvc0, const struct pipe_blit_info *info)
                      NVC0_3D_VERTEX_ATTRIB_FORMAT_SIZE_32 |
                      NVC0_3D_VERTEX_ATTRIB_FORMAT_CONST);
   }
+   if (nvc0->state.instance_elts) {
+      nvc0->state.instance_elts = 0;
+      BEGIN_NVC0(push, NVC0_3D(MACRO_VERTEX_ARRAY_PER_INSTANCE), 2);
+      PUSH_DATA (push, n);
+      PUSH_DATA (push, 0);
+   }
   nvc0->state.num_vtxelts = 2;

   for (i = 0; i < info->dst.box.depth; ++i, z += dz) {
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_vbo.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_vbo.c
@@ -829,6 +829,7 @@ nvc0_draw_indirect(struct nvc0_context *nvc0, const struct pipe_draw_info *info)
   }
   PUSH_DATA(push, nvc0_prim_gl(info->mode));
 #define NVC0_IB_ENTRY_1_NO_PREFETCH (1 << (31 - 8))
+   PUSH_REFN(push, buf->bo, NOUVEAU_BO_RD | buf->domain);
   nouveau_pushbuf_space(push, 0, 0, 1);
   nouveau_pushbuf_data(push,
                        buf->bo, offset, NVC0_IB_ENTRY_1_NO_PREFETCH | size);
--- a/src/gallium/drivers/r300/Android.mk
+++ b/src/gallium/drivers/r300/Android.mk
@@ -33,9 +33,10 @@ LOCAL_SRC_FILES := $(C_SOURCES)
 LOCAL_C_INCLUDES := \
 	$(MESA_TOP)/src/mapi \
 	$(MESA_TOP)/src/glsl \
-	$(MESA_TOP)/src/mesa \
-	$(TARGET_OUT_HEADERS)/libdrm
+	$(MESA_TOP)/src/mesa

+
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_radeon
 LOCAL_MODULE := libmesa_pipe_r300

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/drivers/r300/Automake.inc
+++ b/src/gallium/drivers/r300/Automake.inc
@@ -5,7 +5,7 @@ TARGET_CPPFLAGS += -DGALLIUM_R300
 TARGET_LIB_DEPS += \
 	$(top_builddir)/src/gallium/drivers/r300/libr300.la \
 	$(RADEON_LIBS) \
-	$(INTEL_LIBS)
+	$(LIBDRM_LIBS)

 TARGET_RADEON_WINSYS = \
 	$(top_builddir)/src/gallium/winsys/radeon/drm/libradeonwinsys.la
--- a/src/gallium/drivers/r600/Android.mk
+++ b/src/gallium/drivers/r600/Android.mk
@@ -30,8 +30,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES) $(CXX_SOURCES)

-LOCAL_C_INCLUDES := $(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_radeon
 LOCAL_MODULE := libmesa_pipe_r600

 include external/stlport/libstlport.mk
--- a/src/gallium/drivers/radeon/Android.mk
+++ b/src/gallium/drivers/radeon/Android.mk
@@ -30,8 +30,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES)

-LOCAL_C_INCLUDES := $(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_radeon
 LOCAL_MODULE := libmesa_pipe_radeon

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/drivers/radeonsi/Android.mk
+++ b/src/gallium/drivers/radeonsi/Android.mk
@@ -30,8 +30,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES)

-LOCAL_C_INCLUDES := $(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_radeon
 LOCAL_MODULE := libmesa_pipe_radeonsi

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/drivers/radeonsi/si_shader.c
+++ b/src/gallium/drivers/radeonsi/si_shader.c
@@ -2741,43 +2741,6 @@ static int si_generate_gs_copy_shader(struct si_screen *sscreen,
 	return r;
 }

-static void si_dump_key(unsigned shader, union si_shader_key *key)
-{
-	int i;
-
-	fprintf(stderr, "SHADER KEY\n");
-
-	switch (shader) {
-	case PIPE_SHADER_VERTEX:
-		fprintf(stderr, "  instance_divisors = {");
-		for (i = 0; i < Elements(key->vs.instance_divisors); i++)
-			fprintf(stderr, !i ? "%u" : ", %u",
-				key->vs.instance_divisors[i]);
-		fprintf(stderr, "}\n");
-
-		if (key->vs.as_es)
-			fprintf(stderr, "  gs_used_inputs = 0x%"PRIx64"\n",
-				key->vs.gs_used_inputs);
-		fprintf(stderr, "  as_es = %u\n", key->vs.as_es);
-		break;
-
-	case PIPE_SHADER_GEOMETRY:
-		break;
-
-	case PIPE_SHADER_FRAGMENT:
-		fprintf(stderr, "  export_16bpc = 0x%X\n", key->ps.export_16bpc);
-		fprintf(stderr, "  last_cbuf = %u\n", key->ps.last_cbuf);
-		fprintf(stderr, "  color_two_side = %u\n", key->ps.color_two_side);
-		fprintf(stderr, "  alpha_func = %u\n", key->ps.alpha_func);
-		fprintf(stderr, "  alpha_to_one = %u\n", key->ps.alpha_to_one);
-		fprintf(stderr, "  poly_stipple = %u\n", key->ps.poly_stipple);
-		break;
-
-	default:
-		assert(0);
-	}
-}
-
 int si_shader_create(struct si_screen *sscreen, LLVMTargetMachineRef tm,
 		     struct si_shader *shader)
 {
--- a/src/gallium/state_trackers/clover/core/event.cpp
+++ b/src/gallium/state_trackers/clover/core/event.cpp
@@ -27,7 +27,7 @@ using namespace clover;

 event::event(clover::context &ctx, const ref_vector<event> &deps,
             action action_ok, action action_fail) :
-   context(ctx), _status(0), wait_count(1),
+   context(ctx), wait_count(1), _status(0),
   action_ok(action_ok), action_fail(action_fail) {
   for (auto &ev : deps)
      ev.chain(*this);
@@ -36,36 +36,69 @@ event::event(clover::context &ctx, const ref_vector<event> &deps,
 event::~event() {
 }

+std::vector<intrusive_ref<event>>
+event::trigger_self() {
+   std::lock_guard<std::mutex> lock(mutex);
+   std::vector<intrusive_ref<event>> evs;
+
+   if (!--wait_count)
+      std::swap(_chain, evs);
+
+   return evs;
+}
+
 void
 event::trigger() {
-   if (!--wait_count) {
-      action_ok(*this);
+   auto evs = trigger_self();

-      while (!_chain.empty()) {
-         _chain.back()().trigger();
-         _chain.pop_back();
-      }
+   if (signalled()) {
+      action_ok(*this);
+      cv.notify_all();
   }
+
+   for (event &ev : evs)
+      ev.trigger();
+}
+
+std::vector<intrusive_ref<event>>
+event::abort_self(cl_int status) {
+   std::lock_guard<std::mutex> lock(mutex);
+   std::vector<intrusive_ref<event>> evs;
+
+   _status = status;
+   std::swap(_chain, evs);
+
+   return evs;
 }

 void
 event::abort(cl_int status) {
-   _status = status;
+   auto evs = abort_self(status);
+
   action_fail(*this);

-   while (!_chain.empty()) {
-      _chain.back()().abort(status);
-      _chain.pop_back();
-   }
+   for (event &ev : evs)
+      ev.abort(status);
 }

 bool
 event::signalled() const {
+   std::lock_guard<std::mutex> lock(mutex);
   return !wait_count;
 }

+cl_int
+event::status() const {
+   std::lock_guard<std::mutex> lock(mutex);
+   return _status;
+}
+
 void
 event::chain(event &ev) {
+   std::unique_lock<std::mutex> lock(mutex, std::defer_lock);
+   std::unique_lock<std::mutex> lock_ev(ev.mutex, std::defer_lock);
+   std::lock(lock, lock_ev);
+
   if (wait_count) {
      ev.wait_count++;
      _chain.push_back(ev);
@@ -73,6 +106,15 @@ event::chain(event &ev) {
   ev.deps.push_back(*this);
 }

+void
+event::wait() const {
+   for (event &ev : deps)
+      ev.wait();
+
+   std::unique_lock<std::mutex> lock(mutex);
+   cv.wait(lock, [=]{ return !wait_count; });
+}
+
 hard_event::hard_event(command_queue &q, cl_command_type command,
                       const ref_vector<event> &deps, action action) :
   event(q.context(), deps, profile(q, action), [](event &ev){}),
@@ -93,8 +135,8 @@ cl_int
 hard_event::status() const {
   pipe_screen *screen = queue()->device().pipe;

-   if (_status < 0)
-      return _status;
+   if (event::status() < 0)
+      return event::status();

   else if (!_fence)
      return CL_QUEUED;
@@ -120,6 +162,8 @@ void
 hard_event::wait() const {
   pipe_screen *screen = queue()->device().pipe;

+   event::wait();
+
   if (status() == CL_QUEUED)
      queue()->flush();

@@ -182,8 +226,8 @@ soft_event::soft_event(clover::context &ctx, const ref_vector<event> &deps,

 cl_int
 soft_event::status() const {
-   if (_status < 0)
-      return _status;
+   if (event::status() < 0)
+      return event::status();

   else if (!signalled() ||
            any_of([](const event &ev) {
@@ -207,8 +251,7 @@ soft_event::command() const {

 void
 soft_event::wait() const {
-   for (event &ev : deps)
-      ev.wait();
+   event::wait();

   if (status() != CL_COMPLETE)
      throw error(CL_EXEC_STATUS_ERROR_FOR_EVENTS_IN_WAIT_LIST);
--- a/src/gallium/state_trackers/clover/core/event.hpp
+++ b/src/gallium/state_trackers/clover/core/event.hpp
@@ -23,6 +23,7 @@
 #ifndef CLOVER_CORE_EVENT_HPP
 #define CLOVER_CORE_EVENT_HPP

+#include <condition_variable>
 #include <functional>

 #include "core/object.hpp"
@@ -65,24 +66,29 @@ namespace clover {
      void abort(cl_int status);
      bool signalled() const;

-      virtual cl_int status() const = 0;
+      virtual cl_int status() const;
      virtual command_queue *queue() const = 0;
      virtual cl_command_type command() const = 0;
-      virtual void wait() const = 0;
+      virtual void wait() const;

      const intrusive_ref<clover::context> context;

   protected:
      void chain(event &ev);

-      cl_int _status;
      std::vector<intrusive_ref<event>> deps;

   private:
+      std::vector<intrusive_ref<event>> trigger_self();
+      std::vector<intrusive_ref<event>> abort_self(cl_int status);
+
      unsigned wait_count;
+      cl_int _status;
      action action_ok;
      action action_fail;
      std::vector<intrusive_ref<event>> _chain;
+      mutable std::condition_variable cv;
+      mutable std::mutex mutex;
   };

   ///
--- a/src/gallium/state_trackers/clover/core/queue.cpp
+++ b/src/gallium/state_trackers/clover/core/queue.cpp
@@ -44,6 +44,7 @@ command_queue::flush() {
   pipe_screen *screen = device().pipe;
   pipe_fence_handle *fence = NULL;

+   std::lock_guard<std::mutex> lock(queued_events_mutex);
   if (!queued_events.empty()) {
      pipe->flush(pipe, &fence, 0);

@@ -69,6 +70,7 @@ command_queue::profiling_enabled() const {

 void
 command_queue::sequence(hard_event &ev) {
+   std::lock_guard<std::mutex> lock(queued_events_mutex);
   if (!queued_events.empty())
      queued_events.back()().chain(ev);

--- a/src/gallium/state_trackers/clover/core/queue.hpp
+++ b/src/gallium/state_trackers/clover/core/queue.hpp
@@ -24,6 +24,7 @@
 #define CLOVER_CORE_QUEUE_HPP

 #include <deque>
+#include <mutex>

 #include "core/object.hpp"
 #include "core/context.hpp"
@@ -69,6 +70,7 @@ namespace clover {

      cl_command_queue_properties props;
      pipe_context *pipe;
+      std::mutex queued_events_mutex;
      std::deque<intrusive_ref<hard_event>> queued_events;
   };
 }
--- a/src/gallium/state_trackers/dri/dri_context.c
+++ b/src/gallium/state_trackers/dri/dri_context.c
@@ -233,11 +233,10 @@ dri_make_current(__DRIcontext * cPriv,

   ctx->stapi->make_current(ctx->stapi, ctx->st, &draw->base, &read->base);

-   // This is ok to call here. If they are already init, it's a no-op.
-   if (draw->textures[ST_ATTACHMENT_BACK_LEFT] && draw->textures[ST_ATTACHMENT_DEPTH_STENCIL]
-      && ctx->pp)
-         pp_init_fbos(ctx->pp, draw->textures[ST_ATTACHMENT_BACK_LEFT]->width0,
-            draw->textures[ST_ATTACHMENT_BACK_LEFT]->height0);
+   /* This is ok to call here. If they are already init, it's a no-op. */
+   if (ctx->pp && draw->textures[ST_ATTACHMENT_BACK_LEFT])
+      pp_init_fbos(ctx->pp, draw->textures[ST_ATTACHMENT_BACK_LEFT]->width0,
+                   draw->textures[ST_ATTACHMENT_BACK_LEFT]->height0);

   return GL_TRUE;
 }
--- a/src/gallium/targets/dri/Makefile.am
+++ b/src/gallium/targets/dri/Makefile.am
@@ -55,12 +55,6 @@ gallium_dri_la_LIBADD = \
 	$(LIBDRM_LIBS) \
 	$(GALLIUM_COMMON_LIB_DEPS)

-# XXX: Temporary allow duplicated symbols, as the loader pulls in xmlconfig.c
-# which already provides driParse* and driQuery* amongst others.
-# Remove this hack as we come up with a cleaner solution.
-gallium_dri_la_LDFLAGS += \
-	-Wl,--allow-multiple-definition
-
 EXTRA_gallium_dri_la_DEPENDENCIES = \
 	dri.sym \
 	$(top_srcdir)/src/gallium/targets/dri-vdpau.dyn
--- a/src/gallium/targets/osmesa/Makefile.am
+++ b/src/gallium/targets/osmesa/Makefile.am
@@ -42,7 +42,6 @@ nodist_EXTRA_lib@OSMESA_LIB@_la_SOURCES = dummy.cpp
 lib@OSMESA_LIB@_la_SOURCES = target.c

 lib@OSMESA_LIB@_la_LDFLAGS = \
-	-module \
 	-no-undefined \
 	-version-number @OSMESA_VERSION@ \
 	$(GC_SECTIONS) \
--- a/src/gallium/winsys/freedreno/drm/Android.mk
+++ b/src/gallium/winsys/freedreno/drm/Android.mk
@@ -27,10 +27,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES)

-LOCAL_C_INCLUDES := \
-	$(TARGET_OUT_HEADERS)/libdrm \
-	$(TARGET_OUT_HEADERS)/freedreno
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_freedreno
 LOCAL_MODULE := libmesa_winsys_freedreno

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/winsys/i915/drm/Android.mk
+++ b/src/gallium/winsys/i915/drm/Android.mk
@@ -30,8 +30,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES)

-LOCAL_C_INCLUDES := $(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_intel
 LOCAL_MODULE := libmesa_winsys_i915

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/winsys/intel/drm/Android.mk
+++ b/src/gallium/winsys/intel/drm/Android.mk
@@ -29,8 +29,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES)

-LOCAL_C_INCLUDES := $(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_intel
 LOCAL_MODULE := libmesa_winsys_intel

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/winsys/nouveau/drm/Android.mk
+++ b/src/gallium/winsys/nouveau/drm/Android.mk
@@ -30,8 +30,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES)

-LOCAL_C_INCLUDES := $(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_nouveau
 LOCAL_MODULE := libmesa_winsys_nouveau

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/winsys/nouveau/drm/nouveau_drm_winsys.c
+++ b/src/gallium/winsys/nouveau/drm/nouveau_drm_winsys.c
@@ -120,7 +120,11 @@ nouveau_drm_screen_create(int fd)
 	if (!screen)
 		goto err;

-	util_hash_table_set(fd_tab, intptr_to_pointer(fd), screen);
+	/* Use dupfd in hash table, to avoid errors if the original fd gets
+	 * closed by its owner. The hash key needs to live at least as long as
+	 * the screen.
+	 */
+	util_hash_table_set(fd_tab, intptr_to_pointer(dupfd), screen);
 	screen->refcount = 1;
 	pipe_mutex_unlock(nouveau_screen_mutex);
 	return &screen->base;
--- a/src/gallium/winsys/radeon/drm/Android.mk
+++ b/src/gallium/winsys/radeon/drm/Android.mk
@@ -30,8 +30,7 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := $(C_SOURCES)

-LOCAL_C_INCLUDES := $(TARGET_OUT_HEADERS)/libdrm
-
+LOCAL_SHARED_LIBRARIES := libdrm libdrm_radeon
 LOCAL_MODULE := libmesa_winsys_radeon

 include $(GALLIUM_COMMON_MK)
--- a/src/gallium/winsys/radeon/drm/radeon_drm_bo.c
+++ b/src/gallium/winsys/radeon/drm/radeon_drm_bo.c
@@ -305,14 +305,34 @@ static void radeon_bo_destroy(struct pb_buffer *_buf)
    if (bo->ptr)
        os_munmap(bo->ptr, bo->base.size);

+    if (mgr->va) {
+        if (bo->rws->va_unmap_working) {
+            struct drm_radeon_gem_va va;
+
+            va.handle = bo->handle;
+            va.vm_id = 0;
+            va.operation = RADEON_VA_UNMAP;
+            va.flags = RADEON_VM_PAGE_READABLE |
+                       RADEON_VM_PAGE_WRITEABLE |
+                       RADEON_VM_PAGE_SNOOPED;
+            va.offset = bo->va;
+
+            if (drmCommandWriteRead(bo->rws->fd, DRM_RADEON_GEM_VA, &va,
+				    sizeof(va)) != 0 &&
+		va.operation == RADEON_VA_RESULT_ERROR) {
+                fprintf(stderr, "radeon: Failed to deallocate virtual address for buffer:\n");
+                fprintf(stderr, "radeon:    size      : %d bytes\n", bo->base.size);
+                fprintf(stderr, "radeon:    va        : 0x%016llx\n", (unsigned long long)bo->va);
+            }
+	}
+
+	radeon_bomgr_free_va(mgr, bo->va, bo->base.size);
+    }
+
    /* Close object. */
    args.handle = bo->handle;
    drmIoctl(bo->rws->fd, DRM_IOCTL_GEM_CLOSE, &args);

-    if (mgr->va) {
-        radeon_bomgr_free_va(mgr, bo->va, bo->base.size);
-    }
-
    pipe_mutex_destroy(bo->map_mutex);

    if (bo->initial_domain & RADEON_DOMAIN_VRAM)
--- a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
+++ b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.c
@@ -49,6 +49,8 @@
 #define RADEON_INFO_ACTIVE_CU_COUNT 0x20
 #endif

+#define RADEON_INFO_VA_UNMAP_WORKING	0x25
+
 static struct util_hash_table *fd_tab = NULL;
 pipe_static_mutex(fd_tab_mutex);

@@ -376,6 +378,8 @@ static boolean do_winsys_init(struct radeon_drm_winsys *ws)
            if (!radeon_get_drm_value(ws->fd, RADEON_INFO_IB_VM_MAX_SIZE, NULL,
                                      &ib_vm_max_size))
                ws->info.r600_virtual_address = FALSE;
+            radeon_get_drm_value(ws->fd, RADEON_INFO_VA_UNMAP_WORKING, NULL,
+                                 &ws->va_unmap_working);
        }
 	if (ws->gen == DRV_R600 && !debug_get_bool_option("RADEON_VA", FALSE))
 		ws->info.r600_virtual_address = FALSE;
@@ -461,6 +465,10 @@ static void radeon_winsys_destroy(struct radeon_winsys *rws)
    if (ws->gen >= DRV_R600) {
        radeon_surface_manager_free(ws->surf_man);
    }
+
+    if (ws->fd)
+        close(ws->fd);
+
    FREE(rws);
 }

@@ -662,7 +670,7 @@ radeon_drm_winsys_create(int fd, radeon_screen_create_t screen_create)
        return NULL;
    }

-    ws->fd = fd;
+    ws->fd = dup(fd);

    if (!do_winsys_init(ws))
        goto fail;
@@ -678,7 +686,7 @@ radeon_drm_winsys_create(int fd, radeon_screen_create_t screen_create)
        goto fail;

    if (ws->gen >= DRV_R600) {
-        ws->surf_man = radeon_surface_manager_new(fd);
+        ws->surf_man = radeon_surface_manager_new(ws->fd);
        if (!ws->surf_man)
            goto fail;
    }
@@ -719,7 +727,7 @@ radeon_drm_winsys_create(int fd, radeon_screen_create_t screen_create)
        return NULL;
    }

-    util_hash_table_set(fd_tab, intptr_to_pointer(fd), ws);
+    util_hash_table_set(fd_tab, intptr_to_pointer(ws->fd), ws);

    /* We must unlock the mutex once the winsys is fully initialized, so that
     * other threads attempting to create the winsys from the same fd will
@@ -736,6 +744,9 @@ fail:
        ws->kman->destroy(ws->kman);
    if (ws->surf_man)
        radeon_surface_manager_free(ws->surf_man);
+    if (ws->fd)
+        close(ws->fd);
+
    FREE(ws);
    return NULL;
 }
--- a/src/gallium/winsys/radeon/drm/radeon_drm_winsys.h
+++ b/src/gallium/winsys/radeon/drm/radeon_drm_winsys.h
@@ -55,6 +55,7 @@ struct radeon_drm_winsys {
    enum radeon_generation gen;
    struct radeon_info info;
    uint32_t va_start;
+    uint32_t va_unmap_working;
    uint32_t accel_working2;

    struct pb_manager *kman;
--- a/src/gallium/winsys/svga/drm/Android.mk
+++ b/src/gallium/winsys/svga/drm/Android.mk
@@ -34,9 +34,9 @@ LOCAL_CFLAGS := -D_FILE_OFFSET_BITS=64

 LOCAL_C_INCLUDES := \
 	$(GALLIUM_TOP)/drivers/svga \
-	$(GALLIUM_TOP)/drivers/svga/include \
-	$(TARGET_OUT_HEADERS)/libdrm
+	$(GALLIUM_TOP)/drivers/svga/include

+LOCAL_SHARED_LIBRARIES := libdrm
 LOCAL_MODULE := libmesa_winsys_svga

 include $(GALLIUM_COMMON_MK)
--- a/src/gbm/Makefile.am
+++ b/src/gbm/Makefile.am
@@ -52,7 +52,8 @@ libgbm_dri_la_CFLAGS = \
 	$(LIBDRM_CFLAGS)

 libgbm_la_LIBADD += \
-	libgbm_dri.la $(top_builddir)/src/mapi/shared-glapi/libglapi.la $(LIBDRM_LIBS)
+	libgbm_dri.la \
+	$(LIBDRM_LIBS)
 endif

 TESTS = gbm-symbols-check
--- a/src/gbm/backends/dri/gbm_dri.c
+++ b/src/gbm/backends/dri/gbm_dri.c
@@ -311,6 +311,14 @@ dri_open_driver(struct gbm_dri_device *dri)
   if (search_paths == NULL)
      search_paths = DEFAULT_DRIVER_DIR;

+   /* Temporarily work around dri driver libs that need symbols in libglapi
+    * but don't automatically link it in.
+    */
+   /* XXX: Library name differs on per platforms basis. Update this as
+    * osx/cygwin/windows/bsd gets support for GBM..
+    */
+   dlopen("libglapi.so.0", RTLD_LAZY | RTLD_GLOBAL);
+
   dri->driver = NULL;
   end = search_paths + strlen(search_paths);
   for (p = search_paths; p < end && dri->driver == NULL; p = next + 1) {
--- a/src/glsl/Android.gen.mk
+++ b/src/glsl/Android.gen.mk
@@ -33,11 +33,23 @@ sources := \
 	glsl_lexer.cpp \
 	glsl_parser.cpp \
 	glcpp/glcpp-lex.c \
-	glcpp/glcpp-parse.c
+	glcpp/glcpp-parse.c \
+	nir/nir_constant_expressions.c \
+	nir/nir_opcodes.c \
+	nir/nir_opcodes.h \
+	nir/nir_opt_algebraic.c

 LOCAL_SRC_FILES := $(filter-out $(sources), $(LOCAL_SRC_FILES))

-LOCAL_C_INCLUDES += $(intermediates) $(intermediates)/glcpp $(MESA_TOP)/src/glsl/glcpp
+LOCAL_C_INCLUDES += \
+	$(intermediates) \
+	$(intermediates)/glcpp \
+	$(intermediates)/nir \
+	$(MESA_TOP)/src/glsl/glcpp \
+	$(MESA_TOP)/src/glsl/nir
+
+LOCAL_EXPORT_C_INCLUDE_DIRS += \
+	$(intermediates)/nir

 sources := $(addprefix $(intermediates)/, $(sources))
 LOCAL_GENERATED_SOURCES += $(sources)
@@ -77,3 +89,42 @@ $(intermediates)/glcpp/glcpp-lex.c: $(LOCAL_PATH)/glcpp/glcpp-lex.l

 $(intermediates)/glcpp/glcpp-parse.c: $(LOCAL_PATH)/glcpp/glcpp-parse.y
 	$(call glsl_local-y-to-c-and-h)
+
+nir_constant_expressions_gen := $(LOCAL_PATH)/nir/nir_constant_expressions.py
+nir_constant_expressions_deps := \
+	$(LOCAL_PATH)/nir/nir_opcodes.py \
+	$(LOCAL_PATH)/nir/nir_constant_expressions.py \
+	$(LOCAL_PATH)/nir/nir_constant_expressions.h
+
+$(intermediates)/nir/nir_constant_expressions.c: $(nir_constant_expressions_deps)
+	@mkdir -p $(dir $@)
+	$(hide) $(MESA_PYTHON2) $(nir_constant_expressions_gen) $< > $@
+
+nir_opcodes_h_gen := $(LOCAL_PATH)/nir/nir_opcodes_h.py
+nir_opcodes_h_deps := \
+	$(LOCAL_PATH)/nir/nir_opcodes.py \
+	$(LOCAL_PATH)/nir/nir_opcodes_h.py
+
+$(intermediates)/nir/nir_opcodes.h: $(nir_opcodes_h_deps)
+	@mkdir -p $(dir $@)
+	$(hide) $(MESA_PYTHON2) $(nir_opcodes_h_gen) $< > $@
+
+$(LOCAL_PATH)/nir/nir.h: $(intermediates)/nir/nir_opcodes.h
+
+nir_opcodes_c_gen := $(LOCAL_PATH)/nir/nir_opcodes_c.py
+nir_opcodes_c_deps := \
+	$(LOCAL_PATH)/nir/nir_opcodes.py \
+	$(LOCAL_PATH)/nir/nir_opcodes_c.py
+
+$(intermediates)/nir/nir_opcodes.c: $(nir_opcodes_c_deps)
+	@mkdir -p $(dir $@)
+	$(hide) $(MESA_PYTHON2) $(nir_opcodes_c_gen) $< > $@
+
+nir_opt_algebraic_gen := $(LOCAL_PATH)/nir/nir_opt_algebraic.py
+nir_opt_algebraic_deps := \
+	$(LOCAL_PATH)/nir/nir_opt_algebraic.py \
+	$(LOCAL_PATH)/nir/nir_algebraic.py
+
+$(intermediates)/nir/nir_opt_algebraic.c: $(nir_opt_algebraic_deps)
+	@mkdir -p $(dir $@)
+	$(hide) $(MESA_PYTHON2) $(nir_opt_algebraic_gen) $< > $@
--- a/src/glsl/Android.mk
+++ b/src/glsl/Android.mk
@@ -35,10 +35,10 @@ include $(CLEAR_VARS)

 LOCAL_SRC_FILES := \
 	$(LIBGLCPP_FILES) \
-	$(LIBGLSL_FILES)
+	$(LIBGLSL_FILES) \
+	$(NIR_FILES)

 LOCAL_C_INCLUDES := \
-	$(MESA_TOP)/src \
 	$(MESA_TOP)/src/mapi \
 	$(MESA_TOP)/src/mesa

@@ -59,7 +59,6 @@ LOCAL_SRC_FILES := \
 	$(GLSL_COMPILER_CXX_FILES)

 LOCAL_C_INCLUDES := \
-	$(MESA_TOP)/src \
 	$(MESA_TOP)/src/mapi \
 	$(MESA_TOP)/src/mesa

--- a/src/glsl/ast_array_index.cpp
+++ b/src/glsl/ast_array_index.cpp
@@ -228,24 +228,26 @@ _mesa_ast_array_index_to_hir(void *mem_ctx,
       * dynamically uniform expression is undefined.
       */
      if (array->type->element_type()->is_sampler()) {
-	 if (!state->is_version(130, 100)) {
-	    if (state->es_shader) {
-	       _mesa_glsl_warning(&loc, state,
-				  "sampler arrays indexed with non-constant "
-				  "expressions is optional in %s",
-				  state->get_version_string());
-	    } else {
-	       _mesa_glsl_warning(&loc, state,
-				  "sampler arrays indexed with non-constant "
-				  "expressions will be forbidden in GLSL 1.30 "
-				  "and later");
-	    }
-	 } else if (!state->is_version(400, 0) && !state->ARB_gpu_shader5_enable) {
-	    _mesa_glsl_error(&loc, state,
-			     "sampler arrays indexed with non-constant "
-			     "expressions is forbidden in GLSL 1.30 and "
-			     "later");
-	 }
+         if (!state->is_version(400, 0) && !state->ARB_gpu_shader5_enable) {
+            if (!state->is_version(130, 100)) {
+               if (state->es_shader) {
+                  _mesa_glsl_warning(&loc, state,
+                                     "sampler arrays indexed with non-constant "
+                                     "expressions is optional in %s",
+                                     state->get_version_string());
+               } else {
+                  _mesa_glsl_warning(&loc, state,
+                                    "sampler arrays indexed with non-constant "
+                                    "expressions will be forbidden in GLSL 1.30 "
+                                    "and later");
+               }
+            } else if (!state->is_version(400, 0) && !state->ARB_gpu_shader5_enable) {
+               _mesa_glsl_error(&loc, state,
+                                "sampler arrays indexed with non-constant "
+                                "expressions is forbidden in GLSL 1.30 and "
+                                "later");
+            }
+         }
      }
   }

--- a/src/glsl/ast_to_hir.cpp
+++ b/src/glsl/ast_to_hir.cpp
@@ -2108,7 +2108,7 @@ validate_binding_qualifier(struct _mesa_glsl_parse_state *state,
       *  with an array of size N, all elements of the array from binding
       *  through binding + N - 1 must be within this range."
       */
-      unsigned limit = ctx->Const.Program[state->stage].MaxTextureImageUnits;
+      unsigned limit = ctx->Const.MaxCombinedTextureImageUnits;

      if (max_index >= limit) {
         _mesa_glsl_error(loc, state, "layout(binding = %d) for %d samplers "
--- a/src/glsl/builtin_variables.cpp
+++ b/src/glsl/builtin_variables.cpp
@@ -849,9 +849,9 @@ void
 builtin_variable_generator::generate_gs_special_vars()
 {
   add_output(VARYING_SLOT_LAYER, int_t, "gl_Layer");
-   if (state->ARB_viewport_array_enable)
+   if (state->is_version(410, 0) || state->ARB_viewport_array_enable)
      add_output(VARYING_SLOT_VIEWPORT, int_t, "gl_ViewportIndex");
-   if (state->ARB_gpu_shader5_enable)
+   if (state->is_version(400, 0) || state->ARB_gpu_shader5_enable)
      add_system_value(SYSTEM_VALUE_INVOCATION_ID, int_t, "gl_InvocationID");

   /* Although gl_PrimitiveID appears in tessellation control and tessellation
@@ -919,7 +919,7 @@ builtin_variable_generator::generate_fs_special_vars()
         var->enable_extension_warning("GL_AMD_shader_stencil_export");
   }

-   if (state->ARB_sample_shading_enable) {
+   if (state->is_version(400, 0) || state->ARB_sample_shading_enable) {
      add_system_value(SYSTEM_VALUE_SAMPLE_ID, int_t, "gl_SampleID");
      add_system_value(SYSTEM_VALUE_SAMPLE_POS, vec2_t, "gl_SamplePosition");
      /* From the ARB_sample_shading specification:
@@ -932,11 +932,11 @@ builtin_variable_generator::generate_fs_special_vars()
      add_output(FRAG_RESULT_SAMPLE_MASK, array(int_t, 1), "gl_SampleMask");
   }

-   if (state->ARB_gpu_shader5_enable) {
+   if (state->is_version(400, 0) || state->ARB_gpu_shader5_enable) {
      add_system_value(SYSTEM_VALUE_SAMPLE_MASK_IN, array(int_t, 1), "gl_SampleMaskIn");
   }

-   if (state->ARB_fragment_layer_viewport_enable) {
+   if (state->is_version(430, 0) || state->ARB_fragment_layer_viewport_enable) {
      add_input(VARYING_SLOT_LAYER, int_t, "gl_Layer");
      add_input(VARYING_SLOT_VIEWPORT, int_t, "gl_ViewportIndex");
   }
--- a/src/glsl/glsl_types.cpp
+++ b/src/glsl/glsl_types.cpp
@@ -690,24 +690,27 @@ glsl_type::record_key_compare(const void *a, const void *b)
 }


+/**
+ * Generate an integer hash value for a glsl_type structure type.
+ */
 unsigned
 glsl_type::record_key_hash(const void *a)
 {
   const glsl_type *const key = (glsl_type *) a;
-   char hash_key[128];
-   unsigned size = 0;
-
-   size = snprintf(hash_key, sizeof(hash_key), "%08x", key->length);
+   uintptr_t hash = key->length;
+   unsigned retval;

   for (unsigned i = 0; i < key->length; i++) {
-      if (size >= sizeof(hash_key))
-	 break;
-
-      size += snprintf(& hash_key[size], sizeof(hash_key) - size,
-		       "%p", (void *) key->fields.structure[i].type);
+      /* casting pointer to uintptr_t */
+      hash = (hash * 13 ) + (uintptr_t) key->fields.structure[i].type;
   }

-   return hash_table_string_hash(& hash_key);
+   if (sizeof(hash) == 8)
+      retval = (hash & 0xffffffff) ^ ((uint64_t) hash >> 32);
+   else
+      retval = hash;
+
+   return retval;
 }


--- a/src/glsl/linker.cpp
+++ b/src/glsl/linker.cpp
@@ -345,6 +345,39 @@ private:
   bool uses_non_zero_stream;
 };

+/* Class that finds array derefs and check if indexes are dynamic. */
+class dynamic_sampler_array_indexing_visitor : public ir_hierarchical_visitor
+{
+public:
+   dynamic_sampler_array_indexing_visitor() :
+      dynamic_sampler_array_indexing(false)
+   {
+   }
+
+   ir_visitor_status visit_enter(ir_dereference_array *ir)
+   {
+      if (!ir->variable_referenced())
+         return visit_continue;
+
+      if (!ir->variable_referenced()->type->contains_sampler())
+         return visit_continue;
+
+      if (!ir->array_index->constant_expression_value()) {
+         dynamic_sampler_array_indexing = true;
+         return visit_stop;
+      }
+      return visit_continue;
+   }
+
+   bool uses_dynamic_sampler_array_indexing()
+   {
+      return dynamic_sampler_array_indexing;
+   }
+
+private:
+   bool dynamic_sampler_array_indexing;
+};
+
 } /* anonymous namespace */

 void
@@ -2479,6 +2512,41 @@ check_explicit_uniform_locations(struct gl_context *ctx,
   delete uniform_map;
 }

+/**
+ * This check is done to make sure we allow only constant expression
+ * indexing and "constant-index-expression" (indexing with an expression
+ * that includes loop induction variable).
+ */
+static bool
+validate_sampler_array_indexing(struct gl_context *ctx,
+                                struct gl_shader_program *prog)
+{
+   dynamic_sampler_array_indexing_visitor v;
+   for (unsigned i = 0; i < MESA_SHADER_STAGES; i++) {
+      if (prog->_LinkedShaders[i] == NULL)
+	 continue;
+
+      bool no_dynamic_indexing =
+         ctx->Const.ShaderCompilerOptions[i].EmitNoIndirectSampler;
+
+      /* Search for array derefs in shader. */
+      v.run(prog->_LinkedShaders[i]->ir);
+      if (v.uses_dynamic_sampler_array_indexing()) {
+         const char *msg = "sampler arrays indexed with non-constant "
+                           "expressions is forbidden in GLSL %s %u";
+         /* Backend has indicated that it has no dynamic indexing support. */
+         if (no_dynamic_indexing) {
+            linker_error(prog, msg, prog->IsES ? "ES" : "", prog->Version);
+            return false;
+         } else {
+            linker_warning(prog, msg, prog->IsES ? "ES" : "", prog->Version);
+         }
+      }
+   }
+   return true;
+}
+
+
 void
 link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
 {
@@ -2570,8 +2638,11 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
            link_intrastage_shaders(mem_ctx, ctx, prog, shader_list[stage],
                                    num_shaders[stage]);

-         if (!prog->LinkStatus)
+         if (!prog->LinkStatus) {
+            if (sh)
+               ctx->Driver.DeleteShader(ctx, sh);
            goto done;
+         }

         switch (stage) {
         case MESA_SHADER_VERTEX:
@@ -2584,8 +2655,11 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
            validate_fragment_shader_executable(prog, sh);
            break;
         }
-         if (!prog->LinkStatus)
+         if (!prog->LinkStatus) {
+            if (sh)
+               ctx->Driver.DeleteShader(ctx, sh);
            goto done;
+         }

         _mesa_reference_shader(ctx, &prog->_LinkedShaders[stage], sh);
      }
@@ -2689,6 +2763,16 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog)
      lower_const_arrays_to_uniforms(prog->_LinkedShaders[i]->ir);
   }

+   /* Validation for special cases where we allow sampler array indexing
+    * with loop induction variable. This check emits a warning or error
+    * depending if backend can handle dynamic indexing.
+    */
+   if ((!prog->IsES && prog->Version < 130) ||
+       (prog->IsES && prog->Version < 300)) {
+      if (!validate_sampler_array_indexing(ctx, prog))
+         goto done;
+   }
+
   /* Check and validate stream emissions in geometry shaders */
   validate_geometry_shader_emissions(ctx, prog);

--- a/src/glsl/loop_unroll.cpp
+++ b/src/glsl/loop_unroll.cpp
@@ -100,6 +100,18 @@ public:

   virtual ir_visitor_status visit_enter(ir_dereference_array *ir)
   {
+      /* Force unroll in case of dynamic indexing with sampler arrays
+       * when EmitNoIndirectSampler is set.
+       */
+      if (options->EmitNoIndirectSampler) {
+         if ((ir->array->type->is_array() &&
+              ir->array->type->contains_sampler()) &&
+             !ir->array_index->constant_expression_value()) {
+            unsupported_variable_indexing = true;
+            return visit_continue;
+         }
+      }
+
      /* Check for arrays variably-indexed by a loop induction variable.
       * Unrolling the loop may convert that access into constant-indexing.
       *
--- a/src/glsl/nir/nir_opt_algebraic.py
+++ b/src/glsl/nir/nir_opt_algebraic.py
@@ -101,7 +101,7 @@ optimizations = [
   (('ishr', 0, a), 0),
   (('ishr', a, 0), a),
   (('ushr', 0, a), 0),
-   (('ushr', a, 0), 0),
+   (('ushr', a, 0), a),
   # Exponential/logarithmic identities
   (('fexp2', ('flog2', a)), a), # 2^lg2(a) = a
   (('fexp',  ('flog',  a)), a), # e^ln(a)  = a
--- a/Show More
+++ b/Show More
@@ -1 +1 @@
 .5.3
 .5.9